diff --git a/.claude/skills/pdf/SKILL.md b/.claude/skills/pdf/SKILL.md new file mode 100644 index 0000000..5e341c9 --- /dev/null +++ b/.claude/skills/pdf/SKILL.md @@ -0,0 +1,223 @@ +--- +name: pdf +description: Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill. +--- + +# PDF Processing + +## Quick Start + +```python +from pypdf import PdfReader, PdfWriter + +reader = PdfReader("document.pdf") +print(f"Pages: {len(reader.pages)}") + +text = "" +for page in reader.pages: + text += page.extract_text() +``` + +## Python Libraries + +### pypdf — Basic Operations + +**Merge PDFs:** +```python +from pypdf import PdfWriter, PdfReader + +writer = PdfWriter() +for pdf_file in ["doc1.pdf", "doc2.pdf", "doc3.pdf"]: + reader = PdfReader(pdf_file) + for page in reader.pages: + writer.add_page(page) + +with open("merged.pdf", "wb") as output: + writer.write(output) +``` + +**Split PDF:** +```python +reader = PdfReader("input.pdf") +for i, page in enumerate(reader.pages): + writer = PdfWriter() + writer.add_page(page) + with open(f"page_{i+1}.pdf", "wb") as output: + writer.write(output) +``` + +**Extract Metadata:** +```python +reader = PdfReader("document.pdf") +meta = reader.metadata +print(f"Title: {meta.title}") +print(f"Author: {meta.author}") +``` + +**Rotate Pages:** +```python +reader = PdfReader("input.pdf") +writer = PdfWriter() +page = reader.pages[0] +page.rotate(90) +writer.add_page(page) +with open("rotated.pdf", "wb") as output: + writer.write(output) +``` + +### pdfplumber — Text and Table Extraction + +**Extract Text with Layout:** +```python +import pdfplumber + +with pdfplumber.open("document.pdf") as pdf: + for page in pdf.pages: + text = page.extract_text() + print(text) +``` + +**Extract Tables:** +```python +with pdfplumber.open("document.pdf") as pdf: + for i, page in enumerate(pdf.pages): + tables = page.extract_tables() + for j, table in enumerate(tables): + print(f"Table {j+1} on page {i+1}:") + for row in table: + print(row) +``` + +**Extract Tables to DataFrame:** +```python +import pandas as pd +import pdfplumber + +with pdfplumber.open("document.pdf") as pdf: + all_tables = [] + for page in pdf.pages: + tables = page.extract_tables() + for table in tables: + if table: + df = pd.DataFrame(table[1:], columns=table[0]) + all_tables.append(df) + +if all_tables: + combined_df = pd.concat(all_tables, ignore_index=True) + combined_df.to_excel("extracted_tables.xlsx", index=False) +``` + +### reportlab — Create PDFs + +**Basic PDF Creation:** +```python +from reportlab.lib.pagesizes import letter +from reportlab.pdfgen import canvas + +c = canvas.Canvas("hello.pdf", pagesize=letter) +width, height = letter +c.drawString(100, height - 100, "Hello World!") +c.save() +``` + +**Multi-page PDF:** +```python +from reportlab.lib.pagesizes import letter +from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, PageBreak +from reportlab.lib.styles import getSampleStyleSheet + +doc = SimpleDocTemplate("report.pdf", pagesize=letter) +styles = getSampleStyleSheet() +story = [] + +story.append(Paragraph("Report Title", styles['Title'])) +story.append(Spacer(1, 12)) +story.append(Paragraph("Body content here.", styles['Normal'])) +story.append(PageBreak()) +story.append(Paragraph("Page 2", styles['Heading1'])) +doc.build(story) +``` + +⚠️ **IMPORTANT**: Never use Unicode subscript/superscript characters (₀₁₂₃, ⁰¹²³) in ReportLab — use `` and `` tags instead. + +## Command-Line Tools + +**pdftotext (poppler-utils):** +```bash +pdftotext input.pdf output.txt +pdftotext -layout input.pdf output.txt # preserve layout +pdftotext -f 1 -l 5 input.pdf output.txt # pages 1-5 +``` + +**qpdf:** +```bash +qpdf --empty --pages file1.pdf file2.pdf -- merged.pdf +qpdf input.pdf --pages . 1-5 -- pages1-5.pdf +qpdf input.pdf output.pdf --rotate=+90:1 +qpdf --password=mypassword --decrypt encrypted.pdf decrypted.pdf +``` + +## Common Tasks + +**OCR Scanned PDFs:** +```python +import pytesseract +from pdf2image import convert_from_path + +images = convert_from_path('scanned.pdf') +text = "" +for i, image in enumerate(images): + text += f"Page {i+1}:\n" + text += pytesseract.image_to_string(image) + text += "\n\n" +``` + +**Add Watermark:** +```python +from pypdf import PdfReader, PdfWriter + +watermark = PdfReader("watermark.pdf").pages[0] +reader = PdfReader("document.pdf") +writer = PdfWriter() +for page in reader.pages: + page.merge_page(watermark) + writer.add_page(page) +with open("watermarked.pdf", "wb") as output: + writer.write(output) +``` + +**Password Protection:** +```python +from pypdf import PdfReader, PdfWriter + +reader = PdfReader("input.pdf") +writer = PdfWriter() +for page in reader.pages: + writer.add_page(page) +writer.encrypt("userpassword", "ownerpassword") +with open("encrypted.pdf", "wb") as output: + writer.write(output) +``` + +**Extract Images (CLI):** +```bash +pdfimages -j input.pdf output_prefix +``` + +## Quick Reference + +| Task | Best Tool | +|------|-----------| +| Merge PDFs | pypdf | +| Split PDFs | pypdf | +| Extract text | pdfplumber | +| Extract tables | pdfplumber | +| Create PDFs | reportlab | +| CLI merge | qpdf | +| OCR scanned PDFs | pytesseract + pdf2image | +| Fill PDF forms | scripts/ (see forms.md) | + +## References + +- **Form filling**: See [references/forms.md](references/forms.md) — use bundled scripts in `scripts/` +- **Advanced usage** (pypdfium2, pdf-lib JS, batch processing, troubleshooting): See [references/reference.md](references/reference.md) diff --git a/.claude/skills/pdf/references/forms.md b/.claude/skills/pdf/references/forms.md new file mode 100644 index 0000000..d2612b0 --- /dev/null +++ b/.claude/skills/pdf/references/forms.md @@ -0,0 +1,104 @@ +# PDF Form Filling + +## Initial Assessment + +First check if the PDF has fillable form fields: +```bash +python scripts/check_fillable_fields.py file.pdf +``` + +--- + +## For Fillable Forms + +**Step 1 — Extract field info:** +```bash +python scripts/extract_form_field_info.py input.pdf fields.json +``` +Outputs JSON cataloging all fields (IDs, locations, types: text/checkbox/radio/choice). + +**Step 2 — Visually verify (optional):** +```bash +mkdir pages && python scripts/convert_pdf_to_images.py input.pdf pages/ +``` + +**Step 3 — Create field values JSON:** +```json +[ + {"field_id": "FirstName", "page": 1, "value": "John"}, + {"field_id": "Agree", "page": 1, "value": "/Yes"}, + {"field_id": "Gender", "page": 1, "value": "/Male"} +] +``` +- Checkboxes: use the `checked_value` shown in the field info JSON +- Radio groups: use one of the `radio_options[].value` values +- Choice fields: use one of the `choice_options[].value` values + +**Step 4 — Fill the form:** +```bash +python scripts/fill_fillable_fields.py input.pdf field_values.json output.pdf +``` + +--- + +## For Non-Fillable Forms (Text Annotations) + +### Approach A — Preferred (digitally-created PDFs) + +Extract structural coordinates: +```bash +python scripts/extract_form_structure.py input.pdf structure.json +``` + +Build a `fields.json` with `form_fields` array using PDF coordinates: +```json +{ + "pages": [{"page_number": 1, "pdf_width": 612, "pdf_height": 792}], + "form_fields": [ + { + "description": "Name field", + "page_number": 1, + "label_bounding_box": [72, 100, 150, 115], + "entry_bounding_box": [155, 98, 400, 116], + "entry_text": {"text": "John Smith", "font": "Helvetica", "font_size": 11} + } + ] +} +``` + +### Approach B — Fallback (scanned PDFs) + +Convert to images, determine pixel coordinates visually, then use image-based coordinates: +```json +{ + "pages": [{"page_number": 1, "image_width": 1000, "image_height": 1294}], + "form_fields": [...] +} +``` + +### Hybrid + +Use Approach A for most fields and Approach B for anything extract_form_structure misses. + +--- + +## Validation + +Before generating output, validate bounding boxes: +```bash +python scripts/check_bounding_boxes.py fields.json +``` + +Visually verify a page: +```bash +python scripts/create_validation_image.py 1 fields.json pages/page_1.png validation_page_1.png +``` + +Red boxes = entry areas, Blue boxes = label areas. + +**Fill with annotations:** +```bash +python scripts/fill_pdf_form_with_annotations.py input.pdf fields.json output.pdf +``` + +Then convert output to images to verify text placement. diff --git a/.claude/skills/pdf/references/reference.md b/.claude/skills/pdf/references/reference.md new file mode 100644 index 0000000..36a14b0 --- /dev/null +++ b/.claude/skills/pdf/references/reference.md @@ -0,0 +1,174 @@ +# PDF Advanced Reference + +## pypdfium2 — High-Performance Rendering + +```python +import pypdfium2 as pdfium + +pdf = pdfium.PdfDocument("document.pdf") + +# Render page to image +page = pdf[0] +bitmap = page.render(scale=2) # 2x scale = 144 DPI +pil_image = bitmap.to_pil() +pil_image.save("page_1.png") + +# Extract text with coordinates +textpage = page.get_textpage() +for i in range(textpage.count_chars()): + char = textpage.get_charbox(i) + print(f"Char: {textpage.get_text_range(i, 1)}, Box: {char}") + +pdf.close() +``` + +**When to prefer pypdfium2**: High-quality image rendering, pixel-accurate text extraction, processing large batches. + +--- + +## pdf-lib (JavaScript) + +```javascript +import { PDFDocument, rgb, StandardFonts } from 'pdf-lib'; +import fs from 'fs'; + +// Load and modify +const pdfBytes = fs.readFileSync('input.pdf'); +const pdfDoc = await PDFDocument.load(pdfBytes); + +const page = pdfDoc.getPages()[0]; +const font = await pdfDoc.embedFont(StandardFonts.Helvetica); +page.drawText('Hello World', { x: 50, y: 700, size: 20, font, color: rgb(0, 0, 0) }); + +const output = await pdfDoc.save(); +fs.writeFileSync('output.pdf', output); +``` + +**When to use**: Node.js environments, browser-side PDF generation. + +--- + +## Batch Processing + +```python +import os +from pathlib import Path +from pypdf import PdfReader +import logging + +logging.basicConfig(level=logging.INFO) + +def process_pdfs(input_dir: str, output_dir: str): + Path(output_dir).mkdir(exist_ok=True) + errors = [] + + for pdf_file in Path(input_dir).glob("*.pdf"): + try: + reader = PdfReader(str(pdf_file)) + text = "\n".join(page.extract_text() or "" for page in reader.pages) + + out_path = Path(output_dir) / (pdf_file.stem + ".txt") + out_path.write_text(text, encoding="utf-8") + logging.info(f"Processed: {pdf_file.name}") + except Exception as e: + logging.error(f"Failed: {pdf_file.name} — {e}") + errors.append((str(pdf_file), str(e))) + + return errors +``` + +--- + +## Memory-Efficient Large File Processing + +```python +import pdfplumber + +def stream_text(pdf_path: str): + """Yield text page by page to avoid loading entire PDF into memory.""" + with pdfplumber.open(pdf_path) as pdf: + for i, page in enumerate(pdf.pages): + yield i + 1, page.extract_text() or "" + +for page_num, text in stream_text("large_document.pdf"): + print(f"--- Page {page_num} ---") + print(text) +``` + +--- + +## Handling Encrypted / Corrupted PDFs + +```python +from pypdf import PdfReader + +# Encrypted PDF +reader = PdfReader("encrypted.pdf") +if reader.is_encrypted: + reader.decrypt("password") + +# Corrupted PDF — try qpdf first +import subprocess +result = subprocess.run( + ["qpdf", "--recover", "corrupted.pdf", "repaired.pdf"], + capture_output=True, text=True +) +if result.returncode != 0: + print("qpdf repair failed:", result.stderr) +``` + +--- + +## OCR Fallback for Mixed PDFs + +```python +import pdfplumber +import pytesseract +from pdf2image import convert_from_path + +def extract_with_ocr_fallback(pdf_path: str) -> list[str]: + """Use pdfplumber first; fall back to OCR if page has no text.""" + pages_text = [] + images = None # lazy-load + + with pdfplumber.open(pdf_path) as pdf: + for i, page in enumerate(pdf.pages): + text = page.extract_text() + if text and text.strip(): + pages_text.append(text) + else: + if images is None: + images = convert_from_path(pdf_path, dpi=200) + pages_text.append(pytesseract.image_to_string(images[i])) + + return pages_text +``` + +--- + +## Tool Selection Guide + +| Use case | Recommended tool | +|----------|-----------------| +| Simple text extraction | pdfplumber | +| Tables | pdfplumber | +| High-quality image rendering | pypdfium2 | +| Merge / split / rotate | pypdf | +| Create new PDFs | reportlab | +| JavaScript / browser | pdf-lib | +| CLI text extraction | pdftotext (poppler) | +| CLI manipulation | qpdf | +| OCR | pytesseract + pdf2image | +| Repair corrupted | qpdf --recover | + +## Library Licenses + +| Library | License | +|---------|---------| +| pypdf | BSD | +| pdfplumber | MIT | +| reportlab | BSD | +| pypdfium2 | Apache 2.0 | +| pdf-lib | MIT | +| poppler-utils | GPL-2 | +| qpdf | Apache 2.0 | diff --git a/.claude/skills/pdf/scripts/check_bounding_boxes.py b/.claude/skills/pdf/scripts/check_bounding_boxes.py new file mode 100644 index 0000000..2fc9876 --- /dev/null +++ b/.claude/skills/pdf/scripts/check_bounding_boxes.py @@ -0,0 +1,64 @@ +from dataclasses import dataclass +import json +import sys + + +@dataclass +class RectAndField: + rect: list[float] + rect_type: str + field: dict + + +def get_bounding_box_messages(fields_json_stream) -> list[str]: + messages = [] + fields = json.load(fields_json_stream) + messages.append(f"Read {len(fields['form_fields'])} fields") + + def rects_intersect(r1, r2): + disjoint_horizontal = r1[0] >= r2[2] or r1[2] <= r2[0] + disjoint_vertical = r1[1] >= r2[3] or r1[3] <= r2[1] + return not (disjoint_horizontal or disjoint_vertical) + + rects_and_fields = [] + for f in fields["form_fields"]: + rects_and_fields.append(RectAndField(f["label_bounding_box"], "label", f)) + rects_and_fields.append(RectAndField(f["entry_bounding_box"], "entry", f)) + + has_error = False + for i, ri in enumerate(rects_and_fields): + for j in range(i + 1, len(rects_and_fields)): + rj = rects_and_fields[j] + if ri.field["page_number"] == rj.field["page_number"] and rects_intersect(ri.rect, rj.rect): + has_error = True + if ri.field is rj.field: + messages.append(f"FAILURE: intersection between label and entry bounding boxes for `{ri.field['description']}` ({ri.rect}, {rj.rect})") + else: + messages.append(f"FAILURE: intersection between {ri.rect_type} bounding box for `{ri.field['description']}` ({ri.rect}) and {rj.rect_type} bounding box for `{rj.field['description']}` ({rj.rect})") + if len(messages) >= 20: + messages.append("Aborting further checks; fix bounding boxes and try again") + return messages + if ri.rect_type == "entry": + if "entry_text" in ri.field: + font_size = ri.field["entry_text"].get("font_size", 14) + entry_height = ri.rect[3] - ri.rect[1] + if entry_height < font_size: + has_error = True + messages.append(f"FAILURE: entry bounding box height ({entry_height}) for `{ri.field['description']}` is too short for the text content (font size: {font_size}). Increase the box height or decrease the font size.") + if len(messages) >= 20: + messages.append("Aborting further checks; fix bounding boxes and try again") + return messages + + if not has_error: + messages.append("SUCCESS: All bounding boxes are valid") + return messages + + +if __name__ == "__main__": + if len(sys.argv) != 2: + print("Usage: check_bounding_boxes.py [fields.json]") + sys.exit(1) + with open(sys.argv[1]) as f: + messages = get_bounding_box_messages(f) + for msg in messages: + print(msg) diff --git a/.claude/skills/pdf/scripts/check_fillable_fields.py b/.claude/skills/pdf/scripts/check_fillable_fields.py new file mode 100644 index 0000000..ddeff88 --- /dev/null +++ b/.claude/skills/pdf/scripts/check_fillable_fields.py @@ -0,0 +1,9 @@ +import sys +from pypdf import PdfReader + + +reader = PdfReader(sys.argv[1]) +if reader.get_fields(): + print("This PDF has fillable form fields") +else: + print("This PDF does not have fillable form fields; you will need to visually determine where to enter data") diff --git a/.claude/skills/pdf/scripts/convert_pdf_to_images.py b/.claude/skills/pdf/scripts/convert_pdf_to_images.py new file mode 100644 index 0000000..c35a760 --- /dev/null +++ b/.claude/skills/pdf/scripts/convert_pdf_to_images.py @@ -0,0 +1,31 @@ +import os +import sys + +from pdf2image import convert_from_path + + +def convert(pdf_path, output_dir, max_dim=1000): + images = convert_from_path(pdf_path, dpi=200) + + for i, image in enumerate(images): + width, height = image.size + if width > max_dim or height > max_dim: + scale_factor = min(max_dim / width, max_dim / height) + new_width = int(width * scale_factor) + new_height = int(height * scale_factor) + image = image.resize((new_width, new_height)) + + image_path = os.path.join(output_dir, f"page_{i+1}.png") + image.save(image_path) + print(f"Saved page {i+1} as {image_path} (size: {image.size})") + + print(f"Converted {len(images)} pages to PNG images") + + +if __name__ == "__main__": + if len(sys.argv) != 3: + print("Usage: convert_pdf_to_images.py [input pdf] [output directory]") + sys.exit(1) + pdf_path = sys.argv[1] + output_directory = sys.argv[2] + convert(pdf_path, output_directory) diff --git a/.claude/skills/pdf/scripts/create_validation_image.py b/.claude/skills/pdf/scripts/create_validation_image.py new file mode 100644 index 0000000..8a41eac --- /dev/null +++ b/.claude/skills/pdf/scripts/create_validation_image.py @@ -0,0 +1,35 @@ +import json +import sys + +from PIL import Image, ImageDraw + + +def create_validation_image(page_number, fields_json_path, input_path, output_path): + with open(fields_json_path, 'r') as f: + data = json.load(f) + + img = Image.open(input_path) + draw = ImageDraw.Draw(img) + num_boxes = 0 + + for field in data["form_fields"]: + if field["page_number"] == page_number: + entry_box = field['entry_bounding_box'] + label_box = field['label_bounding_box'] + draw.rectangle(entry_box, outline='red', width=2) + draw.rectangle(label_box, outline='blue', width=2) + num_boxes += 2 + + img.save(output_path) + print(f"Created validation image at {output_path} with {num_boxes} bounding boxes") + + +if __name__ == "__main__": + if len(sys.argv) != 5: + print("Usage: create_validation_image.py [page number] [fields.json file] [input image path] [output image path]") + sys.exit(1) + page_number = int(sys.argv[1]) + fields_json_path = sys.argv[2] + input_image_path = sys.argv[3] + output_image_path = sys.argv[4] + create_validation_image(page_number, fields_json_path, input_image_path, output_image_path) diff --git a/.claude/skills/pdf/scripts/extract_form_field_info.py b/.claude/skills/pdf/scripts/extract_form_field_info.py new file mode 100644 index 0000000..59b7bb6 --- /dev/null +++ b/.claude/skills/pdf/scripts/extract_form_field_info.py @@ -0,0 +1,119 @@ +import json +import sys + +from pypdf import PdfReader + + +def get_full_annotation_field_id(annotation): + components = [] + while annotation: + field_name = annotation.get('/T') + if field_name: + components.append(field_name) + annotation = annotation.get('/Parent') + return ".".join(reversed(components)) if components else None + + +def make_field_dict(field, field_id): + field_dict = {"field_id": field_id} + ft = field.get('/FT') + if ft == "/Tx": + field_dict["type"] = "text" + elif ft == "/Btn": + field_dict["type"] = "checkbox" + states = field.get("/_States_", []) + if len(states) == 2: + if "/Off" in states: + field_dict["checked_value"] = states[0] if states[0] != "/Off" else states[1] + field_dict["unchecked_value"] = "/Off" + else: + print(f"Unexpected state values for checkbox `${field_id}`. Its checked and unchecked values may not be correct; if you're trying to check it, visually verify the results.") + field_dict["checked_value"] = states[0] + field_dict["unchecked_value"] = states[1] + elif ft == "/Ch": + field_dict["type"] = "choice" + states = field.get("/_States_", []) + field_dict["choice_options"] = [{ + "value": state[0], + "text": state[1], + } for state in states] + else: + field_dict["type"] = f"unknown ({ft})" + return field_dict + + +def get_field_info(reader: PdfReader): + fields = reader.get_fields() + + field_info_by_id = {} + possible_radio_names = set() + + for field_id, field in fields.items(): + if field.get("/Kids"): + if field.get("/FT") == "/Btn": + possible_radio_names.add(field_id) + continue + field_info_by_id[field_id] = make_field_dict(field, field_id) + + radio_fields_by_id = {} + + for page_index, page in enumerate(reader.pages): + annotations = page.get('/Annots', []) + for ann in annotations: + field_id = get_full_annotation_field_id(ann) + if field_id in field_info_by_id: + field_info_by_id[field_id]["page"] = page_index + 1 + field_info_by_id[field_id]["rect"] = ann.get('/Rect') + elif field_id in possible_radio_names: + try: + on_values = [v for v in ann["/AP"]["/N"] if v != "/Off"] + except KeyError: + continue + if len(on_values) == 1: + rect = ann.get("/Rect") + if field_id not in radio_fields_by_id: + radio_fields_by_id[field_id] = { + "field_id": field_id, + "type": "radio_group", + "page": page_index + 1, + "radio_options": [], + } + radio_fields_by_id[field_id]["radio_options"].append({ + "value": on_values[0], + "rect": rect, + }) + + fields_with_location = [] + for field_info in field_info_by_id.values(): + if "page" in field_info: + fields_with_location.append(field_info) + else: + print(f"Unable to determine location for field id: {field_info.get('field_id')}, ignoring") + + def sort_key(f): + if "radio_options" in f: + rect = f["radio_options"][0]["rect"] or [0, 0, 0, 0] + else: + rect = f.get("rect") or [0, 0, 0, 0] + adjusted_position = [-rect[1], rect[0]] + return [f.get("page"), adjusted_position] + + sorted_fields = fields_with_location + list(radio_fields_by_id.values()) + sorted_fields.sort(key=sort_key) + + return sorted_fields + + +def write_field_info(pdf_path: str, json_output_path: str): + reader = PdfReader(pdf_path) + field_info = get_field_info(reader) + with open(json_output_path, "w") as f: + json.dump(field_info, f, indent=2) + print(f"Wrote {len(field_info)} fields to {json_output_path}") + + +if __name__ == "__main__": + if len(sys.argv) != 3: + print("Usage: extract_form_field_info.py [input pdf] [output json]") + sys.exit(1) + write_field_info(sys.argv[1], sys.argv[2]) diff --git a/.claude/skills/pdf/scripts/extract_form_structure.py b/.claude/skills/pdf/scripts/extract_form_structure.py new file mode 100644 index 0000000..5e175d9 --- /dev/null +++ b/.claude/skills/pdf/scripts/extract_form_structure.py @@ -0,0 +1,101 @@ +import json +import sys +import pdfplumber + + +def extract_form_structure(pdf_path): + structure = { + "pages": [], + "labels": [], + "lines": [], + "checkboxes": [], + "row_boundaries": [] + } + + with pdfplumber.open(pdf_path) as pdf: + for page_num, page in enumerate(pdf.pages, 1): + structure["pages"].append({ + "page_number": page_num, + "width": float(page.width), + "height": float(page.height) + }) + + words = page.extract_words() + for word in words: + structure["labels"].append({ + "page": page_num, + "text": word["text"], + "x0": round(float(word["x0"]), 1), + "top": round(float(word["top"]), 1), + "x1": round(float(word["x1"]), 1), + "bottom": round(float(word["bottom"]), 1) + }) + + for line in page.lines: + if abs(float(line["x1"]) - float(line["x0"])) > page.width * 0.5: + structure["lines"].append({ + "page": page_num, + "y": round(float(line["top"]), 1), + "x0": round(float(line["x0"]), 1), + "x1": round(float(line["x1"]), 1) + }) + + for rect in page.rects: + width = float(rect["x1"]) - float(rect["x0"]) + height = float(rect["bottom"]) - float(rect["top"]) + if 5 <= width <= 15 and 5 <= height <= 15 and abs(width - height) < 2: + structure["checkboxes"].append({ + "page": page_num, + "x0": round(float(rect["x0"]), 1), + "top": round(float(rect["top"]), 1), + "x1": round(float(rect["x1"]), 1), + "bottom": round(float(rect["bottom"]), 1), + "center_x": round((float(rect["x0"]) + float(rect["x1"])) / 2, 1), + "center_y": round((float(rect["top"]) + float(rect["bottom"])) / 2, 1) + }) + + lines_by_page = {} + for line in structure["lines"]: + page = line["page"] + if page not in lines_by_page: + lines_by_page[page] = [] + lines_by_page[page].append(line["y"]) + + for page, y_coords in lines_by_page.items(): + y_coords = sorted(set(y_coords)) + for i in range(len(y_coords) - 1): + structure["row_boundaries"].append({ + "page": page, + "row_top": y_coords[i], + "row_bottom": y_coords[i + 1], + "row_height": round(y_coords[i + 1] - y_coords[i], 1) + }) + + return structure + + +def main(): + if len(sys.argv) != 3: + print("Usage: extract_form_structure.py ") + sys.exit(1) + + pdf_path = sys.argv[1] + output_path = sys.argv[2] + + print(f"Extracting structure from {pdf_path}...") + structure = extract_form_structure(pdf_path) + + with open(output_path, "w") as f: + json.dump(structure, f, indent=2) + + print(f"Found:") + print(f" - {len(structure['pages'])} pages") + print(f" - {len(structure['labels'])} text labels") + print(f" - {len(structure['lines'])} horizontal lines") + print(f" - {len(structure['checkboxes'])} checkboxes") + print(f" - {len(structure['row_boundaries'])} row boundaries") + print(f"Saved to {output_path}") + + +if __name__ == "__main__": + main() diff --git a/.claude/skills/pdf/scripts/fill_fillable_fields.py b/.claude/skills/pdf/scripts/fill_fillable_fields.py new file mode 100644 index 0000000..a706cd2 --- /dev/null +++ b/.claude/skills/pdf/scripts/fill_fillable_fields.py @@ -0,0 +1,96 @@ +import json +import sys + +from pypdf import PdfReader, PdfWriter + +from extract_form_field_info import get_field_info + + +def fill_pdf_fields(input_pdf_path: str, fields_json_path: str, output_pdf_path: str): + with open(fields_json_path) as f: + fields = json.load(f) + fields_by_page = {} + for field in fields: + if "value" in field: + field_id = field["field_id"] + page = field["page"] + if page not in fields_by_page: + fields_by_page[page] = {} + fields_by_page[page][field_id] = field["value"] + + reader = PdfReader(input_pdf_path) + + has_error = False + field_info = get_field_info(reader) + fields_by_ids = {f["field_id"]: f for f in field_info} + for field in fields: + existing_field = fields_by_ids.get(field["field_id"]) + if not existing_field: + has_error = True + print(f"ERROR: `{field['field_id']}` is not a valid field ID") + elif field["page"] != existing_field["page"]: + has_error = True + print(f"ERROR: Incorrect page number for `{field['field_id']}` (got {field['page']}, expected {existing_field['page']})") + else: + if "value" in field: + err = validation_error_for_field_value(existing_field, field["value"]) + if err: + print(err) + has_error = True + if has_error: + sys.exit(1) + + writer = PdfWriter(clone_from=reader) + for page, field_values in fields_by_page.items(): + writer.update_page_form_field_values(writer.pages[page - 1], field_values, auto_regenerate=False) + + writer.set_need_appearances_writer(True) + + with open(output_pdf_path, "wb") as f: + writer.write(f) + + +def validation_error_for_field_value(field_info, field_value): + field_type = field_info["type"] + field_id = field_info["field_id"] + if field_type == "checkbox": + checked_val = field_info["checked_value"] + unchecked_val = field_info["unchecked_value"] + if field_value != checked_val and field_value != unchecked_val: + return f'ERROR: Invalid value "{field_value}" for checkbox field "{field_id}". The checked value is "{checked_val}" and the unchecked value is "{unchecked_val}"' + elif field_type == "radio_group": + option_values = [opt["value"] for opt in field_info["radio_options"]] + if field_value not in option_values: + return f'ERROR: Invalid value "{field_value}" for radio group field "{field_id}". Valid values are: {option_values}' + elif field_type == "choice": + choice_values = [opt["value"] for opt in field_info["choice_options"]] + if field_value not in choice_values: + return f'ERROR: Invalid value "{field_value}" for choice field "{field_id}". Valid values are: {choice_values}' + return None + + +def monkeypatch_pypdf_method(): + from pypdf.generic import DictionaryObject + from pypdf.constants import FieldDictionaryAttributes + + original_get_inherited = DictionaryObject.get_inherited + + def patched_get_inherited(self, key: str, default=None): + result = original_get_inherited(self, key, default) + if key == FieldDictionaryAttributes.Opt: + if isinstance(result, list) and all(isinstance(v, list) and len(v) == 2 for v in result): + result = [r[0] for r in result] + return result + + DictionaryObject.get_inherited = patched_get_inherited + + +if __name__ == "__main__": + if len(sys.argv) != 4: + print("Usage: fill_fillable_fields.py [input pdf] [field_values.json] [output pdf]") + sys.exit(1) + monkeypatch_pypdf_method() + input_pdf = sys.argv[1] + fields_json = sys.argv[2] + output_pdf = sys.argv[3] + fill_pdf_fields(input_pdf, fields_json, output_pdf) diff --git a/.claude/skills/pdf/scripts/fill_pdf_form_with_annotations.py b/.claude/skills/pdf/scripts/fill_pdf_form_with_annotations.py new file mode 100644 index 0000000..a2ac423 --- /dev/null +++ b/.claude/skills/pdf/scripts/fill_pdf_form_with_annotations.py @@ -0,0 +1,104 @@ +import json +import sys + +from pypdf import PdfReader, PdfWriter +from pypdf.annotations import FreeText + + +def transform_from_image_coords(bbox, image_width, image_height, pdf_width, pdf_height): + x_scale = pdf_width / image_width + y_scale = pdf_height / image_height + + left = bbox[0] * x_scale + right = bbox[2] * x_scale + + top = pdf_height - (bbox[1] * y_scale) + bottom = pdf_height - (bbox[3] * y_scale) + + return left, bottom, right, top + + +def transform_from_pdf_coords(bbox, pdf_height): + left = bbox[0] + right = bbox[2] + + pypdf_top = pdf_height - bbox[1] + pypdf_bottom = pdf_height - bbox[3] + + return left, pypdf_bottom, right, pypdf_top + + +def fill_pdf_form(input_pdf_path, fields_json_path, output_pdf_path): + with open(fields_json_path, "r") as f: + fields_data = json.load(f) + + reader = PdfReader(input_pdf_path) + writer = PdfWriter() + + writer.append(reader) + + pdf_dimensions = {} + for i, page in enumerate(reader.pages): + mediabox = page.mediabox + pdf_dimensions[i + 1] = [mediabox.width, mediabox.height] + + annotations = [] + for field in fields_data["form_fields"]: + page_num = field["page_number"] + + page_info = next(p for p in fields_data["pages"] if p["page_number"] == page_num) + pdf_width, pdf_height = pdf_dimensions[page_num] + + if "pdf_width" in page_info: + transformed_entry_box = transform_from_pdf_coords( + field["entry_bounding_box"], + float(pdf_height) + ) + else: + image_width = page_info["image_width"] + image_height = page_info["image_height"] + transformed_entry_box = transform_from_image_coords( + field["entry_bounding_box"], + image_width, image_height, + float(pdf_width), float(pdf_height) + ) + + if "entry_text" not in field or "text" not in field["entry_text"]: + continue + entry_text = field["entry_text"] + text = entry_text["text"] + if not text: + continue + + font_name = entry_text.get("font", "Arial") + font_size = str(entry_text.get("font_size", 14)) + "pt" + font_color = entry_text.get("font_color", "000000") + + annotation = FreeText( + text=text, + rect=transformed_entry_box, + font=font_name, + font_size=font_size, + font_color=font_color, + border_color=None, + background_color=None, + ) + annotations.append(annotation) + writer.add_annotation(page_number=page_num - 1, annotation=annotation) + + with open(output_pdf_path, "wb") as output: + writer.write(output) + + print(f"Successfully filled PDF form and saved to {output_pdf_path}") + print(f"Added {len(annotations)} text annotations") + + +if __name__ == "__main__": + if len(sys.argv) != 4: + print("Usage: fill_pdf_form_with_annotations.py [input pdf] [fields.json] [output pdf]") + sys.exit(1) + input_pdf = sys.argv[1] + fields_json = sys.argv[2] + output_pdf = sys.argv[3] + + fill_pdf_form(input_pdf, fields_json, output_pdf) diff --git a/book/myst.yml b/book/myst.yml index 617cab2..da3ba92 100644 --- a/book/myst.yml +++ b/book/myst.yml @@ -23,6 +23,8 @@ project: file: why_causal_inference/why_causal_inference_en.ipynb - file: why_causal_inference/why_causal_inference_ko.ipynb hidden: true + - title: Who Should Be Treated? + file: who_should_be_treated/who_should_be_treated_ko.ipynb site: template: book-theme diff --git a/book/who_should_be_treated/data/multi_attribution_sample.csv b/book/who_should_be_treated/data/multi_attribution_sample.csv new file mode 100644 index 0000000..a629466 --- /dev/null +++ b/book/who_should_be_treated/data/multi_attribution_sample.csv @@ -0,0 +1,2001 @@ +Global Flag,Major Flag,SMC Flag,Commercial Flag,IT Spend,Employee Count,PC Count,Size,Tech Support,Discount,Revenue +1,0,1,0,45537,26,26,152205,0,1,17688.363 +0,0,1,1,20842,107,70,159038,0,1,14981.43559 +0,0,0,1,82171,10,7,264935,1,1,32917.13894 +0,0,0,0,30288,40,39,77522,1,1,14773.76855 +0,0,1,0,25930,37,43,91446,1,1,17098.69823 +0,0,1,0,34597,44,51,218703,1,0,17280.70918 +1,0,0,0,40199,21,14,126342,0,0,9153.974376 +0,0,1,1,30454,11,8,96784,0,0,5760.075096 +0,0,0,1,23428,50,55,77298,1,1,17265.11146 +0,0,1,1,7970,131,153,77888,1,1,21953.37179 +0,0,0,1,151143,84,58,417289,1,1,50875.56035 +1,0,1,1,30759,169,136,116016,1,0,22153.82215 +0,0,1,1,19899,279,195,66555,0,0,11972.18818 +0,0,1,1,8420,29,26,36973,1,0,10063.85753 +0,0,1,1,34577,20,15,91689,1,0,10521.9988 +0,0,0,1,1294,18,15,11292,0,1,2296.739654 +1,0,1,1,80629,13,9,466694,0,1,41743.72206 +1,0,1,0,144013,119,124,413240,1,1,56307.34264 +1,0,1,1,17143,24,24,49673,0,1,9673.221778 +0,0,1,1,22729,110,110,117138,0,1,15397.70575 +0,0,1,1,28856,34,34,136831,1,1,21267.51393 +0,0,0,0,14057,24,18,67434,0,1,6321.956387 +0,0,1,0,14682,13,16,45908,0,0,2694.216001 +0,0,0,0,8321,82,50,26999,0,0,2531.466231 +0,0,1,0,22051,170,140,56512,0,0,9116.230185 +0,0,1,0,15052,156,154,41571,0,1,12387.81871 +1,0,0,0,21432,106,122,148865,1,1,27860.93773 +0,0,0,0,21232,221,146,84567,0,0,9072.07955 +0,0,1,1,41352,15,11,126934,1,1,19867.56935 +0,0,1,1,12744,10,10,54468,0,0,2868.606553 +0,0,1,0,13980,46,32,58046,0,1,7126.30626 +1,1,0,0,35303,108,136,120173,0,1,20237.79168 +1,0,1,1,10923,26,21,74202,0,1,11610.05857 +0,0,0,1,16662,11,10,110110,1,1,18647.91666 +0,0,0,1,88422,34,27,295436,1,1,37025.88584 +0,0,0,0,3971,41,30,29867,0,0,1038.615975 +0,0,1,0,22692,17,15,133197,1,1,19821.22882 +0,0,1,1,31488,17,10,112661,0,1,13903.71762 +1,1,0,0,45422,45,56,193432,0,0,15182.77493 +0,0,0,1,62739,108,131,250338,1,0,25477.56325 +1,0,1,0,2168,48,47,11009,1,1,13173.24059 +0,1,0,1,20901,85,80,115902,1,0,15864.39188 +1,0,1,1,4857,44,48,16491,0,1,8488.73425 +0,1,0,0,25002,34,22,99114,1,0,13303.01519 +0,0,0,1,23904,32,20,176769,1,0,15658.03599 +0,0,1,0,5972,28,26,16210,0,1,1866.574021 +0,0,0,1,19486,22,22,116783,0,1,11262.22719 +0,0,1,0,16858,63,45,99214,1,1,18338.09893 +0,0,1,1,37432,98,76,272286,1,1,36192.4073 +0,1,0,1,32794,76,47,101073,0,0,8797.224912 +0,0,0,1,21666,13,14,59452,0,0,3346.04696 +0,0,0,1,2272,33,34,13700,0,0,2984.716846 +0,0,0,1,23677,58,65,93466,1,1,18338.7401 +1,0,1,0,7813,15,15,27740,0,0,5350.908641 +0,0,1,1,117098,13,10,535638,1,1,60479.55167 +0,0,0,1,6590,34,24,19900,0,0,3312.640665 +0,0,0,1,17049,61,62,60369,1,0,12021.49536 +0,1,0,0,88501,127,152,417564,1,1,54047.43047 +1,0,1,1,2937,57,68,10629,0,0,8835.047329 +1,1,0,0,28651,30,28,106147,1,1,23904.30481 +0,1,0,1,7587,50,32,69779,1,1,15366.9466 +0,1,0,1,80706,141,175,244713,0,0,18149.13843 +0,0,0,1,4672,33,24,11762,0,1,3087.227865 +0,0,1,0,41677,26,32,205525,0,1,19301.61921 +0,1,0,0,61286,31,36,179022,1,0,18214.01619 +0,0,1,0,157655,15,18,429946,1,0,28356.19404 +0,1,0,1,24376,20,12,165690,0,1,17103.73349 +0,1,0,1,78105,31,28,429628,1,1,52180.4918 +1,1,0,0,10468,13,14,54499,1,1,17418.02118 +0,0,0,1,14156,59,63,40721,0,1,6480.237874 +1,0,1,0,11328,37,37,56684,1,0,14732.56658 +0,0,1,0,9291,27,33,39421,0,1,6212.661339 +0,0,1,0,47794,51,63,310950,1,1,37274.32798 +0,0,1,1,25131,34,35,129251,0,1,13546.97187 +0,0,0,0,17376,78,60,72051,0,1,6934.633026 +0,0,0,1,45245,34,23,242353,0,1,22886.89951 +0,0,1,1,50999,48,59,128288,1,0,16289.16582 +0,0,0,0,2867,20,16,12868,0,0,427.6482735 +0,0,1,1,67890,15,13,406365,1,1,47587.29878 +1,1,0,1,160244,98,124,454291,1,1,63162.5783 +1,0,0,1,7278,103,93,43359,0,0,8583.56821 +0,0,1,0,61684,14,11,315315,1,1,37902.30016 +0,0,1,1,23501,101,74,59631,0,0,5172.847451 +0,0,1,1,8771,317,268,27460,1,0,20175.58992 +0,0,1,0,10076,27,29,84398,1,1,13690.74117 +0,1,0,1,6669,33,25,34115,1,1,11546.04042 +0,0,1,1,43708,10,7,126892,1,0,13310.68181 +0,0,1,1,24160,32,25,181818,1,0,15695.07674 +0,0,1,1,22691,57,52,149726,1,1,22908.60641 +0,0,1,1,150289,53,56,377721,1,0,28330.76253 +0,1,0,1,24836,120,92,145100,1,0,17501.09664 +0,0,1,0,66191,84,58,176803,1,1,26121.34126 +1,0,1,0,24416,17,18,102243,1,1,18963.73396 +1,0,0,1,6203,38,47,31973,0,1,8998.929994 +0,0,1,1,3176,17,13,21184,0,1,4208.953344 +1,0,0,0,20620,61,68,64891,0,1,11729.14754 +0,0,1,1,10130,42,45,35605,0,1,5698.071425 +1,0,1,0,2082,23,24,11269,1,0,10750.36634 +1,1,0,1,32247,21,24,86154,0,0,10332.25678 +0,0,0,0,4531,83,84,12688,0,0,5205.635952 +0,0,1,1,54076,64,76,211136,1,1,29938.98853 +0,0,1,1,8030,51,50,22089,0,0,3892.89282 +0,1,0,0,22105,58,48,92998,0,1,10775.65117 +0,0,1,0,2446,70,87,12772,1,1,8732.876532 +0,0,0,1,68678,17,18,235152,1,1,30804.12152 +0,0,1,1,28402,143,164,110430,0,1,18359.92468 +0,0,1,1,33525,86,65,98371,1,1,19446.96031 +1,0,1,1,32296,60,51,134989,0,0,10618.8936 +0,1,0,1,25297,27,24,227115,1,0,20154.21729 +0,0,1,1,10143,31,24,81870,1,0,11441.82429 +0,0,0,0,2517,27,17,21453,1,1,8533.242525 +0,0,1,1,8405,73,52,26417,0,0,3883.200484 +1,0,0,1,71141,28,24,210928,0,0,13785.74983 +0,0,1,1,75907,50,42,208770,0,1,20463.75335 +0,0,0,0,2337,159,183,17870,0,0,7309.382264 +1,1,0,1,16756,35,31,70823,1,0,16577.22067 +0,0,0,1,4359,55,48,30794,1,1,11868.75457 +0,1,0,1,22628,14,16,99704,0,1,11800.72088 +0,1,0,1,52238,14,10,150085,1,0,16800.52139 +0,0,0,1,31620,125,146,108026,0,0,11742.24034 +1,0,1,1,20553,174,178,58863,1,0,20339.32765 +0,0,1,1,46828,21,14,121378,1,1,19690.81724 +0,1,0,1,10302,103,98,36116,0,0,8061.36729 +0,0,0,1,11896,20,12,58290,0,1,7112.372385 +0,0,1,0,27461,88,86,77141,1,1,16303.77806 +1,1,0,1,138371,154,95,391156,1,1,55238.15934 +0,1,0,0,17406,14,8,80062,0,1,8856.093234 +0,0,0,1,6672,175,205,17437,0,0,10980.00749 +0,0,0,1,5733,24,20,15552,0,0,1263.065792 +0,1,0,1,4927,71,68,37194,1,1,14424.33442 +0,0,0,0,7712,113,143,35022,0,1,7429.766914 +0,0,0,1,23011,95,68,61522,1,0,10737.71349 +0,0,1,1,37491,17,18,272112,0,1,23602.45932 +0,0,1,0,6444,22,27,46013,1,1,9932.041465 +0,0,1,1,3984,16,10,15835,0,1,4004.123217 +0,1,0,1,94564,59,69,288104,1,0,24503.26207 +0,0,1,0,43904,23,21,199942,0,1,17650.28582 +1,0,1,1,18953,71,53,68308,0,0,9877.285199 +0,0,1,1,22096,30,34,79186,1,1,15465.93262 +0,0,1,1,8555,31,39,31220,1,0,10514.12972 +0,0,0,0,34368,78,52,118289,1,0,14608.44238 +0,0,1,1,15294,120,75,40569,0,0,5958.755283 +0,1,0,0,34126,69,61,187211,0,1,20432.53813 +0,0,1,0,16562,77,66,85917,1,0,12823.25034 +0,1,0,1,9700,109,141,24838,0,1,12434.5908 +1,0,0,0,70265,34,27,330589,1,1,43556.70504 +0,0,0,1,2040,57,68,12825,1,0,10518.12191 +0,0,0,1,8710,105,68,23341,0,1,7931.258914 +0,1,0,1,12776,90,62,41423,1,1,15312.82009 +0,0,0,1,156713,14,13,677214,1,0,40166.67407 +1,1,0,1,3954,20,12,21408,0,0,8699.634473 +0,0,1,1,38797,19,17,326943,0,0,9399.076282 +0,0,1,1,9867,18,11,27353,0,0,2588.025236 +0,1,0,1,11356,11,13,75715,1,1,16257.93618 +0,0,0,1,18931,145,164,50727,0,1,12850.98872 +1,0,0,1,7671,57,35,21975,1,0,11839.8829 +1,0,1,1,17200,150,110,114140,0,1,18577.31238 +1,0,1,1,60359,20,24,459345,1,1,55748.82324 +0,1,0,1,25135,55,45,131549,0,0,8435.037353 +0,0,1,0,2298,19,12,11375,1,0,8415.235112 +0,0,0,0,48536,37,33,335333,1,1,38288.29422 +0,0,0,1,19084,18,19,55362,0,1,6461.177574 +0,0,1,1,12220,66,57,63430,0,0,6489.251169 +0,0,1,0,33933,60,70,157628,0,0,10567.21174 +0,0,1,1,14702,17,16,123893,0,0,5898.268951 +0,0,0,1,27897,35,25,73351,0,0,3352.079607 +1,0,1,1,52982,22,19,159086,0,1,19184.64834 +0,0,1,1,13605,206,156,62175,1,1,19216.95099 +0,0,1,1,5764,53,43,19948,0,1,5020.076872 +0,1,0,1,2244,27,17,15691,0,0,5593.089558 +1,0,0,1,5038,46,32,17721,0,0,6922.064446 +0,0,0,0,68096,13,11,244989,1,1,28475.95454 +1,1,0,1,4849,50,59,14330,0,1,11182.43115 +1,0,0,1,46944,60,72,177787,0,1,22913.23627 +0,0,0,0,24935,80,103,81077,0,1,10502.37863 +0,0,1,1,13657,21,16,44931,1,0,9730.116239 +1,0,1,1,2161,30,37,13652,1,0,12702.69822 +1,0,0,0,12526,89,112,72595,0,1,15586.68329 +1,0,1,0,68732,237,183,224090,1,1,42906.04591 +0,1,0,1,23802,89,70,126710,1,0,14178.61795 +0,0,1,0,117322,41,38,315544,1,1,38604.23011 +0,1,0,1,10745,37,35,28092,1,0,10258.59208 +1,0,1,1,4068,18,13,38388,1,0,13535.50829 +0,0,1,1,3937,61,71,33293,1,0,9309.372487 +0,0,1,1,27937,77,56,155237,0,1,15265.72861 +0,0,1,1,118244,70,71,315199,1,0,24561.74379 +0,0,1,0,22615,14,16,212549,1,0,16571.64054 +0,0,0,1,21515,48,53,75989,0,1,10118.85221 +0,0,0,1,5980,15,14,19643,0,1,2838.325206 +0,0,1,1,3639,56,52,14509,0,0,4098.491438 +0,1,0,0,88237,118,150,229779,1,0,25590.53569 +0,0,1,0,17314,19,18,106862,1,1,18383.28659 +0,1,0,1,8387,175,197,45477,1,0,20337.33835 +0,0,0,1,27267,96,76,104636,0,0,6272.606914 +0,0,1,0,20605,33,33,72497,0,1,5924.766859 +0,0,1,1,12890,29,35,101666,0,0,6657.532861 +0,0,1,1,57910,20,24,268205,1,0,21539.93028 +0,0,0,0,5402,125,149,25775,0,0,7405.369309 +0,0,1,1,69176,48,61,199564,1,1,29080.77548 +0,0,1,1,26470,36,32,94398,1,1,18038.439 +1,0,1,1,10428,18,17,34956,1,1,12826.48382 +0,0,0,0,8917,21,24,24997,0,0,740.2041154 +0,0,1,1,91861,38,23,232176,1,1,30069.56041 +1,0,1,1,5560,14,17,47658,1,1,14984.83452 +0,0,1,1,20889,36,22,64873,0,1,9138.845406 +0,0,0,1,62209,79,97,180677,0,1,21423.36985 +0,0,0,1,40967,64,74,232979,1,1,33710.50043 +0,0,1,1,17317,94,88,77496,0,0,5267.233703 +0,0,0,0,32250,21,22,188642,1,1,25062.01954 +0,1,0,1,5259,27,28,26085,1,0,9904.940439 +1,0,1,1,17184,95,65,120086,1,1,24304.85393 +0,0,1,1,21320,63,57,82308,0,0,6406.755284 +0,0,0,1,2905,41,47,26025,0,1,5116.346422 +0,0,1,1,26657,64,58,161527,0,0,9800.742942 +0,1,0,1,4178,135,107,19275,0,1,9646.977542 +0,0,1,0,1854,33,23,14154,0,0,2925.00813 +0,0,0,1,4586,38,45,21698,1,1,8043.890063 +0,0,1,1,33402,17,13,89124,0,1,10193.08059 +0,0,0,1,15479,90,84,56697,1,0,14903.83914 +0,0,0,0,17982,73,51,52065,1,0,9625.035979 +0,1,0,0,49673,23,28,239833,1,0,19319.62226 +0,0,1,1,18700,30,34,78490,0,1,8096.444424 +0,0,0,1,21059,42,52,115642,1,1,20582.73153 +0,0,0,0,19225,79,75,71811,1,0,11738.91702 +0,0,1,1,76818,13,14,246340,1,0,18364.88524 +0,1,0,1,12616,80,62,57031,0,0,8819.660959 +0,1,0,0,35916,17,21,132169,1,1,22732.21829 +0,0,1,1,33999,95,99,245949,1,1,35096.32977 +0,0,1,0,38213,42,29,338043,0,1,27467.93103 +0,0,1,1,35147,19,18,106306,0,1,10597.60992 +0,0,1,1,6744,12,9,26371,0,1,3115.42718 +0,0,1,1,16928,10,10,85080,1,1,15044.39546 +0,0,1,1,62920,27,30,216075,1,1,29230.15641 +0,0,0,0,4594,76,79,25081,0,1,6347.182107 +0,0,1,1,29370,23,27,143831,1,0,14397.39567 +0,0,1,0,21590,34,21,77756,0,1,6808.504874 +0,0,1,1,18171,33,34,128257,1,1,21128.28362 +0,0,1,1,11026,142,165,54051,0,1,12749.13256 +0,1,0,1,27952,13,16,107788,1,0,14240.70685 +1,1,0,1,21526,11,12,58430,0,1,13134.84644 +0,0,1,1,15825,11,13,59235,0,1,7629.090292 +0,1,0,0,24708,68,64,179604,0,0,9458.421442 +0,0,1,0,6608,35,34,28620,1,0,7657.136936 +0,1,0,1,36287,16,19,112196,1,1,19856.81149 +0,0,0,1,33094,66,55,100932,0,0,5185.256598 +1,0,1,1,4383,83,96,30708,1,1,17481.04576 +0,0,0,1,3176,38,22,13208,1,1,7630.788895 +1,0,1,1,7792,48,49,65744,1,1,21119.4915 +0,0,1,0,5292,15,13,21678,0,0,1678.447341 +0,0,0,1,9937,56,50,55617,0,1,6276.100271 +0,1,0,0,24133,40,50,93204,0,0,6579.419624 +0,1,0,1,26263,20,25,70868,1,1,15935.48222 +1,0,1,0,38842,16,12,126003,1,1,23245.99828 +1,0,0,1,5995,46,57,41312,0,1,11022.39004 +0,0,1,1,20999,84,64,65775,0,1,8633.419379 +0,0,1,0,7930,47,37,41473,1,0,8780.115203 +0,0,1,0,32635,19,20,107181,0,1,11260.84271 +0,1,0,1,24506,14,9,82940,0,0,6434.636636 +0,0,0,1,23291,27,19,84287,1,0,10648.04763 +0,0,1,1,8595,126,113,23124,0,1,8794.347788 +0,0,1,1,11123,51,32,41222,1,1,12292.45826 +0,0,0,1,21949,56,62,59551,1,0,11770.64662 +0,0,1,1,65556,143,119,181708,1,0,20077.4196 +0,0,0,1,3153,34,36,21218,0,0,1691.743608 +0,0,1,0,40992,43,29,120675,0,0,5962.412434 +0,0,1,1,14999,38,41,117742,0,1,14147.67247 +0,0,1,1,7797,31,26,20878,0,0,1364.936583 +0,0,0,0,4365,170,203,22037,1,0,16981.36868 +0,0,0,1,3718,24,20,21031,1,0,6464.915122 +1,0,1,0,47892,41,45,208753,0,1,22672.5671 +1,0,1,1,41496,19,16,188159,1,1,32056.56452 +0,0,1,1,11782,42,40,36314,0,1,5456.59679 +0,0,1,1,15526,73,67,46628,0,0,4740.708498 +0,0,1,1,9601,10,12,27481,0,1,2577.495205 +0,0,1,1,22798,41,25,66720,1,1,13955.58479 +0,0,1,1,5365,22,27,26066,0,1,4995.809424 +0,1,0,1,3045,131,101,21883,0,0,8470.103969 +1,0,0,1,142325,47,49,396317,1,1,51419.71823 +1,1,0,1,9976,18,20,51091,0,1,11592.21297 +0,0,1,0,15861,17,21,104855,0,0,4686.019904 +0,0,1,0,18104,36,27,164427,1,0,15162.89897 +0,0,0,1,106731,103,84,387171,1,0,28861.64052 +0,0,1,0,17323,52,46,65568,0,1,7747.679224 +1,0,0,1,15095,105,124,54340,0,1,15752.30457 +0,0,1,0,19925,19,15,156070,0,0,4339.93969 +0,1,0,1,28974,92,72,110336,1,0,17146.50535 +0,1,0,1,6453,156,187,17014,1,0,15899.73864 +0,0,0,0,1647,29,34,14242,0,1,2675.457403 +0,0,1,1,29419,19,13,146931,0,0,6868.147956 +0,1,0,1,5531,38,40,14398,0,1,7639.256995 +0,0,1,1,55573,31,29,165848,1,1,26572.05176 +0,0,1,1,6886,129,106,17220,0,1,6543.552993 +0,0,0,1,21029,128,93,91299,0,1,12639.67366 +0,1,0,1,2810,535,407,15503,0,0,21445.05937 +0,0,1,1,10824,25,23,42483,0,0,2736.140569 +0,0,1,1,89014,109,140,269738,0,1,29168.27255 +1,0,0,1,132855,77,59,335574,1,0,28664.11676 +0,0,1,0,6807,32,27,25433,1,1,9820.295287 +0,1,0,1,23915,38,30,158725,1,1,23856.34284 +0,1,0,1,13930,131,112,65562,1,0,16241.99888 +0,0,1,1,107773,40,42,346802,1,0,25373.13155 +0,0,0,1,36663,32,33,112622,1,1,17939.21897 +0,1,0,1,8902,104,130,43286,1,1,18184.09885 +0,0,1,0,97203,39,42,288126,1,1,36123.32499 +0,0,1,1,55789,95,60,362323,1,1,43607.018 +1,0,0,1,6370,68,63,51541,1,0,14975.50384 +0,1,0,1,34325,69,69,116687,1,0,18127.12363 +1,1,0,0,5632,34,40,30658,1,0,13068.91705 +1,0,1,1,47377,34,42,129626,1,0,18042.88613 +0,0,0,1,21329,12,12,54067,0,1,4810.106068 +0,0,0,1,43377,106,67,166282,1,0,16043.1834 +0,0,1,0,6150,75,54,47342,0,0,4809.24011 +1,0,1,0,6274,25,31,48622,0,1,9585.262451 +0,1,0,1,12858,18,11,119528,1,1,20842.74918 +0,0,0,1,27159,17,15,134303,0,0,4998.686659 +0,0,0,1,61431,16,14,241590,1,1,30281.73375 +1,0,0,1,12073,32,40,67400,0,0,9195.50682 +0,0,1,0,21109,25,26,147722,0,1,12990.02835 +0,0,1,1,5826,231,270,22805,1,1,18310.04886 +0,0,0,1,4372,43,46,23236,1,0,8068.479679 +0,0,1,0,4162,62,66,25960,0,0,2139.823082 +0,0,0,0,16024,13,8,52378,1,1,10904.29261 +0,1,0,0,4397,55,36,34740,0,1,5827.637873 +0,1,0,1,55877,40,35,246428,1,0,20483.00682 +0,1,0,0,74493,27,21,299245,1,1,37353.16765 +0,0,1,1,20592,20,15,144860,1,0,13779.91158 +0,0,1,1,17654,37,27,161052,0,1,14308.17281 +0,1,0,1,18912,49,53,89464,1,0,15859.34335 +0,0,1,0,28673,113,82,242562,1,1,32562.24665 +0,1,0,0,51176,62,52,140044,1,0,16471.73165 +1,1,0,1,59946,74,74,244004,1,1,40497.96124 +0,0,1,1,13579,180,211,38702,1,0,16523.06401 +1,0,1,1,28764,71,66,88237,0,0,10473.37958 +0,0,1,1,7113,51,45,24939,0,1,5621.495511 +0,0,0,0,44746,82,86,143547,0,1,16116.64649 +1,0,1,1,13932,20,17,80535,1,1,18253.09563 +0,1,0,1,13314,79,82,82226,0,1,12864.58506 +0,0,1,1,2698,10,11,22466,1,0,9765.044709 +0,0,1,1,6696,18,15,16756,0,0,1610.197453 +0,0,1,1,18693,77,91,85722,0,0,8291.163787 +0,1,0,0,79104,150,98,277467,1,0,26538.01097 +0,0,1,1,3610,171,207,28817,0,1,12939.36505 +0,0,1,0,56810,78,69,250355,1,0,20643.79915 +0,0,0,1,13651,35,43,92141,0,1,10595.26695 +0,0,0,0,181671,14,13,456529,1,1,51924.83565 +0,0,1,1,67999,16,16,251152,1,1,31213.60214 +0,1,0,0,25672,16,14,93188,1,0,12642.9538 +0,0,1,1,36456,40,48,117070,0,0,6927.236082 +0,0,1,0,14293,26,31,57400,0,1,5814.425667 +0,0,1,1,44035,12,8,180498,0,0,7340.06039 +0,0,1,1,18056,44,36,70973,1,0,12523.29614 +1,0,1,0,7446,57,54,62155,0,0,8267.670198 +0,0,0,1,60483,67,65,311253,1,1,40723.91115 +0,0,1,1,7166,39,47,20130,1,1,9371.622527 +0,0,1,1,41540,35,36,119009,0,1,14220.32201 +0,0,1,1,5214,27,31,20180,0,0,2267.005255 +0,1,0,1,28456,111,77,84648,1,0,14936.89896 +0,0,0,1,63025,244,184,244034,0,1,29132.85827 +0,0,1,1,26422,42,50,78664,0,1,9850.210724 +0,0,1,1,7722,106,69,31439,1,1,13160.67895 +0,0,1,1,6694,47,33,17661,0,1,5902.711888 +0,0,1,0,16144,49,63,130026,1,0,14066.81751 +0,0,1,0,47156,125,108,203387,1,1,30241.19053 +0,0,1,1,8336,25,22,26640,0,0,3632.90825 +0,0,0,1,17564,121,75,136194,0,0,8156.040297 +0,1,0,1,15949,33,23,105290,0,0,7177.14562 +1,0,1,1,10062,35,33,52179,1,0,14583.99348 +0,1,0,1,9108,92,77,34421,0,0,8403.453102 +0,0,1,1,26039,66,71,81208,0,1,11195.85353 +0,0,1,0,9249,43,51,46714,1,0,11168.71204 +1,0,0,1,17134,83,59,80182,1,1,20948.36407 +0,0,1,1,180249,53,45,523072,1,1,59799.16955 +0,0,1,1,34393,11,10,127796,0,0,5156.378765 +0,0,0,1,34886,36,40,267997,0,1,23392.26768 +0,0,0,1,29790,64,76,102901,1,0,14147.29512 +1,0,1,1,10440,10,11,30476,0,0,6164.337526 +0,1,0,0,5732,101,111,19706,0,1,9219.297591 +0,0,0,0,13071,103,113,41131,0,0,5684.753484 +0,0,0,1,10572,16,15,53019,1,0,9111.931717 +0,0,1,0,8155,63,56,35718,1,0,10277.30621 +0,0,0,0,37613,79,65,111708,1,0,13816.50717 +0,0,0,1,21993,144,98,71301,1,1,17817.12326 +1,0,1,1,2708,92,92,12849,1,0,16609.08389 +0,1,0,1,26275,34,30,67361,0,0,5067.939335 +0,1,0,1,1805,40,49,10101,0,0,4233.988521 +1,0,0,0,4465,23,25,39431,1,0,12541.37395 +0,0,1,1,19104,185,148,63023,0,0,8468.373517 +0,0,1,1,15114,14,16,118715,0,1,10159.10407 +0,0,1,1,5579,131,85,40869,1,0,11627.39836 +1,0,0,0,7080,25,32,35339,1,0,12927.24394 +0,0,0,1,24857,45,38,124626,0,0,7304.186619 +0,1,0,1,62020,13,15,186338,1,0,19648.6002 +1,0,1,1,36893,15,16,285473,1,1,41386.39792 +0,0,1,1,39103,82,97,139197,1,1,25824.9784 +0,0,0,1,45474,58,73,150931,1,0,16145.67606 +0,0,1,1,2976,66,80,10166,0,1,5360.752999 +0,0,1,0,1556,10,11,11288,0,1,3080.489274 +0,0,0,1,11294,30,33,112288,1,0,11639.04193 +0,0,1,1,88673,59,68,354861,0,1,32723.36488 +0,0,1,0,24787,30,20,116506,0,1,11147.73518 +1,0,0,1,31224,160,200,144355,1,1,32341.84295 +0,0,1,1,70994,84,73,278699,1,1,38448.55502 +0,0,1,1,17925,11,7,73279,0,1,7830.856999 +0,0,0,1,11086,47,39,39742,1,0,7762.919592 +0,1,0,1,22586,37,47,67962,0,0,6274.782362 +0,0,1,1,9337,49,36,24599,0,1,6195.65709 +0,0,1,1,4106,14,15,13447,0,1,3287.837466 +0,0,0,1,2820,11,13,11436,0,1,3141.438092 +0,1,0,0,3473,28,19,15797,0,0,2367.974069 +0,0,1,1,62198,18,12,162238,0,0,8267.460899 +0,1,0,1,12529,136,129,100303,1,1,23611.31726 +0,0,1,1,17934,213,223,48109,0,0,13133.45761 +1,0,0,0,26517,27,33,68723,0,0,9594.085757 +1,0,1,1,84183,47,55,211681,1,0,21632.58762 +1,0,0,1,6616,23,14,30771,0,0,4529.899662 +0,0,0,0,19834,57,43,71494,1,1,15139.6409 +0,1,0,0,2920,16,13,10914,1,0,7605.428215 +0,1,0,1,71548,95,66,246262,1,1,34952.13563 +0,0,1,1,7136,30,29,29883,0,1,5157.582288 +0,0,0,1,30996,28,25,77894,0,1,9597.463865 +1,0,1,1,4381,16,15,23762,0,1,7237.045021 +0,0,1,0,48086,13,11,134061,1,1,20640.76421 +1,0,1,1,13047,43,44,50236,0,0,9437.606288 +1,0,0,0,20779,30,29,58572,0,0,6832.944493 +0,0,0,1,3701,100,82,14786,0,0,5994.344433 +1,0,1,0,105657,23,16,365657,1,0,27892.79878 +0,0,1,1,51015,186,172,130256,1,1,26391.25123 +0,1,0,0,9247,92,114,75889,0,1,13929.84848 +0,1,0,1,16894,58,46,98446,0,0,7199.232341 +0,0,0,1,77983,106,105,240548,1,1,34452.27461 +0,0,1,1,43922,34,24,317542,1,1,38424.33029 +0,0,0,1,101715,58,52,441327,1,1,52058.55769 +0,1,0,1,65186,66,54,181351,1,0,21550.5332 +0,0,1,1,34484,12,13,133730,1,1,22621.54606 +1,0,1,0,2263,10,6,22627,0,0,5749.223418 +0,0,1,1,36360,81,78,162718,1,0,17983.01013 +0,0,0,1,42258,83,59,226327,1,1,32179.11067 +0,0,1,1,23536,36,28,107698,0,0,4986.290485 +0,0,0,0,4772,93,74,31472,1,1,12786.3002 +0,0,0,1,8391,107,77,21257,1,0,11037.04497 +0,0,1,1,19917,69,48,111687,0,0,5879.647635 +0,0,1,1,53873,39,37,164836,1,0,17266.07837 +0,0,1,0,2276,10,10,11649,1,0,4619.491361 +1,0,1,0,7709,53,61,37189,0,1,9916.962461 +0,0,1,1,2692,14,16,14203,0,1,2161.660276 +0,0,1,1,90742,97,107,579829,1,1,70405.60593 +0,1,0,1,14130,45,51,59812,0,1,10011.99584 +0,1,0,0,119865,52,31,450829,1,1,54908.52665 +1,0,1,1,24101,35,42,77767,0,1,13610.06466 +0,0,1,1,6533,65,56,32288,0,1,5449.408473 +0,0,0,0,103019,32,37,336370,1,1,39979.63889 +0,0,1,1,84538,50,50,261351,0,1,23226.09424 +0,0,0,0,20640,14,10,55757,0,1,4643.135404 +1,0,1,1,11230,14,11,104708,0,0,8611.12893 +0,0,0,0,16160,10,9,89259,0,0,3531.240563 +0,0,0,1,19215,18,13,148376,1,1,21817.54653 +0,0,1,1,4843,25,26,31145,0,1,5595.403592 +0,0,1,1,20410,29,27,133682,1,1,21018.41974 +0,0,1,1,49747,11,10,150840,0,0,7639.384949 +0,0,1,1,14819,15,12,47163,1,0,9710.63789 +0,0,1,1,21792,211,136,131719,0,1,16740.27343 +0,0,0,0,35220,27,21,184873,0,0,6516.061032 +0,0,1,1,6192,41,43,17203,1,1,10457.70207 +0,0,1,0,27843,184,160,77525,0,0,10263.97187 +0,0,0,1,19419,103,100,65629,0,0,8372.933098 +0,1,0,0,4127,10,8,12863,0,1,5745.206084 +0,0,1,1,47825,23,15,174200,1,0,16693.90473 +1,0,1,0,7680,111,89,61748,1,0,16859.57223 +0,0,1,1,95504,82,60,303089,1,1,39591.00952 +0,0,0,0,17615,16,9,102904,1,1,15570.84483 +0,1,0,1,12249,97,107,39792,0,0,10507.14384 +0,1,0,1,14642,57,63,74178,0,0,9045.136014 +1,0,0,1,9785,14,9,26468,0,0,4921.146726 +0,0,0,0,5737,16,14,46649,0,1,3290.437078 +0,0,1,1,23504,67,40,83551,1,0,11067.72266 +0,0,1,1,20753,46,28,53095,0,0,4100.066835 +0,0,1,1,8840,29,36,72334,0,0,4716.348468 +0,0,1,1,39863,58,67,210389,1,1,31346.13915 +0,0,0,0,9049,68,63,35945,0,0,2342.322272 +0,1,0,1,17463,66,47,140023,1,0,16650.33293 +1,0,1,0,8275,19,16,21935,0,0,6043.553795 +0,0,1,0,37609,76,93,186672,1,0,18216.06353 +0,0,1,1,10049,100,68,53619,1,1,15951.4408 +0,0,1,1,10942,125,154,33710,1,0,15178.07679 +0,0,0,0,72978,46,39,189662,1,1,25200.70828 +0,0,1,0,12363,27,22,31227,1,1,9139.886332 +1,0,1,0,27289,36,31,121945,1,0,16779.91344 +0,0,1,1,7667,44,47,46684,0,1,6629.553622 +0,0,1,0,8679,18,11,28879,0,0,2725.168602 +1,0,1,1,22840,11,6,126504,0,0,9502.755081 +0,1,0,1,9708,89,59,45745,0,0,5006.419232 +1,0,0,0,26739,81,87,120447,0,0,10819.49251 +0,0,0,1,56503,53,48,146965,1,1,21664.35844 +0,0,1,1,1995,112,104,15256,0,1,6100.597739 +0,0,0,1,35938,19,20,95950,0,0,4656.715179 +1,0,0,1,4312,38,41,30816,0,0,8202.154684 +1,0,1,1,11240,13,14,61523,1,0,14287.6097 +1,0,1,1,12051,95,114,32497,1,0,17091.29017 +0,1,0,1,14072,25,16,134313,1,0,13281.92333 +1,1,0,0,10772,30,20,43592,0,0,8978.875867 +0,0,1,0,29398,71,72,103500,0,1,12218.00156 +0,0,1,1,2009,82,83,13299,0,1,6398.359921 +0,0,1,0,55019,75,65,151204,0,1,15393.43583 +0,0,1,1,8024,117,117,24001,1,0,12216.88022 +0,0,1,1,3163,16,10,29186,0,0,3243.766223 +0,0,0,0,9121,17,18,84477,0,0,3859.295896 +0,0,1,1,4566,56,40,13178,0,0,4549.874946 +0,0,1,1,49437,46,44,228672,1,0,18526.7485 +0,0,1,1,52486,15,10,140954,0,0,5858.523325 +0,0,1,1,69370,114,103,184808,0,1,20807.66651 +0,0,0,1,9789,156,197,28010,0,1,13530.65248 +0,0,1,1,8656,80,95,43736,1,0,12710.58625 +0,0,1,1,21175,48,48,102315,1,0,12091.04998 +0,1,0,1,7702,292,312,26266,1,0,22859.58487 +1,0,1,0,60376,124,104,157177,1,1,31904.76107 +0,0,0,0,58435,60,41,191337,0,1,17286.22888 +1,0,1,1,3496,125,161,16201,0,0,12469.58864 +0,0,1,1,7388,40,48,20111,0,1,5416.301687 +0,0,1,0,10081,104,76,25349,1,0,7577.97488 +0,1,0,1,59263,45,33,159316,1,1,26015.1458 +0,0,1,1,39850,117,126,266732,1,1,38457.41874 +0,0,1,1,124422,30,24,426440,0,1,37135.08325 +1,0,1,1,23914,30,35,130900,0,1,16918.90171 +1,0,1,1,7699,27,25,33783,1,0,14934.28956 +1,0,1,1,78624,80,97,213882,1,1,36116.57596 +0,0,1,0,17187,53,60,96134,0,1,10323.35981 +0,1,0,1,54043,18,22,213215,1,1,31239.19889 +0,0,0,0,26325,47,33,152157,1,1,21319.23994 +1,0,1,1,42241,21,25,253465,1,1,36720.15919 +0,0,0,1,5858,15,10,17547,0,0,866.5027041 +1,0,0,1,6384,102,72,35227,1,0,14702.80333 +0,0,1,1,2786,21,17,12121,0,1,2258.131151 +0,0,0,1,82116,31,37,456050,1,1,51880.29928 +0,0,1,1,24232,142,88,157885,1,0,18343.63045 +1,0,1,1,2620,45,30,26001,0,0,5512.498855 +0,0,1,1,3921,31,24,14241,0,0,3313.33739 +1,0,0,1,118363,143,147,304508,1,1,47852.08547 +0,1,0,1,21732,11,12,66918,0,1,8795.355976 +0,0,1,0,8770,50,42,82729,1,1,15661.76372 +1,0,1,1,3432,28,20,17932,1,0,11402.19819 +1,0,0,1,6756,45,34,34062,0,0,8351.665632 +0,0,1,1,30137,74,73,117713,0,0,7571.36749 +0,0,0,0,48923,25,27,131711,0,1,12597.96129 +0,0,0,1,10364,36,33,35162,0,0,2499.167204 +0,0,1,0,15403,11,8,47092,1,1,11240.48073 +0,0,1,1,5730,213,267,47487,0,0,12059.02577 +0,0,0,0,27791,50,47,117492,0,0,5134.198316 +1,0,1,1,11571,23,21,53433,0,1,10629.10275 +0,0,1,1,6523,43,36,18816,1,0,9464.520369 +0,1,0,0,37292,42,42,96920,0,0,7979.394504 +0,0,0,1,91266,90,57,287031,1,1,36334.68281 +0,0,1,1,25181,241,297,70897,1,1,27973.15607 +0,0,1,1,15019,50,55,38278,0,1,6338.094915 +0,1,0,1,5511,12,14,48863,1,1,11006.59102 +1,0,1,1,29488,58,47,125085,1,1,26038.79802 +0,0,1,1,32206,38,36,284796,1,1,36645.63815 +0,0,0,1,124635,35,41,368727,1,1,45373.57403 +0,0,1,0,12544,20,20,91212,0,1,8654.983774 +0,1,0,1,12907,12,13,71272,0,0,5779.293269 +0,1,0,1,20808,156,121,58423,0,1,11890.29075 +0,0,0,0,15182,96,94,91407,1,1,19539.92132 +0,1,0,1,4417,18,20,15004,0,0,3937.913319 +0,0,1,1,27261,39,35,89449,0,0,5573.486266 +0,0,1,0,94970,138,145,270429,0,1,28266.38094 +0,0,1,1,4418,21,16,13274,0,0,3095.427966 +1,0,0,1,8841,37,22,33257,0,0,5563.593837 +0,0,0,0,6372,51,36,25215,1,0,8733.23786 +1,1,0,0,2144,43,38,14802,0,0,8901.799611 +0,0,0,1,4515,23,23,38466,1,1,12135.37366 +0,1,0,0,2454,39,33,16107,0,1,3831.361007 +1,0,0,0,14352,102,88,45394,0,1,11339.28766 +0,1,0,1,19537,106,102,87840,0,1,14775.65201 +0,0,1,0,5145,45,50,25495,0,1,4170.850008 +1,1,0,1,19614,69,56,57081,1,0,17970.53412 +0,0,0,1,8633,49,62,74423,0,0,7596.227578 +0,0,1,1,11846,78,90,54722,1,1,15920.2787 +0,0,1,1,10003,78,67,26486,0,1,4266.673288 +0,1,0,1,9625,29,27,31766,0,1,6432.659745 +0,0,1,1,16926,13,12,99886,1,0,12412.85596 +0,0,1,1,10846,12,13,62702,1,0,8862.558253 +0,1,0,1,30862,11,10,141853,1,1,22100.80708 +0,0,0,1,37431,17,14,100785,1,1,16670.66491 +0,1,0,1,23443,63,61,150607,0,0,10265.69667 +0,1,0,1,5254,45,39,16314,0,0,3674.955232 +0,0,0,1,1484,74,83,10356,0,1,3749.068364 +0,1,0,0,32211,61,61,169634,1,0,17965.99838 +1,0,0,1,24262,18,22,114054,1,0,14901.1801 +0,0,1,1,23411,52,34,61017,1,1,13705.96307 +0,0,1,1,14557,31,29,48400,1,0,10045.32635 +0,0,1,1,3612,31,39,15993,0,1,4451.110194 +0,0,1,0,6048,74,91,26755,0,0,4823.024566 +0,0,0,0,41713,38,31,127759,0,0,3579.628507 +0,0,0,1,3755,110,81,31828,0,0,5080.331732 +0,0,0,1,58730,11,13,156161,1,0,16838.31593 +1,0,1,1,9946,23,18,25537,0,1,8450.373565 +0,0,1,1,62638,29,34,201956,1,0,16353.64152 +0,0,1,0,24886,35,31,140824,1,1,21353.98798 +0,0,0,0,59675,79,49,172544,1,1,23784.014 +1,0,1,1,22479,149,152,57507,0,0,14855.23627 +0,0,1,1,6765,64,55,36235,0,0,3924.85708 +0,0,1,1,4370,163,113,17863,1,0,13067.51613 +1,1,0,0,32908,63,69,182030,1,1,31603.40384 +1,0,0,0,44768,72,78,288131,1,1,41805.45738 +0,0,1,0,51501,19,14,160081,0,1,13253.74293 +0,1,0,1,37973,12,9,118758,0,0,9247.633942 +0,0,0,1,41362,21,26,126915,0,1,12458.48419 +0,0,1,1,100158,397,323,274719,0,1,37514.29356 +0,0,1,1,13531,68,68,41268,0,0,7088.813884 +0,1,0,1,10693,39,38,52841,1,1,15146.54356 +0,0,1,0,9508,55,39,66334,0,1,5978.721145 +0,0,0,1,6476,64,80,23459,0,1,7550.86325 +0,0,1,0,11469,29,35,54717,0,0,3341.823071 +0,0,1,1,50873,17,12,193027,1,1,26327.42469 +0,0,0,1,19255,198,216,48928,0,0,9853.256899 +0,0,0,0,68221,157,133,184800,1,0,18208.87536 +0,0,0,0,6760,42,49,52894,1,1,10408.71595 +1,0,1,1,10816,85,53,27988,0,1,9306.015641 +0,0,1,1,13410,17,19,60339,0,0,2884.160011 +0,1,0,1,20812,47,43,65449,0,1,9734.050914 +0,0,1,0,10132,12,14,75928,0,0,2883.081314 +0,0,0,0,59470,41,27,292631,1,1,36257.64003 +0,0,1,1,112782,38,39,315258,0,1,27193.64001 +0,0,1,1,20229,84,53,58026,0,0,5205.690894 +0,1,0,1,17655,142,98,96204,1,0,15345.31194 +0,0,0,0,18655,23,17,48163,1,1,10107.28003 +0,0,1,1,55725,25,23,166714,1,0,17594.85795 +0,0,0,1,8346,30,32,30487,0,0,3175.214083 +1,0,0,1,24125,92,77,102457,0,0,11334.54085 +0,0,1,0,6244,90,75,51526,0,0,4258.853967 +1,0,1,0,13683,44,56,67320,0,0,8328.713237 +1,0,0,1,6983,18,12,55021,1,1,14420.92728 +1,0,0,1,24811,15,18,102213,0,1,14348.56177 +0,0,1,0,44913,63,44,417163,1,1,47465.3412 +0,0,1,0,4186,12,10,15928,1,1,8648.999344 +0,0,1,1,3885,96,86,17580,1,1,10898.43799 +1,0,1,0,10182,18,20,74347,1,0,12159.90168 +0,0,0,1,11623,16,19,87554,0,1,7233.63393 +0,0,0,1,37849,24,19,108781,1,1,17732.64553 +0,0,1,0,26267,19,24,111482,0,0,5042.940727 +0,0,0,0,36834,84,99,100638,1,0,15057.14152 +0,0,1,1,11158,34,22,32180,0,0,4486.460688 +0,0,1,1,41565,13,9,108630,1,1,17150.95795 +0,0,1,1,8298,73,93,69506,1,0,11789.74401 +0,0,1,1,10833,115,107,33468,0,0,5959.569767 +0,0,0,0,13728,28,26,52748,0,0,2417.57566 +0,0,1,0,43005,16,15,243471,0,1,19579.10068 +0,0,0,1,23209,51,59,64592,0,0,6723.133299 +1,0,1,1,24414,54,36,89895,1,0,16628.32527 +0,0,0,1,12277,22,26,34947,0,0,4656.094392 +1,0,1,0,38269,12,13,107397,1,0,15324.51276 +0,0,1,1,20574,39,40,78478,1,0,9484.149855 +0,1,0,1,8841,10,9,34104,0,0,5098.284531 +0,0,1,0,9260,41,34,87200,0,0,3820.739798 +0,0,1,1,21805,59,54,106506,0,1,11924.61423 +0,1,0,0,27969,36,24,155298,0,1,15583.72657 +0,0,0,1,16583,92,65,58958,1,0,12828.23697 +1,0,1,1,72348,51,35,241828,0,1,26127.9543 +1,0,0,1,14403,27,20,55228,0,0,6713.853191 +0,0,0,1,22024,49,46,118816,1,0,14502.08795 +1,1,0,1,62831,102,115,180594,1,1,35018.90603 +0,1,0,0,82642,25,22,271759,1,0,20509.34615 +0,0,1,1,37274,35,36,140684,0,0,7800.921691 +0,1,0,0,52187,12,9,147369,1,1,20957.4076 +1,0,0,1,2582,97,120,14821,1,0,15673.91672 +0,0,1,1,11106,102,80,49998,0,0,5621.031559 +0,0,0,1,11587,125,114,69017,0,0,9156.330752 +0,1,0,1,41123,62,71,225071,1,0,21751.17177 +0,0,1,1,4741,100,74,39316,1,0,11367.81931 +0,0,1,1,2983,24,18,12027,1,0,7367.999154 +0,0,1,1,67185,19,23,592031,1,1,66294.40273 +0,0,1,1,13886,111,107,58209,1,1,15133.085 +0,0,1,1,4282,70,47,10741,0,1,5618.644444 +0,0,1,1,14278,49,47,50781,1,1,13386.88583 +0,1,0,0,33723,16,20,230303,1,1,31614.4348 +1,0,0,1,108490,49,36,285019,1,1,40717.37495 +1,1,0,1,12425,101,107,31355,1,0,18790.94691 +1,0,1,1,37843,13,10,159448,0,0,10230.59352 +0,0,1,1,4886,22,23,18308,0,0,3440.503781 +0,0,0,1,24118,22,23,85655,0,1,9020.869609 +0,0,0,1,41197,71,85,106057,1,1,19384.67619 +0,0,0,1,36730,86,57,118542,0,1,12495.46075 +1,0,1,1,6270,39,38,35737,0,0,8370.509775 +0,0,0,1,48090,39,41,194165,1,0,17499.98951 +0,0,1,1,4196,46,37,28286,1,1,9859.146847 +0,1,0,0,7210,35,44,49912,0,0,5967.841137 +0,0,1,1,4610,18,22,24152,1,0,7881.626501 +0,0,1,1,82500,149,125,207800,1,0,21643.22548 +0,0,1,1,3152,68,82,15647,0,0,5102.757612 +0,0,1,0,3001,113,104,12047,0,1,4998.714087 +0,0,0,1,36290,11,13,189870,0,1,18641.16406 +0,0,1,1,13968,257,311,66876,0,1,19960.93409 +0,0,0,1,7859,46,43,39429,0,1,7089.908942 +0,1,0,1,51940,178,174,292856,0,1,31379.32219 +1,0,1,1,40464,26,24,185851,1,0,19645.23798 +1,0,0,1,5855,22,15,41749,1,1,13650.70235 +1,0,0,0,20478,70,55,73767,0,0,7603.842548 +0,0,0,0,12950,20,14,63673,1,1,11390.60223 +0,0,1,1,50706,185,191,259323,0,1,29831.33311 +0,0,1,0,12336,66,47,65794,0,0,5295.01662 +0,0,1,1,23842,37,33,92404,1,0,12156.37941 +0,0,0,0,5618,19,22,14347,1,1,5903.90688 +0,0,0,1,28132,93,117,74672,0,0,6671.775593 +1,0,1,1,12270,17,22,31254,0,1,8959.680159 +0,0,0,1,35356,124,107,100331,0,1,12764.66444 +0,1,0,1,39636,23,23,139884,0,1,13465.12537 +0,1,0,1,14221,143,151,51248,1,0,19010.06087 +1,0,0,0,4411,17,17,14701,1,0,10443.93574 +0,0,1,1,35540,19,23,111631,1,0,14659.38464 +0,0,0,1,2436,78,53,11520,0,0,3432.258329 +0,0,1,1,10278,119,122,64211,0,0,9013.157714 +0,0,1,1,23473,30,28,60043,1,1,13186.55818 +0,1,0,1,32605,62,53,130204,1,0,17086.88052 +1,0,1,0,5443,51,53,29779,1,1,12891.60928 +0,1,0,1,12287,14,11,32485,1,0,10804.32922 +1,1,0,1,16551,14,13,72707,0,0,10831.10914 +0,0,0,0,112485,37,33,294377,1,0,20969.14879 +1,1,0,0,43867,15,15,178062,1,1,29001.08064 +0,0,1,0,74769,18,20,211482,1,1,27140.0278 +1,0,1,1,41708,229,147,125470,0,1,21229.99095 +0,1,0,1,14132,32,23,73589,0,1,9436.539748 +0,0,1,1,1645,42,29,10561,1,0,7917.72614 +0,0,1,0,12549,76,57,38509,0,0,4186.247555 +0,0,1,1,41952,132,123,159724,1,0,19038.30905 +0,1,0,1,57496,28,30,180825,1,0,18473.04575 +0,0,0,0,57708,43,55,188921,0,0,8200.044385 +0,0,1,1,26207,63,72,105094,1,1,19052.52686 +0,0,1,0,87395,14,11,265685,1,1,31667.2133 +0,0,0,0,31288,62,66,128180,0,0,7876.537537 +0,0,1,1,22486,28,32,186502,0,1,17617.60425 +0,1,0,1,62431,23,29,170763,1,1,26957.69292 +0,0,1,1,16166,22,28,47458,0,0,4414.542295 +1,0,0,0,10481,66,84,34409,0,1,9644.67258 +0,0,1,1,11046,68,54,39023,0,0,4836.424456 +1,0,1,0,5716,33,27,18758,1,0,10680.46847 +0,0,1,1,31073,177,135,120287,0,0,10224.81206 +0,0,0,1,18683,128,151,93074,1,0,18355.78078 +0,1,0,1,36027,50,32,153666,0,1,17033.61914 +1,0,0,1,57356,59,69,270515,1,1,38720.5649 +1,0,1,1,17026,72,65,68146,0,1,12469.49981 +0,0,0,1,10509,25,27,35371,0,0,3850.495537 +0,0,1,1,71571,18,20,243258,1,1,32101.21364 +0,0,1,1,18663,33,38,111100,0,0,7038.626683 +0,0,0,1,3083,69,54,16495,1,0,9018.927033 +0,0,0,0,8970,14,11,84745,1,1,13071.80872 +0,0,0,1,31243,68,54,136917,1,1,23749.7002 +0,0,1,1,17250,29,24,60243,1,1,13029.4313 +0,0,1,1,4554,18,17,13039,0,0,4193.71714 +0,0,0,0,13831,69,76,90289,0,0,4992.113273 +0,1,0,0,44217,42,51,130825,1,0,16737.69932 +0,0,1,0,10513,29,34,27477,1,1,10363.02502 +0,0,0,0,15967,39,35,89785,1,0,10216.49236 +0,0,1,0,3019,13,8,15849,0,1,1918.472293 +0,0,1,1,7839,127,112,66145,1,1,17110.84006 +1,0,0,1,46790,28,29,202190,0,1,23053.75214 +0,0,0,1,25059,84,95,101515,0,0,8605.208523 +0,0,0,1,60204,67,83,290129,1,1,38259.78067 +0,0,0,0,5553,37,31,24958,0,1,3215.696532 +1,1,0,0,2635,176,170,10540,1,0,18151.19212 +0,1,0,1,11919,73,94,116933,0,1,16588.51885 +0,0,1,1,44548,78,49,139936,1,0,16430.5856 +0,1,0,1,12196,72,58,54559,0,1,10381.80894 +0,0,1,0,30518,58,48,85608,1,1,14940.5702 +1,1,0,1,11235,15,16,33761,0,1,8618.757569 +0,0,1,0,16918,24,20,49967,0,0,2205.347793 +0,0,1,1,51551,19,23,228340,0,1,19487.9771 +0,0,0,1,1161,137,126,10110,0,0,6438.150315 +0,0,1,1,23339,10,6,121433,1,0,11475.91936 +0,0,1,1,11747,13,14,30759,0,0,3952.418228 +0,0,0,1,38292,17,19,276460,1,1,32961.26813 +0,0,0,0,9385,14,12,24120,0,1,2150.764164 +1,1,0,1,22095,90,88,186400,0,1,24938.68313 +0,0,1,0,4524,52,57,11477,1,1,8481.162147 +0,0,1,0,3590,84,65,17743,0,0,3901.467895 +0,0,0,0,57173,69,81,194397,1,0,18470.39534 +0,0,1,1,11691,12,10,30587,0,0,2445.971539 +1,0,0,1,43341,132,82,119059,0,0,12205.08601 +1,0,1,1,39893,27,23,118939,0,1,15217.28724 +0,0,0,1,43130,25,25,155603,1,1,22927.8039 +0,0,1,1,2432,20,22,10707,0,1,3579.826911 +0,0,0,1,4616,99,94,14943,1,0,11813.63627 +0,0,0,1,43744,20,19,123561,0,1,12900.7674 +0,0,1,0,19166,84,96,164619,1,1,24497.47256 +0,0,0,1,12244,106,114,47774,0,1,9585.584733 +0,0,1,1,48686,60,76,335624,1,1,41627.81473 +0,0,0,1,14376,59,75,39382,0,1,6177.235708 +0,0,0,1,23822,21,17,70596,0,1,7828.339295 +1,1,0,1,5162,132,119,51026,1,1,22679.74804 +1,0,1,1,97206,70,85,288765,1,0,28145.2671 +0,0,0,1,7972,56,71,20187,0,0,4358.590607 +0,0,1,0,119282,111,73,381223,1,1,48825.15791 +1,1,0,1,48188,36,24,166317,0,0,13475.01085 +1,0,1,1,21309,75,93,96090,0,0,12753.54049 +0,1,0,1,59161,10,9,177549,0,1,17060.57066 +0,0,1,0,18220,34,28,95970,0,1,8836.230462 +0,0,0,1,58011,81,84,164011,0,1,19483.50593 +0,0,1,1,14562,27,31,50186,0,0,4372.063859 +0,0,0,1,30014,276,295,205223,0,1,29953.98746 +0,0,1,0,25956,43,36,207341,1,0,17356.96793 +0,1,0,1,23191,55,62,115530,1,0,17237.24459 +0,0,0,1,63723,18,17,277788,1,1,33984.33364 +0,0,1,1,34485,23,26,94661,1,0,12052.66583 +0,0,0,1,10832,89,80,36405,1,0,11457.6029 +1,1,0,1,3425,11,7,20511,1,1,14623.94369 +0,0,1,1,48257,25,16,138281,1,1,21028.4986 +0,0,1,1,13363,39,44,44636,0,0,6198.280838 +0,0,1,1,11948,48,45,78276,0,0,8591.515014 +0,0,1,1,24625,80,81,154873,0,1,17730.1473 +0,0,0,0,20477,22,24,150152,0,1,12872.69707 +0,0,1,1,54552,52,58,178123,0,0,9448.094145 +0,0,1,1,12774,91,103,38037,0,0,6335.50074 +0,0,1,1,4302,23,18,37911,1,1,11050.47634 +0,0,1,1,25692,110,93,107027,1,0,15720.36657 +0,0,0,1,12288,91,66,85794,1,1,17085.61706 +1,0,0,0,31423,61,69,149848,1,1,26623.66442 +0,1,0,0,47336,48,50,199343,1,0,20241.82221 +0,1,0,1,56979,15,13,223319,1,1,30246.84776 +0,0,1,0,42895,11,8,409816,1,0,27934.82979 +0,0,1,1,27340,60,45,199973,1,1,28643.98619 +0,1,0,1,18050,96,73,46960,0,0,7386.644608 +0,0,1,1,17590,117,111,64199,1,1,16801.29799 +0,0,1,1,73270,51,35,210451,1,0,17615.61941 +0,0,1,0,19700,60,40,52203,0,0,3702.353016 +1,0,0,1,62574,15,17,211830,0,0,10518.47624 +0,0,1,0,14416,124,114,61017,0,1,10537.92488 +0,0,0,1,3721,61,71,18761,0,0,4274.035841 +0,1,0,0,32014,39,34,168779,1,1,22833.47745 +0,0,1,1,3085,10,11,17143,0,0,1725.572935 +0,0,1,1,7698,67,67,21786,0,1,4493.399538 +1,0,0,1,25514,16,18,79330,0,0,8614.599349 +0,0,0,0,80766,13,8,409216,0,1,32385.27363 +0,0,1,1,9472,23,15,70972,0,1,7159.863106 +0,0,1,1,14786,173,120,51263,0,1,9207.940664 +0,0,1,1,12143,16,10,84849,1,1,15010.17268 +0,0,1,1,34725,35,31,151778,0,1,15802.91663 +0,1,0,1,38681,25,17,234451,1,1,30319.0011 +1,0,0,1,89769,32,22,267633,1,1,38036.22259 +0,0,1,0,25641,13,11,69995,1,0,11522.14837 +1,1,0,1,44697,79,56,237676,1,0,25230.62232 +1,0,1,1,32141,13,7,126438,1,0,16778.87418 +1,1,0,1,22923,268,312,194055,1,1,43769.61503 +0,0,0,0,2684,77,58,17673,0,0,3511.944554 +0,0,0,1,9460,68,52,68510,0,0,5695.145297 +0,0,0,1,12360,14,14,67594,0,0,3685.758896 +1,1,0,1,53037,21,15,211044,1,0,24571.00018 +0,0,0,1,34956,153,147,142812,0,1,19931.75744 +1,0,1,1,20072,111,98,108150,0,1,19529.78425 +0,1,0,1,8014,109,120,52152,1,1,19136.5617 +0,0,1,1,18900,11,7,74524,0,0,4241.815778 +1,0,0,1,41220,50,44,275648,1,1,39463.25931 +0,0,0,1,12064,114,141,58648,0,0,7356.327876 +0,0,0,1,2348,66,61,11965,0,0,3747.519848 +0,0,1,1,50879,130,115,278253,1,0,25501.66844 +0,0,1,1,106257,60,42,273111,1,1,34219.61616 +0,0,1,1,19632,77,82,130469,1,0,15473.29918 +0,1,0,0,11690,19,15,111676,0,0,6369.082518 +0,1,0,1,32570,45,34,110579,1,0,14653.6345 +0,0,1,1,8791,34,27,22367,1,1,9900.519725 +0,1,0,0,16224,40,27,40810,1,0,10836.31695 +1,0,1,1,70877,37,38,182144,0,0,10354.31599 +0,1,0,0,43168,11,12,164172,1,0,17868.72587 +0,0,0,1,5976,33,30,22015,0,0,2909.544956 +0,1,0,0,37102,21,20,102313,0,1,13090.73802 +0,1,0,1,23036,43,49,80175,1,1,17671.33507 +1,1,0,1,23049,99,122,63075,1,1,23058.83592 +1,0,0,1,4835,60,44,21741,0,1,8089.685727 +1,0,1,1,12372,18,20,73617,0,0,7871.632576 +0,0,0,0,167714,32,19,532991,1,1,59494.49477 +0,0,0,0,23597,18,20,111714,0,1,10550.17236 +0,0,0,1,3226,66,47,29406,0,0,3692.980006 +0,0,1,0,32170,156,162,206448,0,0,13529.61913 +1,0,0,1,95361,130,121,325754,1,1,48460.80509 +0,0,0,0,64591,109,98,284399,1,0,23513.40276 +0,1,0,0,4407,72,49,20032,1,0,11115.66682 +1,1,0,1,18638,97,73,114884,1,1,26747.56274 +0,1,0,1,82656,32,27,226035,1,1,31697.45921 +0,0,1,0,9507,26,15,92186,1,0,11675.94125 +0,0,1,0,32765,53,56,91646,1,0,12127.16554 +0,1,0,1,9417,41,38,24161,1,1,13055.80096 +0,0,1,1,9982,115,131,96959,0,1,13698.22229 +0,0,1,0,38505,33,27,134791,0,0,4399.388273 +0,0,0,0,63185,43,28,213695,0,1,17689.83084 +0,1,0,0,16589,56,34,73505,1,0,12810.15442 +0,0,0,1,27012,20,24,73207,0,0,4864.281094 +0,0,1,1,9818,33,38,77656,0,0,6199.903732 +0,0,0,1,14712,31,35,40750,0,0,2730.122799 +0,0,1,1,21903,40,46,81450,1,0,13859.8355 +0,0,1,0,7400,69,60,28434,1,0,9762.730488 +1,1,0,0,11898,10,12,92882,1,0,15522.97777 +0,0,1,1,12334,115,91,84071,0,1,11690.24518 +1,0,1,1,25406,64,56,66919,0,1,14219.43138 +1,0,1,1,7389,14,11,53856,1,0,12486.7371 +1,1,0,0,86368,62,43,224519,1,1,35715.53248 +0,0,0,0,117187,20,19,324495,0,1,25494.5016 +0,1,0,0,33745,72,47,173644,1,0,18047.44692 +0,0,1,1,18768,104,108,78648,1,1,15877.81596 +0,1,0,0,2701,43,44,17103,1,0,10857.14235 +1,1,0,1,12999,20,23,98378,0,0,11892.40261 +0,1,0,1,19686,179,231,156726,1,1,33337.30767 +0,0,0,0,43841,77,49,161258,0,1,14596.43392 +0,0,0,1,69194,208,216,178354,1,0,23894.58049 +0,0,0,1,2474,33,34,24532,0,0,2858.632027 +0,1,0,1,14448,34,21,86961,0,0,5878.004417 +0,0,1,1,26173,42,25,99311,0,1,9894.79667 +0,0,1,0,2137,161,157,16364,1,0,11303.72537 +0,0,0,1,26067,135,131,155046,1,1,28146.75714 +1,0,1,1,11034,10,12,46612,0,1,8520.844793 +0,0,0,1,13843,132,135,92344,1,0,15489.22167 +0,0,1,0,13819,34,25,35881,1,1,8913.420934 +0,0,1,0,24105,23,27,65307,0,1,5669.662516 +0,0,1,0,54221,63,55,222019,1,1,28354.79066 +1,0,0,1,84700,200,190,377756,1,1,53778.42969 +0,1,0,0,2368,105,81,21188,0,1,9076.91555 +0,0,0,1,2363,27,20,10165,1,0,6417.365157 +0,1,0,1,70745,72,73,199052,1,1,30427.48377 +0,1,0,1,23133,17,17,116491,1,0,13790.5221 +0,1,0,1,8661,55,60,67170,0,0,8687.583676 +0,0,1,1,21920,55,51,82313,1,0,12226.4228 +0,0,1,1,7778,97,114,61796,1,0,14462.92373 +0,0,1,0,18925,52,41,105305,0,1,10454.9325 +0,0,1,1,4115,68,73,16836,0,0,5430.179286 +1,0,1,1,23519,71,43,126031,1,1,23700.03716 +0,0,1,1,63336,67,83,268036,1,1,37758.74083 +0,0,1,1,19677,50,34,102561,0,0,5443.590796 +1,1,0,1,16814,61,49,63214,1,1,20121.49264 +0,0,0,1,33617,21,22,135596,0,0,5190.790215 +0,0,1,1,42113,127,76,212139,1,1,27746.21581 +1,0,1,1,4631,317,250,25033,1,1,22180.12798 +1,0,1,1,86415,34,39,315105,1,1,44069.65938 +1,0,0,1,5281,86,104,33488,1,1,16338.14719 +1,0,0,0,3797,10,8,10273,0,0,5414.094946 +0,0,1,1,5677,17,15,23129,0,0,4098.004662 +0,1,0,1,7971,178,123,25390,0,0,8156.575502 +0,0,1,1,3632,69,56,24876,0,1,5957.777274 +0,0,1,0,27567,51,50,118976,1,0,15984.9995 +0,0,1,0,55529,35,35,175141,1,0,14314.20721 +1,0,1,0,27218,46,38,74856,1,0,13268.24686 +1,1,0,0,5909,127,115,21842,1,0,16346.18441 +1,0,1,1,23480,26,16,98484,1,0,16978.09629 +0,1,0,0,41426,120,110,150732,1,0,20274.3054 +1,0,1,1,3100,28,33,21758,0,0,9822.09884 +0,0,0,0,4836,50,53,19446,0,1,3234.454299 +0,0,0,0,1829,33,37,11692,0,0,374.4919543 +0,1,0,1,97609,16,16,251115,1,1,34324.18839 +0,0,1,1,57828,62,55,253448,1,1,34590.33593 +0,1,0,1,14845,42,28,109708,0,0,6290.856127 +0,0,1,0,24978,27,28,93186,0,1,8775.30422 +0,0,0,0,11430,96,72,83583,1,0,14556.84843 +0,0,1,0,8300,111,75,32700,0,1,5379.372413 +0,0,0,0,6110,29,22,48991,1,1,11711.37175 +0,1,0,0,12882,35,37,96742,0,1,11139.01919 +0,1,0,0,13966,132,136,47141,1,1,16030.57305 +1,0,1,1,47068,72,53,179251,0,1,20728.17565 +0,0,1,1,15723,11,8,61958,0,1,6943.252728 +1,0,1,1,49000,69,64,132441,0,0,11586.90069 +0,0,1,1,7533,68,73,31732,0,0,5844.905558 +0,0,1,0,36929,41,33,187615,1,0,15022.90035 +0,0,1,0,32258,129,104,101546,1,1,19915.63252 +0,0,0,0,30929,47,40,79995,0,1,7886.598072 +0,0,1,1,19383,144,183,95841,0,1,17049.23992 +0,0,1,1,44333,11,10,186906,0,0,6885.897147 +0,0,1,0,4121,24,23,26389,1,0,6905.691011 +0,0,1,0,10302,93,112,86224,1,1,17494.19729 +0,0,1,1,23366,117,111,126904,0,0,8984.496102 +1,0,0,1,5885,72,51,33055,0,0,8370.312256 +0,0,0,0,3411,61,56,17393,0,0,927.7440511 +0,0,1,1,37285,142,146,180782,1,1,29339.67577 +1,0,1,1,55002,22,15,319537,1,1,42297.84965 +1,0,1,0,26594,17,19,140264,1,1,24339.567 +0,0,1,0,3672,100,120,16333,1,0,10969.97368 +0,1,0,1,33596,23,23,104648,0,0,6289.555494 +0,0,1,1,7898,15,18,26614,0,1,3894.353505 +0,0,1,0,120053,16,20,348026,1,0,25019.12665 +1,0,0,1,26374,155,101,156262,1,0,21733.72894 +0,0,1,1,11169,74,59,39394,0,0,4933.718505 +0,0,1,0,10343,117,101,29140,0,0,4516.709131 +0,0,0,0,119811,74,79,338780,1,1,41893.04823 +0,0,1,1,55155,41,28,209808,1,1,28148.2893 +0,0,1,0,25486,43,37,200461,0,1,18717.08297 +0,0,0,1,17133,115,114,43800,1,0,13719.34886 +0,0,1,0,10986,22,25,55203,0,0,3877.036058 +0,0,1,1,3720,120,118,29719,1,1,14819.01866 +0,0,0,0,37833,220,204,101318,1,0,19501.87294 +1,1,0,0,47449,11,12,251510,1,1,34882.6487 +1,0,1,1,8760,131,84,81243,1,0,19114.5469 +1,0,1,0,35447,135,124,110438,0,0,12688.82126 +0,0,1,1,13150,56,43,70848,0,1,8498.405233 +0,0,0,1,47170,32,24,163505,1,1,23199.82368 +0,0,1,1,16848,30,24,131324,1,0,13032.75978 +1,0,0,1,27695,11,7,237921,0,1,25148.51695 +0,0,1,0,77625,74,94,512799,1,1,61842.53886 +0,0,0,1,134562,15,18,343625,1,1,40063.98894 +0,0,1,1,47989,152,179,255746,1,1,38496.75612 +0,0,1,1,4747,14,11,13521,0,0,2955.875518 +1,0,1,1,29087,61,45,97119,1,0,17603.9525 +1,0,1,0,59544,54,45,218979,0,1,25110.63315 +0,0,1,1,8135,85,74,58753,0,0,8419.202725 +0,0,0,1,85468,22,25,314297,1,1,38741.92019 +0,0,1,1,12137,62,59,101757,1,1,19272.74509 +0,0,1,0,33471,17,17,205233,1,1,27122.56479 +0,1,0,1,4448,22,24,13398,0,0,3398.11952 +0,0,0,1,41467,31,29,264171,1,1,32723.72263 +0,0,0,1,12207,99,59,60497,0,0,4879.220052 +1,0,0,1,63941,32,29,235279,1,1,34810.55066 +1,1,0,1,11578,41,33,32790,0,1,11304.1594 +1,0,1,0,169712,22,23,426030,0,0,18095.77891 +0,0,1,1,20110,49,42,105859,1,0,13418.95409 +0,0,0,1,84131,83,59,247807,0,0,11316.10913 +0,0,1,0,2335,79,102,15524,0,1,6788.539469 +0,0,1,1,22064,106,64,161425,1,1,26358.09252 +1,0,0,0,64113,46,48,204894,1,0,20667.64091 +0,0,1,0,50248,171,140,208390,1,1,32069.98303 +1,0,1,1,8338,63,79,48841,1,1,16647.35755 +0,0,0,1,19899,38,45,73329,0,0,5064.05926 +1,1,0,1,110002,264,226,275491,1,1,50307.33259 +0,0,1,1,95362,33,25,256642,0,1,23432.22518 +0,0,0,0,11397,26,31,30444,0,1,5145.141759 +1,0,0,1,43817,213,271,379290,1,1,60203.89043 +0,0,1,0,16580,88,56,99898,0,0,4183.562699 +0,0,0,0,12933,51,47,53317,1,1,11588.65023 +0,0,1,1,5449,20,19,15266,0,0,3658.278053 +1,1,0,1,100167,48,51,358940,1,1,52019.16027 +1,0,1,0,74555,22,17,226384,1,1,32896.6055 +0,0,0,0,47194,152,155,126111,1,0,17573.86443 +0,0,1,0,3926,55,54,10454,0,1,3504.503079 +1,1,0,1,104055,18,13,305435,1,1,43612.60606 +1,0,0,0,16056,28,28,49332,1,1,15206.90912 +0,0,1,1,54424,67,60,253921,1,0,21686.95205 +0,0,1,1,18775,26,27,124740,1,1,18199.69751 +0,0,1,1,14302,43,42,84235,0,0,6303.212166 +0,1,0,1,16525,50,41,91894,0,1,13907.25597 +0,1,0,0,9501,53,50,38192,1,0,10861.04743 +1,0,1,0,28947,146,134,89687,1,1,23753.70897 +0,0,0,0,38258,338,327,227586,1,0,30043.81505 +1,1,0,1,5411,110,101,30244,0,0,10369.50698 +0,0,0,1,26835,243,187,99838,1,0,19400.48493 +0,1,0,1,12232,75,49,38683,0,1,8939.53286 +1,0,1,0,3209,120,145,15965,0,0,11643.57355 +0,0,0,0,31661,127,157,117083,1,1,24038.8003 +0,0,0,1,11598,19,12,33497,1,1,10677.20017 +0,0,0,1,83791,18,22,273756,0,1,25331.78145 +1,0,0,0,50206,12,15,149350,1,0,17083.90622 +1,0,0,1,45543,10,10,140943,0,0,10211.61067 +0,0,0,1,27402,54,44,79984,0,0,3774.022142 +0,0,0,1,67216,13,13,369854,1,1,42810.83364 +0,0,1,1,35233,30,18,90039,1,0,10666.32347 +0,1,0,1,10080,21,18,79275,0,0,6123.187067 +1,0,0,0,15824,138,84,86526,1,0,17595.12246 +0,0,1,1,7156,44,55,24417,1,0,10982.72842 +0,0,1,1,22383,30,33,133443,1,0,13957.71172 +0,1,0,0,31957,21,26,151588,1,0,14916.32438 +0,0,1,1,7142,76,90,38019,0,1,8745.555866 +1,0,0,0,25718,135,137,69886,0,0,12375.25549 +0,0,1,1,25393,70,61,81306,0,1,9984.021432 +0,0,1,1,4272,244,278,18389,0,0,12912.11771 +1,0,0,0,14482,30,21,89187,1,1,19112.70034 +1,1,0,1,36515,27,31,91536,0,1,15703.48185 +0,0,1,1,83954,22,13,414028,1,1,48411.73584 +1,0,1,1,5359,24,18,17539,1,0,12403.82279 +0,0,0,0,25396,93,81,169520,1,1,24682.17914 +0,0,1,1,8218,93,103,21545,0,1,8013.702893 +1,0,1,1,20775,24,14,71600,0,0,8119.431132 +0,0,0,1,17762,243,237,70104,1,1,25916.50473 +0,0,0,0,49517,11,11,189560,1,1,23457.60884 +0,0,1,1,15628,25,15,71621,1,0,10537.24554 +0,0,1,1,13660,38,24,46070,0,0,5594.256812 +0,1,0,1,5732,36,46,27540,1,1,12872.95036 +1,1,0,1,10033,74,65,56750,1,0,16952.31794 +0,1,0,1,3970,12,10,21244,0,1,7616.587121 +1,0,0,1,54925,102,85,171154,1,0,22461.02533 +0,0,0,1,21284,37,22,103826,0,1,11795.08376 +0,0,0,1,12314,37,43,33933,0,0,4730.494451 +0,0,1,1,26324,170,133,74431,1,1,18602.41447 +0,1,0,0,77559,71,57,290795,1,1,37418.4234 +1,0,1,1,3976,42,25,14692,1,1,12010.44682 +0,0,1,1,65118,14,9,214925,1,1,29424.86975 +0,0,1,0,73434,11,10,254975,1,0,19171.56902 +0,0,1,1,52217,21,18,171420,1,1,24924.6775 +0,0,1,1,20795,15,14,61856,1,0,10738.01935 +0,0,1,1,10102,247,306,55864,0,0,16188.5893 +0,0,1,1,53540,99,80,184557,1,0,19937.30604 +0,0,1,1,3516,134,132,26005,1,0,12624.47883 +0,0,0,1,31130,75,52,138529,1,1,22232.06116 +0,0,1,1,49152,98,126,161032,1,1,29057.13155 +1,0,1,1,6429,34,42,47778,1,0,15575.59092 +0,0,0,1,24143,19,17,150038,1,1,22069.98835 +0,0,1,0,10347,177,154,29299,1,0,12110.95703 +0,0,1,1,25178,31,20,90695,0,1,10792.43108 +0,0,1,0,14148,33,23,48945,0,0,3046.58905 +0,1,0,0,113561,59,67,291874,1,1,38101.29937 +0,0,1,1,17240,17,16,62576,1,0,10924.6916 +0,0,1,1,3641,74,56,15051,1,0,12652.01364 +0,0,1,0,12569,47,60,61120,1,0,11758.11622 +0,0,1,1,74494,28,22,224557,1,0,21570.27881 +1,0,1,0,75145,13,13,436643,1,1,51481.47076 +1,0,1,1,31558,25,28,128626,1,1,25264.31538 +0,0,0,1,18540,30,27,72997,0,1,9055.466622 +0,1,0,0,34325,41,40,120679,0,0,8536.313565 +0,0,1,1,5720,57,61,26848,0,0,6769.439448 +0,1,0,1,12020,203,234,30600,0,0,12671.6971 +0,0,0,1,20576,66,50,152654,1,0,16145.62882 +0,1,0,1,16597,34,33,73011,0,0,6803.452299 +0,0,0,0,25021,39,33,159410,0,0,5687.746612 +1,0,1,1,41644,146,93,244406,1,1,37938.17821 +0,1,0,0,12911,49,62,39582,1,1,14715.84123 +0,0,1,0,24486,84,108,91734,1,0,13786.72061 +0,0,1,1,24412,80,51,124649,1,0,13916.59201 +0,0,1,1,35015,40,30,89385,1,1,17680.56379 +0,0,1,1,81286,11,13,308051,1,1,36980.83711 +0,1,0,1,36647,155,108,196940,0,1,23306.00006 +0,0,1,1,26933,101,76,123624,1,0,16293.66391 +0,0,0,0,32174,45,31,240885,1,1,31262.08172 +1,0,1,1,41344,41,49,115047,0,0,9719.121829 +0,0,1,1,4077,13,12,19584,1,0,9121.758337 +0,0,0,0,2000,27,24,13828,0,0,1950.003882 +0,0,0,1,19511,42,47,147592,1,1,21766.20787 +0,1,0,1,15166,203,125,42304,0,0,9369.386154 +1,0,1,0,33239,13,13,99772,1,1,19402.91401 +0,1,0,1,68209,50,59,180279,0,1,20803.41705 +0,0,1,0,4131,103,83,20340,1,1,10261.71063 +0,0,1,1,20910,134,170,143777,0,1,19449.43461 +0,1,0,1,4243,16,13,21518,1,0,12566.93159 +0,1,0,1,3108,110,70,17153,0,0,6432.405158 +0,1,0,0,31160,20,21,146114,0,0,6955.52316 +0,0,1,1,2924,71,56,21425,1,0,11287.25064 +0,0,1,1,6267,130,160,22855,0,0,8529.479435 +0,0,1,1,2398,49,34,23887,0,0,3459.079997 +0,0,1,1,38047,15,15,180869,1,0,14293.74411 +0,0,1,1,125515,34,42,621723,1,1,69502.99747 +1,1,0,1,47333,118,137,363714,1,0,35334.1679 +0,0,1,0,32360,50,50,178707,1,0,15949.85578 +1,1,0,0,49230,19,11,185925,0,0,13743.69156 +0,0,0,0,86318,27,24,415986,1,1,47684.93267 +0,0,0,1,23830,330,232,83006,0,1,17790.00111 +0,1,0,1,14024,49,31,111457,1,1,19360.29741 +0,0,1,0,45845,159,110,157802,1,0,18662.28483 +1,0,1,0,19093,29,27,189251,0,1,21412.61012 +0,0,1,1,4333,16,17,11813,1,0,7057.716994 +0,0,0,0,8334,113,110,59012,0,0,5815.313267 +0,0,1,0,7503,77,86,39893,0,1,6041.691655 +0,0,1,1,3344,37,35,20632,0,0,2487.28966 +0,0,1,1,8425,43,39,25626,0,1,4957.736644 +0,1,0,0,19158,18,14,80909,1,0,12110.20072 +0,0,1,0,19452,52,65,56500,1,0,9610.462044 +1,0,1,0,21621,213,133,68031,1,1,22110.9188 +0,0,1,1,27904,18,13,160297,0,0,8022.861962 +1,0,1,0,19168,31,30,146996,1,1,23680.3172 +0,0,0,1,19530,51,52,133215,0,1,12813.65007 +0,0,0,0,2295,87,85,11652,1,1,12866.98374 +0,0,0,1,10431,111,109,43425,1,1,14834.34432 +0,1,0,0,1914,130,129,18960,1,0,14030.81194 +0,1,0,1,11769,10,12,101790,1,0,13709.73817 +0,1,0,1,40599,11,13,391334,1,1,47007.10273 +0,0,1,1,3673,93,66,29982,0,1,5465.561091 +0,0,1,0,43813,300,291,117866,0,1,22034.16676 +0,0,0,1,24922,28,36,238370,1,1,30232.00281 +0,1,0,1,29650,75,58,129825,1,1,26170.72392 +0,0,1,1,57779,117,100,155279,0,1,16874.24697 +0,0,1,1,8710,18,19,37352,0,0,4343.695185 +0,0,1,0,12372,31,37,87831,1,0,10074.31003 +1,0,1,1,25210,152,93,82789,0,1,16246.96505 +1,0,0,1,35232,95,93,88361,0,0,11747.05758 +0,0,1,1,2910,28,22,19665,1,0,7535.963856 +0,0,0,1,4697,11,13,19975,0,0,3153.999581 +1,0,1,0,22028,139,142,60867,1,0,18548.92683 +1,0,1,1,20085,42,48,55217,0,0,8837.617126 +0,1,0,1,25176,16,11,75388,1,1,15671.0671 +0,0,0,1,55601,128,152,364693,1,1,48422.49158 +0,0,1,0,60042,34,42,180787,0,0,7457.534234 +0,0,0,1,43814,44,28,139225,0,1,14923.19864 +1,0,1,1,8175,175,174,57797,1,0,20795.30641 +1,0,1,1,20911,23,27,157897,1,0,18476.45679 +1,0,1,1,7020,92,57,36048,1,1,16736.85308 +0,1,0,0,24809,16,13,82969,1,1,18157.45591 +0,0,1,1,5037,27,22,44426,0,1,5106.006203 +1,1,0,1,8911,37,25,36052,1,0,15058.3902 +0,0,0,1,6184,66,52,34816,0,1,6351.385349 +0,0,1,0,54124,26,24,360237,1,1,41718.57662 +0,0,1,0,6964,12,8,25839,0,1,2562.555572 +0,0,1,1,37921,11,6,112185,0,1,9782.389138 +0,0,1,1,36046,94,91,123950,0,1,14914.87993 +1,0,1,0,20439,34,32,53451,1,0,13411.55529 +0,0,0,1,34471,70,77,112570,0,0,8116.201911 +0,0,1,1,26216,80,75,83602,1,0,14465.35069 +0,0,1,1,72410,113,98,213094,0,1,23650.45983 +0,0,0,1,24612,11,6,83804,1,1,14458.94926 +0,0,0,1,20156,50,44,105410,0,0,5360.756553 +0,0,0,1,65471,131,86,312961,1,1,41559.2825 +0,0,1,1,5085,52,64,19455,0,1,5690.371584 +0,0,0,1,26726,35,35,80233,1,1,15856.81648 +0,0,0,1,28145,100,91,75364,0,0,8782.850133 +1,0,0,1,6326,88,106,19067,0,0,11346.83188 +1,1,0,1,5595,200,168,32218,0,0,16131.02992 +0,0,1,1,8670,175,151,63222,0,1,13452.202 +0,0,0,1,48104,110,136,121870,0,1,16826.19827 +0,0,0,1,2715,50,52,18984,0,0,3276.441955 +0,0,1,1,7555,81,62,25721,0,0,5182.933638 +0,0,0,0,8202,52,50,37711,1,0,10393.92094 +1,0,1,0,60229,16,10,169389,0,1,17528.55139 +1,0,1,1,76500,16,11,303731,1,1,39653.45191 +0,0,1,1,187785,125,132,486587,1,1,60619.11284 +0,0,1,1,18389,15,14,84984,0,0,4554.40667 +0,1,0,1,6423,16,14,61674,0,0,7472.361427 +0,0,0,1,2961,42,34,10950,0,0,3248.018934 +0,0,1,0,44693,189,159,138383,0,1,20930.42881 +1,0,1,0,25553,146,118,72761,1,1,22416.93314 +0,0,1,0,12754,29,31,47655,0,1,5814.054048 +0,0,1,1,5905,15,16,26206,1,0,8830.422136 +0,0,1,1,16602,179,219,78992,0,0,12512.88604 +0,1,0,1,9423,39,35,64041,0,1,9039.266551 +0,0,0,1,18287,55,61,76185,1,1,15442.85173 +0,1,0,0,7816,18,15,25347,1,0,9855.571089 +1,1,0,1,2889,103,73,23204,0,0,10374.1152 +1,0,0,0,2238,31,25,17128,0,1,6439.1138 +0,0,1,0,17184,13,9,45179,0,0,3240.677761 +0,1,0,1,71207,22,15,324583,1,1,41789.21678 +1,1,0,1,95864,126,100,313896,1,1,48722.19318 +0,0,0,0,61565,88,82,155731,1,1,23979.2456 +0,0,1,0,52056,84,78,249586,0,1,24583.70491 +0,0,1,0,42209,287,368,132812,0,1,26763.89677 +0,0,1,1,30881,350,346,78447,0,0,19044.33861 +0,0,1,1,32098,76,47,94511,0,1,10933.44821 +0,0,0,0,125643,17,15,353921,1,1,40255.44235 +0,0,0,1,15553,40,39,102679,0,0,6468.444106 +0,1,0,1,27005,129,130,116103,1,0,18377.2377 +0,1,0,0,42988,167,172,206798,0,0,14047.61075 +0,0,0,0,7672,32,39,33964,0,1,3792.469079 +0,0,0,0,8166,36,33,35400,0,1,4509.107352 +1,0,1,0,31533,158,166,160557,1,1,31782.88533 +0,0,1,1,12678,82,51,51160,0,1,6702.151585 +0,1,0,1,10525,28,27,90436,0,1,11455.40518 +0,1,0,1,14493,111,83,47701,0,0,8361.661382 +0,0,1,0,9689,23,20,30535,1,0,7362.893813 +0,1,0,1,42165,35,43,182719,0,1,20467.30031 +0,0,1,1,37048,86,64,128094,0,0,8159.382761 +0,0,1,1,148792,28,31,387643,1,1,46328.9537 +0,1,0,1,22286,30,26,64435,1,1,17184.19603 +0,0,0,1,17740,58,55,60088,1,1,14089.96524 +0,0,0,1,9428,26,16,35205,1,1,11627.70748 +0,0,1,1,62294,142,179,236867,1,1,36651.53787 +0,0,1,1,26526,151,195,85307,1,0,16851.32486 +0,1,0,0,9974,36,35,26139,0,1,6011.221209 +0,0,1,1,16751,12,8,60287,0,0,2161.067162 +0,0,1,0,2017,16,11,10767,0,1,1075.306142 +0,0,1,0,29482,40,51,287594,0,1,26263.23327 +1,0,0,1,66115,14,11,175283,0,1,20758.60175 +0,0,0,0,18516,54,56,108860,1,0,12850.2833 +0,0,0,1,10060,158,145,35440,0,0,8283.110327 +0,0,1,1,3280,71,54,16764,0,0,5008.528603 +0,0,1,1,20957,178,196,70511,1,1,23263.80294 +1,0,1,1,11821,48,35,58965,0,0,8584.94953 +0,0,1,1,14031,12,14,62941,1,0,10127.11694 +0,0,1,1,52153,41,27,204358,0,0,9614.969945 +0,0,1,1,40238,184,122,355100,1,1,46075.22695 +1,0,1,1,5342,44,44,18610,0,1,8958.329522 +1,1,0,0,49513,35,34,273202,1,0,26057.50382 +0,0,1,1,36169,154,184,171836,0,1,22141.74528 +0,0,0,1,5712,10,12,30801,1,1,9849.800909 +0,0,1,1,14457,51,63,101818,0,0,7633.370285 +1,0,0,1,11786,24,28,69099,0,0,8669.574323 +0,0,1,0,9982,117,95,27981,0,1,7712.577363 +0,0,1,1,28616,32,36,111547,1,1,18512.21463 +0,0,0,1,8003,149,89,33333,0,1,5532.672742 +0,0,1,1,13539,24,22,65113,0,1,5200.339559 +0,0,1,1,21443,13,9,93531,0,0,4438.57358 +1,0,1,0,55566,71,51,380317,1,1,50858.10683 +0,0,1,0,2780,74,45,14419,0,0,2272.127562 +0,0,0,1,37114,36,43,157622,1,1,23926.54642 +0,0,1,0,6722,67,82,29573,1,1,11457.61484 +0,0,1,1,6531,56,47,36447,1,1,12740.23798 +0,0,1,0,11264,36,32,28892,0,1,4429.202694 +0,0,1,1,43847,19,19,157285,0,0,8283.264431 +0,0,1,1,5582,59,62,17240,0,0,6302.147101 +1,0,1,0,4154,37,26,14490,0,1,7030.601816 +0,0,0,1,85025,13,12,227981,1,1,28421.08444 +0,0,1,1,12019,58,68,44029,1,0,12317.583 +0,0,1,1,27506,188,135,219199,0,1,25131.74074 +0,0,1,1,64011,155,108,191203,1,1,31791.75098 +0,0,1,0,10387,130,123,60886,0,1,11196.20622 +0,0,1,1,54804,104,90,230590,1,0,22851.47202 +0,1,0,1,3739,136,100,10517,0,1,7861.118021 +0,0,1,0,43218,90,59,240968,1,1,31185.63017 +1,0,0,1,133727,58,57,343435,1,0,29186.81641 +0,0,0,1,2145,80,74,19592,0,1,7154.496475 +0,0,1,0,18340,19,19,46709,0,0,1788.257783 +0,0,1,1,40388,23,28,163068,1,1,22514.99273 +0,0,1,0,27683,14,8,79956,0,1,8201.242114 +1,0,0,0,37653,10,10,169230,0,0,9536.348658 +1,0,1,1,43133,88,89,126433,1,0,22077.31349 +0,0,0,0,4783,32,23,14964,0,1,2205.170846 +0,1,0,1,10860,46,53,31509,0,1,8237.521679 +0,0,1,1,14286,73,47,39297,0,0,2805.656965 +0,0,1,1,7035,24,28,27539,0,0,3451.888589 +0,0,1,1,185820,45,36,561249,1,0,38349.62419 +0,0,0,1,65137,122,147,237129,1,0,24103.33576 +0,0,0,0,11679,50,41,99386,1,1,16866.80194 +0,0,1,0,6399,25,21,16146,1,1,7812.085994 +0,0,1,1,72852,71,83,227646,1,1,32424.35195 +0,1,0,1,5047,24,29,28839,1,0,10910.47314 +0,0,0,1,6566,21,17,19778,0,0,1526.97737 +0,0,1,1,63185,282,292,163772,1,1,33763.74732 +0,0,1,1,1923,15,18,17196,0,0,3599.09488 +1,0,0,0,21061,23,24,134540,0,1,14828.91803 +0,0,0,1,6541,20,15,39541,0,0,2689.880991 +0,0,1,0,24886,34,35,98687,1,0,12058.38907 +0,0,0,1,7147,22,14,39237,0,0,3793.687727 +0,0,0,1,20846,64,56,112843,0,1,12762.26916 +1,0,1,1,23796,48,49,60288,1,0,14779.28191 +0,0,1,1,17231,58,48,74124,1,1,15862.94307 +0,1,0,1,17530,28,20,115727,0,1,13850.06656 +0,0,1,1,6434,20,20,36841,0,0,2466.731959 +0,0,1,0,12337,75,45,53062,1,0,10506.62169 +0,0,1,1,15110,22,15,44091,0,1,5042.730342 +0,0,1,0,14587,82,83,46146,0,1,8148.839687 +0,0,1,1,9466,51,64,44456,0,1,6819.859368 +0,0,0,1,10197,105,108,37298,0,1,8699.677047 +0,0,1,1,88324,20,21,248728,0,0,10088.00987 +0,0,0,1,12919,16,17,89955,0,1,9269.945628 +0,1,0,1,20893,47,50,66235,1,1,15915.97157 +0,0,0,1,34937,27,18,204745,1,1,27843.71872 +1,0,0,1,21028,64,55,118109,1,0,16707.76303 +1,0,1,1,35716,176,145,258150,1,1,43287.31095 +0,1,0,1,26653,10,12,196556,1,0,17609.17247 +0,0,1,1,21425,81,69,93682,0,0,5730.848545 +0,1,0,1,132165,128,117,369686,1,1,50130.92151 +0,1,0,1,10282,12,8,93054,1,0,12674.01825 +0,0,1,0,7932,68,61,64418,0,0,4660.333545 +0,0,1,1,41186,31,31,133654,0,1,13759.22737 +1,0,1,0,3082,33,27,30297,1,0,12782.91284 +0,1,0,0,44056,23,23,152713,1,1,24418.0332 +0,0,0,1,34057,38,27,119427,1,1,19079.72898 +0,0,1,1,3888,54,43,29769,0,1,6113.80667 +0,0,0,1,13275,80,54,41946,0,1,6875.8485 +0,0,1,0,73103,39,42,224032,1,1,28756.2661 +0,0,1,0,8505,62,55,29014,1,0,7046.27083 +0,0,1,1,18993,82,96,153617,1,0,17746.91304 +1,0,0,1,25259,50,47,105468,0,1,15388.45275 +1,0,0,1,5900,67,83,25413,0,0,9931.152373 +0,0,1,1,4799,15,13,25828,0,0,1808.608641 +1,0,1,0,6233,93,97,55279,0,0,8637.988017 +1,0,0,1,1855,82,76,13145,1,0,14090.36477 +0,0,0,1,16861,55,34,153278,0,0,6254.98617 +0,0,0,1,43118,32,35,127226,0,0,6983.619267 +0,0,1,0,23597,264,220,73249,1,1,22659.50419 +0,0,0,0,22245,12,10,96594,0,1,8108.028387 +0,0,0,0,12420,90,78,37446,0,0,4271.401033 +0,0,1,1,78109,13,12,232974,1,0,18953.51594 +0,0,1,1,24163,24,17,162345,1,1,21509.95807 +0,1,0,0,17457,23,19,57875,0,0,4471.087288 +1,0,1,1,16434,37,25,82730,1,0,16182.27909 +0,0,1,0,5115,19,12,17494,0,1,2799.743004 +0,0,0,0,13959,40,49,101399,0,0,4744.473057 +1,0,0,1,9149,75,71,53781,1,1,18086.028 +1,0,1,1,5180,105,90,13802,1,0,13681.09963 +0,0,1,1,14921,136,132,47549,0,0,9496.221711 +0,0,0,0,19206,22,22,67538,1,1,12374.67423 +0,0,1,1,4865,120,118,32082,0,0,7445.272025 +1,0,0,1,7157,57,68,17917,0,1,9655.977756 +0,0,1,1,55763,72,70,274217,1,1,36554.17464 +0,0,0,1,45926,32,25,121688,1,0,13102.90052 +0,0,0,0,12303,53,59,83824,1,1,15267.41038 +0,0,1,1,12864,15,17,64482,1,1,14042.46102 +0,0,0,0,15689,21,15,41964,0,0,1113.172781 +0,0,0,1,49295,85,58,131982,1,0,15754.24544 +1,0,0,1,10933,56,38,57713,0,1,10785.29754 +0,0,0,1,42488,74,55,143637,1,1,24193.11289 +0,0,1,1,17395,105,117,48496,0,1,9994.092275 +0,0,1,1,50126,34,43,125504,1,1,20932.81831 +1,0,0,1,19693,73,65,75585,1,0,16648.60247 +0,0,0,1,44522,88,104,216508,1,1,30794.48969 +0,0,1,1,71691,12,11,329923,1,1,38545.90173 +1,0,0,0,27011,34,29,96403,1,1,19490.37588 +1,0,1,1,41441,29,20,105035,0,0,10722.2157 +0,1,0,1,32052,51,44,101540,0,1,13513.02957 +0,0,1,0,10763,39,26,29301,0,1,4918.077477 +1,0,1,1,17253,62,72,48319,1,1,18683.07929 +0,0,1,1,48633,37,34,126911,0,0,5982.366767 +0,1,0,1,4028,46,31,22089,1,1,11764.24383 +0,1,0,1,30229,92,113,248850,1,1,37524.42834 +0,0,1,1,2029,13,12,18641,0,1,5965.460187 +1,0,1,0,19544,14,17,60230,1,0,12723.99488 +1,0,1,0,3438,110,108,31135,1,1,17993.10381 +0,0,0,1,38621,99,107,293510,1,1,39756.50128 +1,1,0,1,38283,73,48,121796,0,1,20599.22185 +1,0,1,1,4328,125,119,16090,0,0,9928.621027 +0,0,1,1,11517,33,22,55005,0,0,3714.651945 +0,1,0,1,2778,80,65,13682,0,0,5325.555068 +0,0,0,1,72346,14,10,194938,0,0,7866.984969 +0,0,1,1,37139,227,221,134594,1,1,29605.25558 +0,0,0,1,34975,16,15,140578,0,1,12305.625 +0,0,0,1,25050,34,36,70904,1,1,15090.02796 +1,0,0,1,17714,16,11,48556,0,0,6507.064194 +0,0,1,1,16104,16,12,47197,0,0,4627.528321 +0,0,1,1,26560,46,49,87562,0,0,6624.822918 +1,0,1,1,13485,16,11,69563,1,0,15812.20706 +0,1,0,1,71203,21,26,245123,0,1,25313.38013 +1,1,0,1,8016,61,46,25613,0,1,10625.00185 +0,1,0,1,7958,38,42,77649,1,1,17474.30546 +0,0,1,1,9141,65,40,59573,1,0,11450.96151 +0,0,1,0,73924,58,65,316663,0,0,13905.29862 +1,0,1,1,3798,11,11,15615,0,0,6929.75123 +0,0,0,1,47051,55,61,198478,1,1,28593.84247 +0,0,0,1,30457,39,39,77466,1,1,17704.23462 +1,0,1,1,12036,22,18,47870,1,0,12267.90934 +0,1,0,1,6063,45,31,17352,0,0,4157.376807 +0,1,0,0,16212,170,140,63835,0,0,11058.92666 +0,0,0,1,25627,68,64,73479,0,0,4728.372297 +0,0,1,1,2595,134,169,12752,0,1,9363.243188 +1,1,0,0,10994,19,21,44245,1,1,15004.20719 +0,0,1,1,11023,93,92,100300,1,0,12888.18501 +1,0,0,0,23399,82,87,63825,0,1,10876.15546 +0,1,0,1,66950,78,72,229300,0,1,26308.99261 +0,0,0,0,9442,41,45,35149,0,0,3897.27888 +1,0,1,1,7350,59,43,56526,0,1,11354.20289 +1,1,0,1,22699,21,25,89220,1,1,21847.66321 +0,0,0,1,15888,53,40,134756,1,1,20548.47139 +0,1,0,0,90030,71,56,227082,1,0,21740.35477 +1,0,0,1,15860,29,28,65533,1,0,15546.20046 +1,0,1,1,9837,87,99,49832,1,1,18958.14154 +0,0,0,1,16449,122,96,52638,1,1,16276.81724 +0,0,1,1,52276,23,15,175190,1,0,15262.2085 +0,0,0,1,30669,24,17,125355,1,1,20924.00185 +0,0,1,1,49274,40,41,146268,1,1,24373.30679 +0,1,0,0,9733,70,60,34800,1,0,11246.90037 +0,0,1,1,9333,71,65,29377,0,1,5236.862906 +0,0,1,1,91066,20,12,239951,1,1,31190.52552 +0,0,1,0,6268,210,222,25450,0,0,9725.253749 +0,1,0,1,31784,31,27,118649,1,1,23041.97953 +0,0,1,1,12565,49,59,101747,1,1,18578.82007 +0,0,0,1,5898,39,37,22433,0,1,4665.17618 +0,0,0,1,12432,21,17,68448,1,1,13605.75904 +0,0,1,1,24080,15,14,99400,1,0,13929.53007 +0,0,1,0,50785,23,16,170686,1,1,21878.81944 +1,0,1,1,46151,40,24,228243,1,1,34461.63062 +0,0,1,1,29833,19,16,200222,1,1,25888.90355 +0,0,0,1,6223,141,155,47986,1,0,15543.72503 +1,0,0,1,5389,21,22,46579,1,0,13139.07023 +0,1,0,1,9687,10,12,73454,0,0,5929.673647 +0,0,0,1,13982,181,186,68710,0,1,13725.05944 +0,0,1,0,28634,27,19,86259,1,1,14586.45649 +0,0,0,1,19698,223,194,68434,0,0,12146.46301 +1,1,0,1,48400,46,52,152151,1,1,29080.33788 +1,0,0,0,36848,95,89,154038,1,1,29094.56132 +1,1,0,1,18391,25,19,48419,0,0,9525.496591 +0,1,0,1,11356,26,25,33958,0,1,6615.064745 +0,0,0,0,4571,28,22,12955,0,1,889.9756528 +0,0,1,1,12171,12,13,42237,0,0,1408.456289 +0,0,1,1,3613,35,30,30750,0,1,6131.616399 +0,0,0,1,20584,41,49,52063,1,0,10898.83007 +0,0,1,0,5651,22,15,14223,0,0,1608.868027 +0,0,0,1,7610,14,16,19559,0,0,2836.787536 +0,0,1,1,19834,21,22,141636,1,1,21384.87729 +0,0,1,0,32146,78,61,88939,1,1,17804.1785 +0,1,0,1,12634,25,15,89284,1,1,18148.86182 +0,0,1,1,216847,58,59,576931,1,0,39189.79226 +1,0,1,1,8892,18,13,79262,0,0,5216.817245 +1,0,0,1,36354,47,34,334533,0,1,34620.77186 +0,0,0,0,15505,236,154,78785,0,0,10416.76761 +0,0,1,1,43928,14,9,142408,0,0,6604.038947 +0,0,1,0,5047,34,38,21189,0,0,1496.571961 +0,0,1,0,33099,32,21,159814,1,1,20189.74348 +0,0,0,0,9390,27,19,51588,0,1,5638.55887 +1,0,0,1,26849,82,89,100878,1,0,18010.79571 +0,0,1,0,5514,39,31,53677,0,1,7323.304046 +0,0,1,1,9182,25,26,25319,1,1,7832.328389 +0,1,0,1,90715,215,134,252573,0,1,30418.1706 +0,0,1,0,29291,22,25,219386,1,1,28408.8204 +1,0,1,1,2769,22,20,12621,0,1,7731.261504 +1,0,0,1,24130,39,39,238170,0,1,23439.61324 +0,0,1,0,22121,206,242,94992,1,0,21355.66519 +0,1,0,0,15892,79,49,48896,0,0,5184.767479 +0,0,0,0,14824,23,18,49678,1,0,5816.155015 +1,1,0,1,55185,114,146,195651,1,0,25214.32095 +0,1,0,1,4890,32,27,19932,0,0,4906.737437 +0,0,0,0,4574,45,56,11464,1,0,7700.722688 +0,0,1,0,32713,42,36,89221,0,1,6824.258074 +0,0,1,0,84172,83,82,289225,1,1,37057.45653 +0,1,0,1,18823,22,25,144786,1,0,16952.15638 +0,0,1,1,26311,26,15,106429,0,1,11563.57478 +0,0,1,1,14319,49,62,49910,0,0,4846.826271 +0,1,0,0,8243,53,67,20861,0,0,6473.5128 +0,0,1,1,2457,24,17,12147,0,0,1816.05151 +0,0,0,0,32519,23,14,121554,0,0,3944.826915 +0,0,1,1,14614,53,34,46570,1,0,9455.624525 +0,1,0,1,31770,123,117,79885,1,0,17573.13979 +0,0,1,1,26622,65,61,76943,0,1,10611.93383 +0,0,0,0,43574,26,25,122852,0,0,5382.789214 +0,0,1,1,20046,72,55,107275,1,0,12982.07293 +0,0,1,1,22681,56,44,71702,0,1,9683.228815 +0,0,1,1,33502,48,42,100847,0,1,12161.26982 +0,0,1,0,7014,12,7,48094,1,0,8736.408395 +0,0,1,1,31031,68,55,87512,0,0,5908.813014 +0,0,0,1,55835,47,31,339298,1,1,39957.12295 +0,1,0,1,26338,56,34,91201,1,0,13443.9427 +0,0,0,1,54085,18,15,164706,1,1,23309.57527 +0,0,1,0,4195,78,51,11390,1,0,7885.634278 +0,0,0,1,7331,11,11,21523,0,1,3581.785919 +1,1,0,1,7216,39,33,45321,1,0,15447.03202 +0,0,1,0,23982,64,47,60092,1,1,12370.46019 +0,0,1,1,6898,77,86,58994,0,1,10328.88638 +0,0,1,1,10665,55,69,64120,1,0,11766.00932 +0,0,1,0,63286,38,42,160582,0,1,16046.82812 +0,0,0,1,14194,34,26,54669,0,1,7219.732899 +0,0,0,1,12785,77,98,38887,1,1,14129.6421 +1,1,0,1,13980,26,18,38892,0,0,9816.213152 +0,0,1,1,36150,12,12,115412,0,1,10856.31406 +0,0,1,1,17826,63,52,50404,1,0,12470.55267 +1,0,1,0,16838,59,56,64514,0,0,9496.053381 +0,0,1,0,18778,35,31,55095,1,1,12093.38079 +0,0,1,0,4459,55,63,25729,1,0,9576.578823 +0,0,0,0,135692,178,213,618047,1,1,77528.81537 +0,0,1,1,3675,26,17,25096,1,1,9942.700669 +0,0,1,0,42914,14,15,254051,1,1,31201.35727 +0,0,0,0,108334,218,192,291538,1,1,43629.75392 +0,0,0,0,2444,27,28,15688,1,0,7456.745373 +0,1,0,1,24130,26,31,86989,0,0,7454.659858 +1,0,0,1,42664,41,35,134974,1,0,18862.37313 +0,0,0,0,9113,59,61,52406,0,0,5423.927958 +0,0,0,1,11283,55,50,70135,0,0,6679.766799 +0,0,1,1,30243,169,203,247768,1,0,28086.67604 +0,1,0,1,15694,24,30,84643,1,0,12468.53816 +0,0,1,1,31333,42,36,94666,0,0,5333.651027 +1,0,1,1,11357,59,55,75101,0,0,10147.14027 +1,0,0,0,10117,20,22,31883,1,0,11395.4986 +0,0,1,1,11262,17,18,39822,0,1,6271.528516 +1,0,1,0,33583,85,77,161560,1,1,29537.65233 +0,0,0,1,8268,121,142,29476,1,0,13640.71751 +0,0,0,1,9502,15,17,92715,1,1,15155.16726 +1,1,0,1,41557,114,98,332003,1,1,48728.22763 +0,0,0,1,3933,24,27,10768,1,0,8371.741955 +0,0,1,0,18772,52,43,51145,1,0,10488.15887 +0,0,0,1,50921,20,21,188000,0,1,15941.99873 +0,1,0,1,8786,40,37,68130,0,0,5894.35398 +0,0,0,0,59529,24,15,181704,1,0,14601.54169 +0,0,1,0,6277,20,12,30417,0,0,1835.451631 +0,0,0,1,60133,129,160,172913,1,1,32519.54847 +0,1,0,1,34596,42,42,105828,0,1,13968.27771 +0,0,0,1,12412,26,26,44954,0,0,4081.900584 +0,0,1,0,22916,77,58,62639,0,0,4171.2487 +1,1,0,1,32039,11,12,166808,1,0,21418.21206 +0,1,0,1,7566,103,74,64992,1,1,16998.38856 +0,0,1,1,4319,24,16,22989,1,1,7655.750618 +0,0,1,1,31472,51,38,125245,1,1,21036.03817 +0,1,0,0,41368,70,79,188113,0,0,10174.99537 +0,0,1,1,59887,24,14,176089,1,1,24340.71476 +0,0,0,1,11400,36,27,96026,0,0,5364.934704 +1,1,0,0,5190,18,22,23923,0,0,7416.942714 +0,0,1,0,21753,12,10,57036,0,1,5541.2102 +0,1,0,1,31244,33,28,161287,1,0,17158.78318 +0,1,0,1,14349,57,42,70116,1,1,16646.46136 +0,0,1,0,11225,72,57,45233,0,1,6225.88462 +0,0,0,0,58172,32,22,256127,1,1,32063.72955 +1,1,0,1,8119,71,75,47949,0,0,10124.27551 +0,0,1,1,8935,75,84,34192,0,0,6418.173838 +0,0,1,0,5955,11,13,56655,1,0,9543.824782 +0,1,0,1,21769,25,25,115008,0,0,7619.890571 +0,1,0,0,54710,57,37,145035,1,1,23480.37139 +1,0,1,0,20201,15,17,62075,0,0,5279.11638 +0,1,0,1,13168,24,16,70591,1,1,14673.046 +0,0,1,1,40516,43,45,142076,1,0,15228.48508 +1,0,1,0,10601,125,95,30445,1,0,15968.83095 +0,1,0,1,13643,63,44,89773,1,1,19043.65555 +0,0,0,1,17353,14,8,46146,0,1,6127.903977 +0,0,0,1,35304,15,9,123038,0,1,11212.04662 +0,1,0,1,29430,12,7,77690,0,0,5465.409643 +0,0,1,1,3181,66,52,23386,0,1,4155.17803 +0,0,0,0,58779,87,54,298788,1,1,39230.68034 +0,0,1,0,13745,29,32,52031,1,0,8031.738977 +1,0,1,1,12862,30,21,45223,0,1,10476.25359 +1,0,0,0,33526,15,17,112292,0,1,14707.84229 +0,0,0,1,22332,68,66,70057,1,0,13970.97992 +0,1,0,1,10233,51,65,26333,0,0,5677.481518 +1,0,1,0,92769,37,35,266060,1,0,22936.35093 +0,0,1,0,16204,56,68,44187,1,1,13024.00882 +0,0,0,1,7781,67,53,20585,1,0,10420.06826 +0,0,0,1,77157,66,47,260624,0,0,11088.22641 +0,0,1,1,2020,114,125,11821,0,0,7004.24437 +1,0,1,0,4208,23,14,25922,0,0,6203.790479 +0,0,1,1,38290,121,105,193140,1,0,20883.98191 +1,0,1,1,33872,26,17,85308,0,0,8423.4853 +0,0,0,1,72987,54,48,257270,1,1,33902.75033 +0,0,1,0,6755,42,35,48614,1,0,9096.476076 +0,1,0,1,22843,14,16,127307,1,0,15132.06439 +0,0,0,0,24694,49,53,63961,0,0,4710.660484 +0,0,0,1,7651,38,46,55372,0,1,7548.081392 +0,0,1,1,19961,146,128,87993,0,0,10474.25199 +0,0,1,1,19379,127,84,68897,1,1,17654.38141 +1,0,1,0,92517,107,89,525030,1,1,65050.49043 +0,0,1,1,24728,210,247,143155,1,1,32194.46318 +0,0,1,0,43604,20,15,193262,0,0,6513.837374 +0,0,1,1,3718,87,69,12453,1,0,10546.46439 +0,0,1,1,2656,79,100,15633,0,0,6646.573074 +0,0,1,1,13651,65,59,43059,0,0,4597.490173 +0,0,0,1,17527,20,20,115576,0,1,11554.45132 +0,1,0,0,62843,110,133,313708,0,1,34085.18712 +1,1,0,1,35866,28,32,125668,1,1,25430.41466 +0,0,1,0,47793,35,28,149714,0,1,13584.77301 +0,0,1,1,11016,58,55,35876,1,1,11025.45128 +0,0,1,0,4853,73,46,19814,1,0,9097.568257 +0,0,0,1,56856,25,16,143099,1,0,15055.79704 +1,0,1,0,22924,18,19,140933,1,0,17543.84502 +0,0,0,1,16492,57,52,105671,0,1,13203.11481 +0,1,0,1,3837,14,17,14864,0,1,4378.016683 +0,0,1,0,5218,15,10,42069,0,0,1439.392273 +0,0,1,1,2236,13,14,15777,0,0,1983.398267 +0,0,1,1,14765,57,65,62142,0,1,9307.400678 +0,0,0,0,16950,101,93,131327,1,0,15252.16749 +0,0,1,1,19100,134,97,80943,0,0,5675.826269 +0,0,1,1,110805,38,25,478518,0,1,41818.39213 +0,1,0,1,2410,27,23,19804,0,0,3395.495012 +1,0,0,1,43058,83,94,175626,1,1,32215.16098 +1,0,0,1,10751,17,10,37642,0,0,5915.114035 +0,0,1,0,15375,25,32,58423,1,0,9731.674473 +1,0,0,1,9136,39,24,77650,1,1,18608.01124 +0,0,0,1,45636,190,214,138040,0,1,21710.83319 +0,0,0,0,17132,22,13,56766,1,0,7688.134548 +0,0,1,0,4451,19,22,15751,0,0,2342.172593 +0,1,0,1,21171,24,18,68592,1,1,16261.05254 +0,0,0,1,53513,35,21,171389,1,0,15341.2549 +0,0,0,1,65868,20,17,202041,1,1,27864.98293 +0,0,0,1,12421,100,81,94980,1,1,17726.35317 +0,0,1,1,44950,183,233,120480,0,1,19661.73817 +0,0,1,0,23056,34,29,191204,1,0,17610.66107 +0,0,0,1,12015,104,131,74728,0,1,12023.65435 +0,1,0,1,39637,17,21,309604,1,1,37891.172 +0,1,0,0,32529,49,37,90631,0,0,6647.296616 +0,0,0,1,77350,15,12,303470,1,1,35467.78765 +0,0,1,1,61039,66,43,196930,1,1,28597.51471 +0,0,0,1,16226,90,72,54075,0,0,7220.87191 +0,0,1,0,12813,158,188,97736,1,1,22883.8854 +0,0,1,1,52575,153,155,143587,0,1,21617.33461 +1,0,0,1,9349,34,20,39847,0,1,9099.395522 +0,1,0,1,32920,10,11,192001,1,1,27934.92302 +0,0,1,1,7853,94,92,52369,1,1,14899.93712 +0,0,0,1,2740,44,55,10867,1,0,9150.068397 +0,0,0,0,115796,13,8,418788,1,1,46668.59688 +0,0,1,1,40637,85,100,243944,1,1,33519.77889 +0,0,1,1,2964,43,42,26090,0,1,6237.467069 +0,0,1,1,29107,13,12,90923,1,0,12351.10675 +0,1,0,1,8140,16,11,70111,0,0,6673.257513 +0,0,0,0,18938,35,29,54633,0,1,5766.333262 +0,0,1,1,1755,51,65,12679,1,0,8678.13295 +0,0,0,1,86106,52,35,267114,0,0,8834.50154 +0,0,1,0,94521,72,92,301844,0,1,28800.33804 +0,1,0,1,5344,147,137,17498,0,0,9284.071213 +1,0,1,1,11215,33,22,37339,0,1,9379.400485 +0,0,0,0,8397,76,86,78647,1,1,16291.33259 +0,0,1,1,51343,32,36,159946,0,0,7940.947136 +0,1,0,0,13017,118,84,54033,1,1,15732.86598 +0,0,1,0,13512,16,16,34524,0,0,1657.60029 +0,0,0,0,5153,86,53,24055,0,0,1768.879168 +1,1,0,1,32540,14,12,200125,0,1,21874.94337 +0,0,1,1,87157,33,36,236683,1,0,20887.12542 +0,1,0,1,95183,29,33,281831,0,0,15365.64843 +0,0,1,0,35818,55,70,286403,1,1,35776.06122 +0,0,0,1,40342,53,60,124441,0,1,14019.02331 +0,0,1,1,7565,17,17,21557,0,0,2572.29145 +0,0,1,1,39299,30,25,121658,1,0,13440.23615 +0,1,0,0,3695,52,57,27252,0,0,6427.628148 +0,0,1,1,34997,76,82,124451,0,0,6864.032046 +1,0,1,1,2228,10,10,22187,0,0,6828.300545 +0,0,1,0,34935,37,24,312042,1,1,37676.723 +0,0,1,0,11982,52,40,71221,0,1,8315.512754 +1,0,1,1,15945,114,109,52711,0,1,14559.37835 +0,0,0,1,22829,19,21,122986,0,0,5520.08532 +0,0,0,1,41241,53,61,218244,0,0,10153.0231 +0,0,0,1,93644,18,20,369755,1,1,45167.61109 +0,0,1,1,26096,31,40,112556,0,0,7420.574514 +0,0,1,1,33283,37,47,113427,0,0,8036.636786 +0,0,0,0,12263,64,53,76953,0,1,9011.419677 +0,0,1,1,44908,38,49,171011,1,1,25780.19642 +0,0,1,1,63404,35,25,186296,1,1,26795.14017 +0,1,0,1,29286,40,37,80408,1,1,17264.45423 +0,0,1,0,50350,17,17,170110,1,1,22058.75821 +1,1,0,1,18843,35,41,112058,1,0,18925.58012 +0,0,1,0,41753,21,22,174845,1,0,16353.77671 +0,0,1,1,12170,26,16,34745,1,0,9797.931115 +0,0,1,1,8425,21,24,69801,1,0,11471.60871 +0,0,1,1,89351,42,49,235865,1,0,20195.00107 +0,0,1,1,40103,34,41,262309,1,0,19585.5975 +0,0,1,1,3153,57,43,10245,0,0,4387.135137 +0,1,0,1,11462,204,204,35991,0,0,13619.82974 +0,0,1,1,10851,72,44,28921,1,1,12724.2847 +0,1,0,1,3446,77,52,31509,0,0,5947.278689 +0,0,0,1,5816,25,31,25646,1,0,9107.229617 +0,0,1,1,3757,72,62,21188,0,0,6550.153588 +1,0,1,1,12841,158,149,86687,0,0,14162.78827 +0,1,0,0,26798,30,24,98948,0,0,7180.914803 +0,0,1,1,66632,74,76,193061,0,1,20889.74858 +1,0,0,1,7182,27,23,23792,1,1,13954.74776 +0,0,0,1,4857,240,196,27719,1,1,17028.09659 +1,0,1,1,26045,92,95,72144,1,0,17047.86739 +0,1,0,0,13861,35,23,34800,1,0,9965.380853 +1,0,1,0,8576,84,94,27701,1,0,16472.20159 +0,0,0,0,11033,26,30,60672,0,1,5748.774662 +1,0,1,0,35919,24,29,97257,1,1,21762.75624 +0,0,1,0,3726,173,106,31027,0,1,7139.476089 +0,0,1,1,14114,29,30,56959,0,1,7991.801988 +0,0,0,1,4919,51,44,38379,0,0,3672.918732 +0,0,0,1,30702,153,145,95179,1,1,21795.06614 +0,0,0,1,16510,184,219,46455,0,1,13878.99218 +0,0,1,1,7157,16,14,69576,1,1,14189.7407 +0,0,1,1,2284,122,158,12521,0,1,10381.90741 +0,0,1,0,47012,38,39,143772,0,0,6887.731045 +0,0,1,0,73855,141,139,188319,1,1,30681.06136 +0,1,0,1,9381,83,85,42495,1,0,13716.10988 +0,0,0,1,38673,28,30,142276,1,1,21574.47709 +0,0,0,0,15243,26,32,38690,1,1,10110.00069 +0,1,0,1,8698,22,23,46078,1,1,14498.91261 +0,0,1,1,7373,33,40,24073,0,0,2680.364289 +0,0,1,1,4251,33,29,38335,0,0,5480.749596 +0,1,0,1,34672,129,160,126165,1,0,19823.71177 +1,0,0,1,103613,74,52,259317,0,1,29737.78514 +0,0,0,1,23857,30,23,73164,1,1,14328.8149 +0,0,0,1,8915,38,27,59774,0,0,5781.273289 +0,0,1,1,19374,39,34,66436,1,0,10420.16974 +0,0,1,1,18630,266,325,100278,0,1,23166.06052 +0,0,1,1,4127,32,31,15569,0,1,2921.20266 +0,1,0,1,8279,38,30,52969,0,1,10113.77289 +1,0,1,1,6508,33,39,38371,0,1,10330.24255 +0,1,0,0,43287,26,33,207662,1,0,19666.70339 +0,0,0,0,65671,13,14,171025,1,0,15897.04316 +0,0,1,0,62591,57,59,205256,1,1,27906.4413 +0,1,0,1,21605,70,52,59139,0,0,5977.569192 +0,0,1,1,4809,21,26,34871,0,1,5239.444732 +0,0,1,1,75639,30,19,210431,1,0,19023.10679 +0,1,0,1,20232,29,27,92846,0,0,6904.359815 +0,1,0,1,3315,67,66,11681,0,0,5434.443462 +0,1,0,0,42932,23,15,125437,1,0,15123.5866 +0,0,0,0,1640,15,11,11267,0,0,-616.5724513 +0,0,0,1,56765,54,53,160784,1,1,23112.51345 +0,0,1,1,4644,56,49,45005,0,0,5500.479797 +0,0,0,1,6217,56,41,36316,1,0,7328.944312 +0,1,0,1,10443,60,68,99629,1,1,19656.17731 +0,0,1,1,17254,49,62,46979,0,1,8632.529171 +0,0,1,1,10801,80,98,32666,1,0,12627.10734 +0,0,1,1,19955,30,26,60593,0,1,7854.669318 +1,0,0,1,7690,20,20,22101,1,0,13912.19876 +0,0,0,0,35600,66,76,171952,0,1,17500.87214 +1,0,0,0,63483,142,92,204411,1,1,33250.97377 +1,1,0,1,17853,94,94,84566,0,1,17705.53082 +0,0,1,0,4721,24,14,17807,1,0,6322.44198 +0,0,1,0,11592,59,65,77087,0,0,4565.842911 +0,1,0,1,133052,36,40,399228,1,0,28104.52976 +0,0,0,0,49286,22,14,150010,1,1,20132.08548 +1,0,1,1,29684,21,15,160862,1,1,26783.82687 +0,0,1,1,8922,92,65,62289,1,1,15612.3162 +0,0,1,1,64836,112,127,197133,1,1,31799.31029 +1,1,0,1,46758,91,96,170928,1,1,33713.93212 +0,0,1,1,50599,107,67,229859,1,1,33989.91842 +0,0,0,1,5394,13,12,30169,1,0,7777.676957 +0,0,1,0,2115,16,17,11709,0,1,2117.546395 +0,0,1,0,19769,197,231,55240,1,0,17419.95201 +0,0,0,0,54806,123,145,283472,0,1,28592.42148 +0,0,1,0,59798,178,203,558513,1,1,70110.41759 +1,1,0,0,24172,70,72,81282,1,1,22154.12055 +0,0,1,1,10226,10,7,92253,0,1,8899.070061 +0,1,0,1,3121,16,16,14226,1,1,11642.06011 +0,0,0,1,4568,21,14,14755,1,0,5933.833852 +0,1,0,1,15788,42,44,97815,0,1,13636.95272 +0,1,0,1,13014,45,37,93611,0,1,12127.64035 +0,0,1,1,6018,27,20,19427,1,0,8389.650104 +0,0,0,1,84149,71,88,310718,1,1,38438.5399 +0,0,1,0,32655,36,44,135278,1,1,21340.90633 +1,0,1,0,18844,70,90,55184,0,1,13191.2218 +0,1,0,1,58594,55,62,490270,1,1,58714.84971 +0,0,0,1,10630,23,27,55771,0,0,3980.851169 +0,0,0,1,8220,19,15,72357,1,0,10492.48497 +0,0,1,1,9317,211,254,39983,0,0,14542.64742 +1,0,1,0,28734,48,38,269252,1,1,37507.32502 +1,0,0,0,12037,42,42,61609,0,0,6937.60052 +0,0,0,1,49427,28,33,288358,1,1,35979.81746 +1,0,0,1,5752,37,46,54873,0,0,6987.138356 +1,0,0,1,7096,91,86,65357,0,0,11234.4588 +0,1,0,1,13359,18,12,77553,1,0,11368.25645 +1,0,0,0,21722,59,48,136034,0,1,15218.78818 +0,0,1,0,24113,27,27,160121,1,0,13816.97789 +0,0,1,1,17293,29,37,80727,0,0,7133.681372 +0,0,0,0,12688,11,10,42969,0,1,2549.01298 +0,0,1,0,28207,125,95,76872,1,1,18576.01452 +0,0,1,1,71943,18,14,248566,0,1,20998.0432 +1,0,0,0,84320,54,65,225946,0,1,23284.55072 +0,0,1,1,22042,14,14,68408,1,0,9544.957874 +0,1,0,0,18815,80,81,61704,0,0,7972.853529 +0,0,1,0,10897,24,28,74510,0,1,7692.030915 +0,0,1,0,8332,21,13,48657,1,0,7872.254385 +1,0,1,0,60860,227,214,165733,1,0,26875.78401 +0,0,0,0,23524,92,111,216948,1,1,30042.24231 +0,0,1,0,1501,22,18,12243,0,0,3118.114431 +0,0,1,1,12010,31,30,42780,0,1,7446.547831 +0,0,0,1,59331,323,338,260230,1,1,46530.42113 +0,0,1,0,31294,17,14,100087,1,0,9617.575449 +0,0,0,1,43525,79,101,201749,1,1,32638.07335 +1,0,1,0,21105,34,21,146090,1,1,25425.5671 +0,1,0,1,42867,48,56,191941,1,1,29134.66939 +0,1,0,0,12838,68,41,48972,1,0,14081.51429 +0,1,0,0,21339,20,25,98120,1,0,12868.98743 +0,1,0,1,36471,16,13,101443,1,1,18642.82558 +0,1,0,1,38793,139,165,192679,1,1,34058.60358 +0,0,0,1,112566,89,97,306079,1,1,38693.56767 +0,1,0,1,1727,65,42,14266,0,1,4852.886305 +0,0,1,1,15814,32,35,79801,0,0,6464.14138 +0,1,0,1,54198,21,22,276950,1,0,24003.60504 +1,0,1,0,6176,77,88,24603,1,1,16244.06651 +0,0,1,1,75071,71,89,369362,1,1,49636.20396 +1,1,0,1,17200,29,24,163971,1,0,21486.33935 +1,0,0,1,8656,28,22,43930,1,0,13369.71389 +0,1,0,0,1883,30,34,11621,1,1,10157.94009 +0,0,0,1,45309,39,43,137828,1,1,20308.17965 +0,0,0,1,15369,26,16,61966,0,1,7223.827695 +0,0,0,1,107765,163,155,357482,0,1,36079.39157 +0,0,1,1,28942,109,125,146101,0,0,9826.859963 +0,0,0,1,4176,76,72,24504,0,0,4875.567785 +0,0,1,1,11363,49,52,105082,0,1,11544.48893 +0,0,1,1,12749,22,26,75798,0,0,4021.204818 +1,0,1,1,37759,20,15,136223,0,1,16741.71293 +1,1,0,1,45201,74,78,158032,1,1,31342.41506 +1,0,1,1,23587,152,104,103485,0,0,13710.65014 +1,1,0,1,12588,234,271,41540,0,0,21326.34189 +0,0,0,0,76824,88,99,556966,1,1,64641.33112 +1,0,1,1,8236,14,15,26162,0,1,8053.329576 +0,0,0,0,4874,22,15,32888,0,1,3085.171282 +0,0,1,1,16975,17,13,104309,1,0,10897.16033 +0,0,1,1,27236,10,7,90537,1,0,11655.52611 +0,0,1,0,12979,14,16,43358,0,1,3744.097328 +0,0,0,1,6638,21,21,29166,0,1,3814.303434 +0,0,1,0,38213,40,28,100002,0,1,9744.287477 +1,0,1,1,52165,78,53,316043,1,1,43156.33691 +0,1,0,1,4837,74,80,45004,1,1,16722.52173 +0,1,0,1,46825,37,35,239934,1,0,20709.49016 +0,1,0,1,62909,15,19,185341,0,0,11098.19468 +0,0,0,0,34929,25,23,98476,0,1,9266.628165 +0,0,1,0,49604,18,14,216230,1,1,28216.15072 +0,0,0,1,9678,73,47,24449,0,0,3398.862946 +0,0,1,0,38592,226,198,115829,1,0,19909.6354 +0,0,0,1,60156,30,18,273062,0,1,23779.99628 +0,0,0,1,10476,24,23,47249,0,1,7022.488822 +0,0,0,1,86093,65,75,285143,1,1,38203.11784 +0,0,0,1,7414,32,33,34819,1,1,9752.831101 +0,0,1,1,45546,41,50,128179,0,0,10259.69591 +0,0,1,1,16846,84,87,49484,0,1,10842.52317 +0,0,1,1,11467,100,121,31288,0,0,7444.036432 +0,0,1,1,11169,11,9,42955,1,0,8119.356813 +0,0,0,1,14193,38,35,75949,0,0,5396.306639 +0,0,0,1,20864,54,66,61250,0,0,5077.933824 +0,1,0,0,53006,26,33,262468,1,1,34569.82493 +0,0,0,0,19280,59,59,56606,0,0,4302.232812 +0,0,1,1,15826,148,97,45361,1,1,16062.76019 +1,0,1,1,4897,55,37,12435,0,0,8105.305773 +0,0,0,1,17992,57,36,68079,0,0,3393.358658 +0,0,0,1,55622,57,73,278418,1,1,36875.60652 +0,0,1,0,43553,36,26,114094,1,1,16561.71637 +0,0,0,0,1959,43,40,14795,1,0,7434.790067 +0,0,1,1,28795,23,27,162999,0,1,16569.3359 +0,0,1,1,21700,51,56,182905,0,0,6948.191324 +0,0,1,1,15220,283,188,43339,1,0,16128.40668 +0,1,0,1,6288,12,7,35941,1,0,9764.999051 +0,0,0,1,4131,60,69,28178,0,0,4651.985014 +0,1,0,1,12338,14,16,60982,1,1,14309.01002 +0,0,1,1,9858,13,9,65645,1,1,11808.9937 +0,1,0,0,49639,34,39,236428,1,0,22720.78752 +0,0,0,1,46007,38,34,151678,0,1,15213.31743 +0,1,0,1,8458,56,63,29407,0,1,8268.107679 +1,0,1,0,31635,88,100,96045,0,0,11546.25512 +0,0,1,0,20130,35,34,180710,1,0,17637.81626 +0,0,0,1,22448,108,65,85971,1,0,12121.85281 +0,0,1,0,9400,214,225,38115,0,1,13017.22591 +1,0,1,1,125862,30,35,349699,0,1,34921.74153 +0,0,1,0,3029,75,69,13965,1,0,9034.125983 +0,0,1,0,3899,141,150,14010,1,0,10357.27798 +0,0,1,0,7801,76,50,51789,1,1,12466.48763 +0,0,1,1,8357,34,26,24866,1,0,8358.920998 +0,0,1,0,12245,16,16,70499,1,1,13222.14717 +0,1,0,1,37314,44,56,189686,0,0,13143.71507 +0,0,0,1,6231,10,9,16430,0,0,1299.102236 +0,0,1,0,17341,17,17,47528,0,0,3814.032543 +0,0,1,1,7188,53,34,26282,0,1,4400.598975 +1,0,1,1,2161,109,75,15424,0,0,7953.444885 +1,0,1,1,17351,16,12,50651,0,1,10928.05785 +1,0,1,1,50693,75,74,221486,1,1,36120.18731 +0,0,1,1,57128,150,99,158373,1,1,28399.87285 +0,0,0,1,27799,45,41,97169,1,1,18611.64529 +0,1,0,1,15352,50,60,136254,0,0,9350.540121 +0,0,0,1,1903,39,30,13700,0,1,5861.649012 +1,1,0,0,34179,21,18,143612,0,1,17312.76386 +0,0,1,1,11925,99,83,58553,0,1,9550.739313 +0,1,0,0,15724,20,17,39611,1,0,9581.700883 +0,1,0,0,73832,60,38,287082,1,1,37591.60621 +0,0,0,0,53577,16,14,152576,0,0,4783.476068 +0,0,1,1,18461,25,28,52829,1,0,9443.426405 +0,0,1,1,20166,62,70,59649,1,0,12419.66799 +0,0,1,1,94161,41,24,276626,1,0,21897.14378 +0,1,0,1,259808,58,38,766485,1,1,86006.92445 +0,0,0,1,28400,58,52,82686,1,0,10331.16381 +0,0,1,1,8755,47,43,47511,0,0,3946.520963 +0,0,1,0,3319,90,106,21859,0,0,5554.865993 +0,1,0,1,12435,37,42,108507,0,1,13941.19144 +1,0,0,1,8712,74,60,41868,1,0,14081.63434 +0,1,0,1,19731,10,6,60657,0,1,7952.221689 +0,0,1,0,5900,186,170,23567,1,1,14249.44051 +0,0,1,0,7499,13,9,21854,0,0,1880.998166 +0,0,1,0,3231,12,11,23226,0,0,869.5679452 +0,0,1,1,23285,101,125,82213,0,0,9610.946665 +1,0,0,0,11851,20,13,94601,1,1,18687.84542 +1,0,0,1,44676,44,49,140724,1,1,26882.98389 +0,0,0,1,18777,65,62,60298,1,0,10622.92178 +1,0,1,0,27383,25,29,127925,0,1,15879.43813 +0,0,0,1,60902,16,19,161836,1,0,14719.98451 +0,0,0,1,13097,30,20,47851,0,0,1680.262467 +0,0,0,1,75199,93,100,193057,1,0,21270.02348 +0,0,1,1,47956,71,87,150646,0,1,17077.46962 +0,1,0,1,30668,12,14,76685,0,1,11054.46492 +0,1,0,1,24828,32,34,66207,0,1,9464.745128 +0,0,1,0,9588,24,23,64208,0,0,1576.904992 +0,0,1,1,15014,25,18,46293,1,1,13356.46203 +0,0,0,1,55478,32,31,150647,1,1,21544.90395 +0,0,0,0,26110,61,65,88044,1,1,16860.49338 +0,0,0,0,58097,14,11,200479,0,1,15998.36376 +0,0,1,1,1981,87,105,13886,0,1,6807.419547 +1,1,0,0,19786,105,80,61173,1,1,21710.26054 +0,1,0,1,4494,15,12,41218,0,1,6891.681471 +0,1,0,0,11548,194,178,32745,1,0,15415.53276 +1,1,0,1,33526,21,26,98231,0,0,12487.33731 +0,0,1,1,6034,74,79,17817,1,0,11138.53115 +0,0,0,1,37957,30,36,109245,0,1,12192.21526 +1,0,0,1,19390,29,37,74918,0,1,12960.35675 +1,0,1,0,8551,45,53,40703,0,0,8123.101703 +0,1,0,1,13988,49,33,50219,0,0,8672.851562 +1,0,1,1,42616,107,94,122328,1,1,28184.29894 +0,1,0,1,78554,56,71,228506,1,0,23682.07355 +0,1,0,1,52583,106,88,137381,0,1,20940.13031 +1,0,1,1,42105,38,41,229669,1,1,34305.23927 +0,0,1,1,18524,65,62,137759,1,1,23509.18852 +1,0,0,1,92410,41,35,422327,0,1,40789.41791 +0,0,0,1,69328,11,10,209910,0,1,18452.77644 +1,0,1,1,5864,71,86,41264,0,0,11360.06959 +0,0,0,1,34784,26,25,221889,0,1,18096.93978 +0,0,1,1,36554,184,127,138290,1,0,20023.38286 +0,0,1,1,34041,87,56,201086,1,1,28382.05576 +0,0,0,1,4290,60,69,21262,0,1,5242.13016 +0,0,0,1,34241,84,76,130518,1,1,22971.45448 +0,0,1,0,18514,91,58,57277,0,0,5125.167803 +1,0,1,0,76835,68,63,244594,1,0,24399.42021 +0,0,0,1,7216,11,10,33670,1,0,6363.877963 +0,0,0,1,19414,59,76,137979,1,1,23419.19662 +0,1,0,0,43270,119,99,117234,1,1,23452.52164 +0,0,0,0,32414,48,52,107806,1,0,11894.77167 +0,0,1,1,7683,37,24,20041,1,0,6912.385715 +0,0,1,1,36966,10,11,156425,0,0,6046.072857 +0,0,1,0,7598,41,26,22896,0,0,3173.607208 +0,0,0,0,11723,79,87,43580,0,1,6342.437748 +1,0,0,1,16234,202,183,77772,0,1,18668.96923 +0,1,0,1,34980,23,20,130407,1,1,20730.45615 +0,0,0,1,74432,12,8,278883,1,0,20267.32063 +0,0,0,1,15286,21,13,68312,0,1,7316.069738 +0,1,0,1,1864,15,9,15110,0,0,3960.120835 +0,0,1,0,4402,18,13,14696,1,0,5994.52877 +0,0,1,1,44773,15,14,125674,1,1,19955.65869 +1,0,0,0,73881,49,43,190863,1,0,22145.54476 +1,0,0,0,54554,248,312,470588,1,1,68429.09507 +0,1,0,1,11124,70,71,43253,1,0,14168.48187 +0,0,0,1,16223,61,59,57955,0,1,9284.721324 +1,0,1,1,27346,59,55,151249,1,1,27932.65921 +0,0,1,1,30256,54,67,85065,0,0,5937.476304 +1,0,1,1,21063,62,40,65840,0,1,12962.64014 +0,0,0,1,23500,11,7,91670,1,0,13413.1494 +1,0,0,1,7939,16,11,65674,0,0,7322.936783 +0,0,0,1,73202,24,16,252970,0,0,6949.114958 +0,1,0,1,13139,14,16,73949,0,1,10114.24582 +1,0,1,1,20919,28,23,160371,0,0,11384.83831 +0,0,0,0,47645,104,76,163568,0,0,7956.801304 +0,0,1,1,41860,28,16,106287,0,1,8605.681492 +0,0,0,1,103125,55,45,318677,1,1,39303.1489 +0,0,0,1,59409,50,58,211780,0,1,19458.78897 +0,1,0,1,38481,30,34,104335,1,0,15149.97842 +1,0,1,0,30814,70,54,88879,1,1,19835.77323 +0,0,0,1,11045,47,50,38137,0,0,6500.022999 +1,0,1,1,69117,11,10,215342,1,1,33756.74093 +0,0,1,1,3348,20,18,30282,1,0,8886.536678 +0,0,1,1,2327,66,70,15085,0,0,7924.059399 +0,0,1,0,2982,68,43,16262,0,0,2613.223994 +0,0,1,1,10988,106,102,109237,0,1,13443.79982 +0,0,1,0,57973,69,63,153445,0,0,9036.567702 +0,0,1,1,14793,32,40,42943,0,1,8199.699674 +0,0,1,0,36920,204,143,96848,0,0,8235.537343 +0,0,1,1,9682,224,286,84905,0,1,21328.54713 +0,0,0,1,4440,64,47,20368,0,0,2290.876869 +0,0,0,0,2346,45,52,17580,1,0,7918.8815 +1,0,0,1,14772,63,74,54855,0,0,8930.495022 +0,0,0,1,3504,321,334,15802,1,0,21081.75578 +0,0,0,1,25058,161,139,127963,1,0,18045.75122 +1,0,1,1,105538,97,117,267677,0,1,32454.81798 +0,1,0,1,20893,70,47,112880,1,0,15249.06291 +0,0,1,1,4716,27,32,18150,0,0,4607.795027 +0,0,0,1,17504,28,30,124231,1,0,13802.46645 +1,0,1,1,22440,33,29,99953,1,0,17706.83971 +0,0,1,1,46186,74,48,141579,0,1,13930.12862 +0,0,1,0,39683,12,13,111848,0,0,4753.072214 +0,1,0,0,4195,14,17,11924,0,0,2161.745939 +1,0,0,1,10664,68,47,40037,1,1,17694.82079 +0,1,0,1,21618,30,20,152869,0,0,9412.468409 diff --git a/book/who_should_be_treated/who_should_be_treated_ko.ipynb b/book/who_should_be_treated/who_should_be_treated_ko.ipynb new file mode 100644 index 0000000..81ce860 --- /dev/null +++ b/book/who_should_be_treated/who_should_be_treated_ko.ipynb @@ -0,0 +1,2142 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "id": "1ae5a6c4", + "metadata": {}, + "source": [ + "# 누구에게 어떤 프로모션을 제공해야 할까요?\n", + "\n", + "인과추론의 많은 문제는 평균 처치 효과(ATE)에 집중합니다. 예를 들어 “프로모션을 진행할 것인가, 하지 않을 것인가?”라는 질문은 모든 고객에게 동일한 프로모션을 제공하는 정책과 아무에게도 제공하지 않는 정책을 비교하는 문제입니다.\n", + "\n", + "하지만 예산이 제한되어 있고 그 안에서 최대 효과를 내야 한다면 이야기가 달라집니다. 프로모션 효과가 고객마다 크게 다르다면, 모두에게 동일하게 제공하는 전략이 최선이 아닐 수 있습니다.\n", + "\n", + "- 이미 구매할 고객(always-converters)에게 불필요한 비용을 쓰게 되고,\n", + "- 마케팅에 부정적으로 반응하는 고객(sleeping dogs) 때문에 오히려 손실이 발생하며,\n", + "- 효과가 클 고객(high-responders)을 제대로 우선순위에 두지 못해 잠재 매출을 놓칩니다.\n", + "\n", + "따라서 핵심 질문은 다음과 같습니다.\n", + "\n", + "_\"어떤 고객에게 집중할 때 순이익이 가장 커질까?\"_\n", + "\n", + "정책학습(policy learning)은 가능한 처치 규칙들을 탐색하면서 전체 평균 순수익을 최대화하는 방법을 학습합니다. 즉, “누구를 처치해야 하는가?”라는 질문에 답하는 접근입니다." + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "id": "fc2741f9", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:46.211163Z", + "iopub.status.busy": "2026-05-09T16:49:46.211076Z", + "iopub.status.idle": "2026-05-09T16:49:48.058022Z", + "shell.execute_reply": "2026-05-09T16:49:48.057674Z" + } + }, + "outputs": [], + "source": [ + "import os\n", + "import tempfile\n", + "import warnings\n", + "from pathlib import Path\n", + "\n", + "os.environ.setdefault(\"MPLCONFIGDIR\", str(Path(tempfile.gettempdir()) / \"matplotlib\"))\n", + "os.environ.setdefault(\"MPLBACKEND\", \"Agg\")\n", + "os.environ.setdefault(\"LOKY_MAX_CPU_COUNT\", \"8\")\n", + "\n", + "import matplotlib.pyplot as plt\n", + "import numpy as np\n", + "import pandas as pd\n", + "import seaborn as sns\n", + "from scipy import stats\n", + "from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor\n", + "from sklearn.model_selection import train_test_split\n", + "\n", + "from econml.dr import DRLearner\n", + "from econml.policy import DRPolicyForest, DRPolicyTree\n", + "\n", + "warnings.filterwarnings('ignore')\n", + "\n", + "SEED = 1\n", + "rng = np.random.default_rng(SEED)\n", + "np.random.seed(SEED)\n", + "sns.set_theme(style='whitegrid')" + ] + }, + { + "cell_type": "markdown", + "id": "8b741c26", + "metadata": {}, + "source": [ + "## 정책학습 문제 정의\n", + "\n", + "한 소프트웨어 판매 회사는 할인이나 기술지원 같은 인센티브가 고객의 소프트웨어 구매를 실제로 늘리는지, 그리고 어떤 고객에게 더 효과적인지 알고 싶어합니다.\n", + "\n", + "이상적으로는 고객마다 서로 다른 인센티브를 무작위로 배정하는 실험(RCT)을 수행하는 것이 가장 좋지만 실제 비즈니스 환경에서는 비용 문제, 영업 전략, 대형 고객을 놓칠 위험 등으로 인해 이러한 실험을 수행하기 어렵습니다. 따라서 이번 데이터는 무작위 실험 데이터가 아닌 관찰 데이터(observational data)입니다.\n", + "\n", + "데이터에는 약 2,000명의 고객 정보가 포함되어 있으며, 다음과 같은 변수들로 구성됩니다.\n", + "\n", + "- 고객 특성: 각 고객의 산업 분야, 규모, 매출, 기술 프로필에 대한 세부 정보.\n", + "- 개입: 고객에게 제공된 인센티브에 대한 정보.\n", + "- 결과: 인센티브 제공 이후 1년 동안의 소프트웨어 구매 금액\n", + "\n", + "| 변수명 | 타입 | 설명 |\n", + "| --------------- | -- | -------------------------------------------------------------- |\n", + "| Global Flag | X | 고객이 글로벌 오피스(해외 지사)를 보유하고 있는지 여부 |\n", + "| Major Flag | X | 고객이 해당 산업에서 대규모 소비 고객인지 여부 (SMC 또는 SMB와 구분) |\n", + "| SMC Flag | X | 고객이 중소기업(SMC, Small Medium Corporation)인지 여부 (Major 및 SMB와 구분) |\n", + "| Commercial Flag | X | 고객의 사업 유형이 상업용(Commercial)인지 여부 (공공 부문과 구분) |\n", + "| IT Spend | X | IT 관련 구매에 사용한 금액 |\n", + "| Employee Count | X | 직원 수 |\n", + "| PC Count | X | 고객이 사용하는 PC 수 |\n", + "| Size | X | 연간 총매출 기준 고객 규모 |\n", + "| Tech Support | T | 고객이 기술 지원(Tech Support)을 받았는지 여부 (이진값) |\n", + "| Discount | T | 고객이 할인 혜택을 받았는지 여부 (이진값) |\n", + "| Revenue | Y | 고객의 소프트웨어 구매 금액 기준 매출(Revenue) |\n", + "\n", + "\n", + "`Tech Support`와 `Discount`는 모두 개입(intervention)이므로, 이를 조합해 다음과 같이 네 가지 처치로 정의합니다.\n", + "\n", + "$$\n", + "A_i \\in \\{0, 1, 2, 3\\}\n", + "$$\n", + "\n", + "각 처치는 다음을 의미합니다.\n", + "- $A_i = 0$: 아무 개입도 제공하지 않음\n", + "- $A_i = 1$: 기술지원만 제공\n", + "- $A_i = 2$: 할인만 제공\n", + "- $A_i = 3$: 기술지원과 할인을 모두 제공" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "id": "477aed7a", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:48.059813Z", + "iopub.status.busy": "2026-05-09T16:49:48.059518Z", + "iopub.status.idle": "2026-05-09T16:49:48.071045Z", + "shell.execute_reply": "2026-05-09T16:49:48.070709Z" + } + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "(2000, 13)\n" + ] + }, + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
Global FlagMajor FlagSMC FlagCommercial FlagIT SpendEmployee CountPC CountSizeTech SupportDiscountRevenuearmarm_name
010104553726261522050117688.363002discount_only
1001120842107701590380114981.435592discount_only
20001821711072649351132917.138943discount_plus_support
30000302884039775221114773.768553discount_plus_support
40010259303743914461117098.698233discount_plus_support
\n", + "
" + ], + "text/plain": [ + " Global Flag Major Flag SMC Flag Commercial Flag IT Spend \\\n", + "0 1 0 1 0 45537 \n", + "1 0 0 1 1 20842 \n", + "2 0 0 0 1 82171 \n", + "3 0 0 0 0 30288 \n", + "4 0 0 1 0 25930 \n", + "\n", + " Employee Count PC Count Size Tech Support Discount Revenue arm \\\n", + "0 26 26 152205 0 1 17688.36300 2 \n", + "1 107 70 159038 0 1 14981.43559 2 \n", + "2 10 7 264935 1 1 32917.13894 3 \n", + "3 40 39 77522 1 1 14773.76855 3 \n", + "4 37 43 91446 1 1 17098.69823 3 \n", + "\n", + " arm_name \n", + "0 discount_only \n", + "1 discount_only \n", + "2 discount_plus_support \n", + "3 discount_plus_support \n", + "4 discount_plus_support " + ] + }, + "execution_count": 3, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "data = pd.read_csv('./data/multi_attribution_sample.csv')\n", + "data.columns = data.columns.str.strip()\n", + "\n", + "covariates = [\n", + " 'Global Flag',\n", + " 'Major Flag',\n", + " 'SMC Flag',\n", + " 'Commercial Flag',\n", + " 'IT Spend',\n", + " 'Employee Count',\n", + " 'PC Count',\n", + " 'Size',\n", + "]\n", + "outcome = 'Revenue'\n", + "\n", + "ARM_NAMES = {\n", + " 0: 'none',\n", + " 1: 'tech_support_only',\n", + " 2: 'discount_only',\n", + " 3: 'discount_plus_support',\n", + "}\n", + "ARM_LABELS = [ARM_NAMES[i] for i in range(4)]\n", + "\n", + "required_columns = covariates + ['Tech Support', 'Discount', outcome]\n", + "for col in required_columns:\n", + " data[col] = pd.to_numeric(data[col], errors='coerce')\n", + "\n", + "policy_df = data[required_columns].dropna().copy()\n", + "policy_df['arm'] = (\n", + " 2 * policy_df['Discount'].astype(int)\n", + " + policy_df['Tech Support'].astype(int)\n", + ").astype(int)\n", + "policy_df['arm_name'] = policy_df['arm'].map(ARM_NAMES)\n", + "\n", + "n = len(policy_df)\n", + "X = policy_df[covariates].to_numpy()\n", + "Y = policy_df[outcome].to_numpy(dtype=float)\n", + "A = policy_df['arm'].to_numpy(dtype=int)\n", + "\n", + "print(policy_df.shape)\n", + "policy_df.head()\n" + ] + }, + { + "cell_type": "markdown", + "id": "8bfa27b6", + "metadata": {}, + "source": [ + "### 비용 설정\n", + "\n", + "실제 운영 환경에서는 인센티브 효과뿐 아니라 비용도 함께 고려해야 합니다. 또한 같은 인센티브라도 고객 특성에 따라 비용이 달라질 수 있습니다.\n", + "\n", + "예를 들어 기술지원은 직원 수가 많은 기업일수록 더 많은 지원 시간이 필요할 수 있고, 할인 역시 규모가 큰 고객일수록 더 큰 할인 폭을 요구할 수 있습니다.\n", + "\n", + "하지만 이 데이터에는 실제 비용 정보가 포함되어 있지 않기 때문에 고객 특성에 따라 달라지는 비용을 시뮬레이션해 사용합니다.\n", + "\n", + "기술지원 비용은 직원 수가 많을수록 증가하도록 설정하고, 할인 비용은 고객 규모가 클수록 증가하도록 설정합니다.\n", + "\n", + "이떄 로그정규분포(LogNormal)를 사용해 비용은 항상 양수이고, 일부 고객에서 큰 비용이 발생할 수 있는 현실적인 분포를 반영합니다.\n", + "\n", + "$$\n", + "C_i(\\text{tech}) \\sim \\text{LogNormal}(\\log C_{\\text{tech}} + 0.5 \\cdot \\tilde{e}_i,\\ 0.3)\n", + "$$\n", + "\n", + "$$\n", + "C_i(\\text{disc}) \\sim \\text{LogNormal}(\\log C_{\\text{disc}} + 0.4 \\cdot \\tilde{s}_i,\\ 0.3)\n", + "$$\n", + "\n", + "여기서 $\\tilde{e}_i$는 표준화된 직원 수, $\\tilde{s}_i$는 표준화된 회사 규모입니다. \n", + "\n", + "이렇게 정의한 고객별 비용을 반영해, 최종 순이익(net outcome)은 다음과 같습니다.\n", + "\n", + "$$\n", + "Y_i^{net} = Y_i - C_i(A_i)\n", + "$$" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "id": "5174ddd2", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:48.572223Z", + "iopub.status.busy": "2026-05-09T16:49:48.572150Z", + "iopub.status.idle": "2026-05-09T16:49:48.576620Z", + "shell.execute_reply": "2026-05-09T16:49:48.576207Z" + } + }, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
armmean_coststd_costmean_revenuemean_net_revenue
0none0.0000000.0000006585.8917926585.891792
1tech_support_only5037.8158944917.64135815104.11153410066.295640
2discount_only5181.5703343043.34317212247.9359537066.365619
3discount_plus_support12665.2832967375.49012526784.12464914118.841353
\n", + "
" + ], + "text/plain": [ + " arm mean_cost std_cost mean_revenue \\\n", + "0 none 0.000000 0.000000 6585.891792 \n", + "1 tech_support_only 5037.815894 4917.641358 15104.111534 \n", + "2 discount_only 5181.570334 3043.343172 12247.935953 \n", + "3 discount_plus_support 12665.283296 7375.490125 26784.124649 \n", + "\n", + " mean_net_revenue \n", + "0 6585.891792 \n", + "1 10066.295640 \n", + "2 7066.365619 \n", + "3 14118.841353 " + ] + }, + "execution_count": 4, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "COST_TECH_SUPPORT = 4_000.0 # baseline cost for tech support\n", + "COST_DISCOUNT = 5_000.0 # baseline cost for discount\n", + "\n", + "\n", + "rng_c = np.random.default_rng(SEED + 99)\n", + "emp_idx = covariates.index('Employee Count')\n", + "size_idx = covariates.index('Size')\n", + "\n", + "emp_z = (X[:, emp_idx] - X[:, emp_idx].mean()) / (X[:, emp_idx].std() + 1e-8)\n", + "size_z = (X[:, size_idx] - X[:, size_idx].mean()) / (X[:, size_idx].std() + 1e-8)\n", + "\n", + "# Tech support: 직원 수가 많을수록 더 많은 지원 비용 발생\n", + "c_tech = COST_TECH_SUPPORT * np.exp(0.5 * emp_z + 0.3 * rng_c.standard_normal(len(policy_df)))\n", + "c_tech = np.clip(c_tech, 200.0, 40_000.0)\n", + "\n", + "# Discount: 규모가 큰 기업일수록 더 큰 할인 비용 발생\n", + "c_disc = COST_DISCOUNT * np.exp(0.4 * size_z + 0.3 * rng_c.standard_normal(len(policy_df)))\n", + "c_disc = np.clip(c_disc, 200.0, 30_000.0)\n", + "\n", + "C_obs = np.select(\n", + " [A == 1, A == 2, A == 3],\n", + " [c_tech, c_disc, c_tech + c_disc],\n", + " default=0.0,\n", + ")\n", + "\n", + "Y_net = Y - C_obs\n", + "\n", + "pd.DataFrame({\n", + " 'arm': ARM_LABELS,\n", + " 'mean_cost': [C_obs[A == a].mean() if (A == a).sum() > 0 else 0.0 for a in range(4)],\n", + " 'std_cost': [C_obs[A == a].std() if (A == a).sum() > 0 else 0.0 for a in range(4)],\n", + " 'mean_revenue': [Y[A == a].mean() for a in range(4)],\n", + " 'mean_net_revenue': [Y_net[A == a].mean() for a in range(4)],\n", + "})" + ] + }, + { + "cell_type": "markdown", + "id": "024e5803", + "metadata": {}, + "source": [ + "### 데이터 탐색 및 진단\n", + "\n", + "정책학습에 들어가기 전에, 먼저 네 가지 처치 조합($A \\in {0,1,2,3}$)에 대해 데이터의 구조를 점검합니다.\n", + "\n", + "1. 처치별 표본 수: 각 처치에 충분한 관측치가 존재하는가?\n", + "2. 처치별 고객 특성 평균: 어떤 고객군이 특정 처치를 받는 경향이 있는가?\n", + "3. 처치별 Revenue 분포: Revenue의 규모, 분산, 이상치 분포가 처치마다 어떻게 다른가?\n", + "4. Positivity Check: 고객 특성 전반에서 각 처치가 충분히 함께 관측되는가?" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "id": "3d0f329e", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
countshare
arm
none5170.2585
tech_support_only4620.2310
discount_only4770.2385
discount_plus_support5440.2720
\n", + "
" + ], + "text/plain": [ + " count share\n", + "arm \n", + "none 517 0.2585\n", + "tech_support_only 462 0.2310\n", + "discount_only 477 0.2385\n", + "discount_plus_support 544 0.2720" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "arm_counts = policy_df['arm_name'].value_counts().reindex(ARM_LABELS).rename_axis('arm').to_frame('count')\n", + "arm_counts['share'] = arm_counts['count'] / len(policy_df)\n", + "\n", + "display(arm_counts)" + ] + }, + { + "cell_type": "markdown", + "id": "8e911f32", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:48.072480Z", + "iopub.status.busy": "2026-05-09T16:49:48.072405Z", + "iopub.status.idle": "2026-05-09T16:49:48.076539Z", + "shell.execute_reply": "2026-05-09T16:49:48.076070Z" + } + }, + "source": [ + "실제 표본 수는 `none` 517명, `tech_support_only` 462명, `discount_only` 477명, `discount_plus_support` 544명입니다. 각 처치의 비중은 23.1%에서 27.2% 사이로 비교적 고르게 분포되어 있으며, 특정 처치에 표본이 극단적으로 부족한 문제는 보이지 않습니다.\n", + "\n", + "따라서 네 가지 처치를 비교하는 정책학습을 진행하기에 표본 분포는 안정적인 편입니다." + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "id": "d6488c62", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "arm_covariate_means = (\n", + " policy_df\n", + " .groupby('arm_name')[covariates]\n", + " .mean()\n", + " .reindex(ARM_LABELS)\n", + ")\n", + "\n", + "plt.figure(figsize=(10, 5))\n", + "sns.heatmap(arm_covariate_means.T, annot=True, fmt='.2f', cmap='YlGnBu', cbar_kws={'label': 'Mean'})\n", + "plt.xlabel('Arm')\n", + "plt.ylabel('Covariate')\n", + "plt.title('Mean customer characteristics by arm')\n", + "plt.tight_layout()" + ] + }, + { + "cell_type": "markdown", + "id": "c49e9856", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:48.078246Z", + "iopub.status.busy": "2026-05-09T16:49:48.078155Z", + "iopub.status.idle": "2026-05-09T16:49:48.194453Z", + "shell.execute_reply": "2026-05-09T16:49:48.193850Z" + } + }, + "source": [ + "처치별 고객 특성은 완전히 동일하지 않습니다. `Size` 평균은 `none` 그룹의 약 70,943에서 `discount_plus_support` 그룹의 약 171,467까지 큰 차이를 보이며, `IT Spend` 역시 약 18,056에서 41,371까지 차이가 납니다. 즉, 규모가 큰 고객일수록 기술지원과 할인을 함께 제공받는 경향이 있습니다.\n", + "\n", + "이는 처치 배정이 고객 특성과 독립적이지 않음을 의미합니다. 따라서 이후 정책학습에서는 propensity model과 outcome model을 사용해 이러한 차이를 보정해야 합니다." + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "id": "c925a3d8", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
armarm_namecountmeanmedianstdminmax
00none5176585.8917926123.1870673363.163462-616.57245121445.05937
11tech_support_only46215104.11153414483.7193205400.3198584619.49136140166.67407
22discount_only47712247.93595310454.9325007472.178476889.97565341818.39213
33discount_plus_support54426784.12464923560.25289013124.9680835903.90688086006.92445
\n", + "
" + ], + "text/plain": [ + " arm arm_name count mean median \\\n", + "0 0 none 517 6585.891792 6123.187067 \n", + "1 1 tech_support_only 462 15104.111534 14483.719320 \n", + "2 2 discount_only 477 12247.935953 10454.932500 \n", + "3 3 discount_plus_support 544 26784.124649 23560.252890 \n", + "\n", + " std min max \n", + "0 3363.163462 -616.572451 21445.05937 \n", + "1 5400.319858 4619.491361 40166.67407 \n", + "2 7472.178476 889.975653 41818.39213 \n", + "3 13124.968083 5903.906880 86006.92445 " + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxAAAAGACAYAAAA9AISXAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAdwdJREFUeJzt3QmczPX/B/D3HrP3rmVd677KUZIjkrNIQiIpCaWkkyJnJEJIEQoRUdKFlKIU6fDLLR2O3JFdx6497Dm7O//H67O+85/ZnbVz7Vz7ej4e29iZ73zmO9/57vR5fz/v9+fjZzAYDEJERERERGQFf2s2IiIiIiIiYgBBREREREQ24QgEERERERFZjQEEERERERFZjQEEERERERFZjQEEERERERFZjQEEERERERFZjQEEERERERFZjQEEERGRFbjuKhFRPgYQREQuMnDgQKlfv77ZT4MGDaRZs2Zy3333yZdffsnPwsTOnTvVMcKtq40bN07uuOMO4+9btmyRsWPHeu37ISJypkCntkZERNfUqFEjeeWVV4y/5+bmSnx8vKxYsULGjBkj0dHR0qFDBx5FN3vmmWdk0KBBxt/x+RARUT4GEERELhQRESE333xzofvbt28vrVu3lnXr1jGA8AA1atRw9y4QEXkspjAREXmA4OBgCQoKEj8/P+N9eXl5smTJErnzzjvlxhtvlLvuuks+/PBD4+OLFy9W9ycnJ5u1havlN9xwgyQkJKjfz507JyNHjpSWLVtKkyZN5JFHHpGDBw8atz979qxKrdm0aZMMHz5cmjZtqradOHGipKenG7fDNgsWLDB7LfyO+03t2bNHBgwYoF4L7SD1JzExsdhj8Mknn6j3eNNNN6nnY78LctZ7+euvv9RzmzdvrrZ59NFH5ffff7eYwoTUs127dqkftP2///1P2rZtKy+++GKh/evSpYt6rWs5duyY9O/fXxo3bqw+W9PPFPuMYBKfvakJEyaoY1MUHN8pU6bI7bffrs4JvOdnn31WHQ8N3seoUaPUayCIHTx4sPF4ffvtt2rUBfffdtttsnDhQrly5Yq89NJL6hjhvtmzZ7MOhIgUBhBERC4uxM3JyTH+ZGVlyYkTJ2T8+PGSlpYm9957r3HbyZMny/z586Vnz54qWOjatau89tpr8s4776jH77nnHtXG5s2bzV7jm2++UR3cmJgY1bHs16+f/P333/Lyyy/Lm2++qTqnDz/8sBw/ftzseUitqlq1quo8Pv7447JmzRpZtGiRTe9v9+7dqjMeEhIib731luqAouONdKDMzMwin7dq1Sr1+kjfwusjOMD+mnLWe0HHeMiQIVK2bFkVAM2dO1cyMjLUdqmpqYX2DW0h9Qw/n376qQpwevXqJT/88INqS7N37145ffq0qme5lhkzZqiOOvanXbt2Mm3aNFm5cqV67P7775fz58+b1UnguKGD37t37yLPqSeffFK2b9+uAoRly5bJc889J7/99ptZuhwgsAoPD1evjWOgQdBz/fXXq/sxEjZv3jy1L/gc3377bRUYvffee2o/iIjwxUNERC4wYMAAw/XXX1/op379+oZ77rnHsGnTJuO2J06cUPe/++67Zm3MnTvX0LhxY0NiYqKxzUGDBhkfP336tGrzm2++Ub/PmTNHbX/27FnjNllZWYZOnToZhg0bpn4/c+aMes6oUaPMXmvgwIGGHj16GH/HNvPnzzfbBr/jfs2DDz6onpOTk2P2Xho2bGhYtWqVxeOSl5dnaN26teGFF14wu3/SpEmq7R07djj1vezfv19ts3fvXrPj9vrrrxvi4uLU72PHjjXcfvvtxsdxnPFj+p7Qxpo1a4z3TZw40dClSxdDUfA+8JwpU6aY3f/MM8+o95+bm6t+2rdvbxgzZozx8a+++srQoEED474VFB8fr97f7t27ze6fOnWq4cYbbzR7D02aNFHHTKMdL9Njf/HiRXVf//79zT6jZs2aGaZNm1bk+yOi0oMjEERELoTUIlwNxw+ujuOqb61atdTVeowwaHbs2KGuLCONxnTEAr9j1AJXuwGjE7jqf/HiRePoA+ostPQbXIVu2LChVKpUydiGv7+/SpNBKo6pgrUZlStXNkv7KQ6u4h84cECNIpiOtFSvXl3q1q2rrpBbghEYpFsh/cbU3Xffbfa7s97LddddJ+XKlZOnnnpKJk2aJN9//72UL19eRo8erbazRu3atVVqjzZzFkYJcHW/uNEH6Natm9nvSGPC+8dxwPvBSANGlXA84YsvvlApREXtG47HBx98oPYHKUk4zkiL2rdvn2RnZ5ttW6dOHZUqVxDSuDQ4FoCRFg1S68qUKWNxhIaISh8WURMRuRDSR5D7rkGqDoKAxx57TBVQo2MLSUlJ6rZ79+4W20GaCyDomDp1quq8Ik0IAQRy5ZF6orWDtBoELpZonVQIDQ01ewydWVvWPkhJSVEpRUuXLlU/luo8LNFqOJBSZKpChQpmvzvrveAz+Oijj1S6Do4b0pJwvJA+hlQeSx1sS5DigxStuLg4FdAhBQ2pTcXROugapJqZHoc+ffqolDUEEbfeeqsKnN54441rtvnVV1/JnDlz1L5gJi8EWto5YArv3RIEnQWFhYUV+16IqHRiAEFE5EboTOIq+PPPPy/Tp09Xef0QFRWlbpEbb6nTV6VKFXUbGRmpRhvQEUZn8+jRo2a1A3gcBbWYItYSazvLptPOmjIdocB+4ko1aiAsBT4FO/UaLXDQir41WhBVEu8FV+JRFIz388cff6iRhI8//ljNvmRaG3AtCN5Qv4C6ABSOt2nTRo0GFKdg0fulS5fMAgmM2OB94jPFMUDnvnPnzkW2h9dGoTqKpFHHoe3D66+/bhypIiJyJqYwERG5GTqiKKb9+uuvVcExtGjRQt1evnxZjVhoPygkRoGraecaV84xgxA6wAgs0PnU4N8nT55UKTem7aDDjDSqgIAAq/cTHVlt5EODNBnTx1FojFQc09dCyhCKlYtaQA0pXLGxsYUKdH/88Uez3531XvA6CLaQ9oXnIH0HBesI2izN/KSNYFi6Qo90JHxuSBuyJn0Jtm3bZvY7Ro3w/mvWrGk2uoG0LLSN1yhq9Ab279+vRn6GDRtmDB4QGGlpXQVndCIichQDCCIiD4BUGJ1Op65oo/OHqTWR2oTRBMx+g5oIBAjI00cQgU63BsEH0laQioOZmUyngsVoADqQuN24caNKh0GbyJFHR9wWHTt2VJ1dTLeKdrAvSCkyhSlWf/31VzXF6U8//SRbt25VV/SxfVGpR9hfzB6EgAEpRHg+Zv7B+zXlrPeClb/RDqY5xUxKaAejQMjvx2xDliC4QPCCbU1HENDRx5SwGP241iiBKewvPlN08DE96y+//CIjRoww+9yQhoagAaMjSGm6Fq1W4dVXX1XnyXfffaemaD18+LC635Y6FiIiazCAICLyAEipQQrKkSNHjB1nTPeJjiA67OiEIy8eV6OXL19udrU9MDBQpQwh8EDQYQpXpPF8TGmKq+woHEanFOlS6IjbAlPNotB51qxZai0BXIEvuBYCpo/FNKJYXRvbIN0I+/r+++9bXEBP06NHDzWdKkZSnn76aRVMoENcEu+lYsWKqgOPlCh04DEFKqaGxSgJRiYswVSxCPCeeOIJ+fnnn4334z0heMPxtzaFSkt7Gjp0qBrBQe2C6fS9gOAB+1KvXj2zYmZLWrVqpQIgjERg/2bOnKlGohCEAdOYiMjZ/DAVk9NbJSIiKgUw69QDDzyg0qgaNGjgtHYxqxNms8LibljwjojIk7CImoiIyEao58DP+vXr1aiLs4KH//77T03bivQmpDQVl75EROQOTGEiIiKyEYrbkZaFWbSQkuS0/yn7+6saCaSAIaXL0vSqRETuxhQmIiIiIiKyGkcgiIiIiIjIagwgiIiIiIjIagwgiIiIiIjIapyFyUNg/m7MqIt5xomIiIiIXEmv16vZ35o2bVrsthyB8BAIHjxlSQ7sR3Z2tsfsD3kHnjfE84b4XUOeiv+Pcm5flCMQHkIbeWjcuLG7d0XS09Pl0KFDagVUrDRLxPOG+H1DnoL/jyKeNyXjzz//tHpbjkAQEREREZHVGEAQEREREZHVGEAQEREREZHVGEAQEREREZHVGEAQEREREZHVOAsTEREReZ3c3FzJzMx0926Ql8jKyjLe+vv7l8rZPgMCApzWHgMIIiIi8hraPPX//vtvqewIkn3y8vIkMDBQzp07V2rPm+joaKlcubJaLM5RDCCIiIjIayQkJKirqRUrVpSIiAindIaodIxYYfQhODjYqVfivSXoxvopFy5cUL/HxsY63CYDCCIiIvKaTmBqaqqUL19eypUrV+o6guTYuQMhISGl8rwJDQ1VtwgiEHw7egxK5xgOEREReR29Xq9ucRWZiGwTFhZm9nfkCAYQRERE5FWYtkTk3r8bpjCR22WcOye5GfbNpBEQGiKhVao4fZ+IiIiIyDIGEOT24GHf08McaqPZogUMIoiIiIhchAEEuZU28lDt/j4SXLG8Tc/NunBJzq5Za/foBRERERHZjgEEeQQED0xFIiIiIvJ8LKImIiIi8jJYhfvNN9+ULl26yI033ijNmjWTwYMHy6FDh9Tj48aNk0ceeUReeeUV9Vi3bt3UVKb169eXjz/+WD3evHlzadmypUybNk21N2vWLLn11lulVatWMmHCBOPqzdYaOHCget6SJUukY8eO0rhxY+nXr5/88ccfZtv98MMP0r9/f2natKna965du8pHH31kfHznzp1qP3/77TfV5k033aTa+/zzz9U0pM8995x6bocOHWTFihVmbSclJcmkSZPktttuU6//wAMPqHbIuRhAEBEREXmZMWPGyNq1a2Xo0KGyfPlyGT9+vBw9elRefPFF42rde/bskbi4OHnnnXfU/drc/7Nnz5agoCB5++23pVevXvLhhx+qW2z7xhtvqE77mjVr1P22+u6772TLli0yceJEmTNnjly6dEmGDRtmXIdh27Zt8uyzz8oNN9wgCxculAULFkj16tXl1VdflQMHDpi1NXLkSLnjjjvk3Xffldq1a6tgaNCgQXLdddep5yKwmDFjhjFAQcCDoAmvP2LECPX+sPLykCFDZMeOHU446qRhChMRERGRF8nOzpa0tDTVScfIAmAk4cqVKzJz5kzVaYecnBzVMUcn2lS9evXU/drzcGUfawMgeAgMDJS2bduqQGDfvn027xtec9myZWqVcMB+jh07Vo2MYLTh2LFj0rt3bzVSocFoAkY9MPLQpEkT4/19+vRRoyraGgYYTUDQ8Pzzz6v7GjRoIJs3b1b7ifu//PJLOXz4sHz22WfGdtq3b68CIgQzH3zwgc3vhyxjAEFERETkRTB6gE46nD9/Xk6ePCmnTp2SH3/80RhgQHR0dKHgQeuwazAqUbZsWTUigOBBg+di1W9bITjRggeoVKmSus3IyFC3GA3QAgvs97///it//vmn2X5b2s+YmBh1axpgYL9B20+kKlWoUEG9FwQymttvv11ef/11SUlJUStRk+MYQBARERF5mV9++UVee+01OXHihISHh6ur8dpKw1oKE+63xLSDr9Ge66jQ0FCz3/3987Pl8/Ly1G1iYqJKRUIdBBY2q1mzprRo0cJsv6+1nwXbL1j/cPHiRRVAWIKRmYoVK9rxrsgjAwicVMhTwxAaoshbbrlFFcAgJ86Sy5cvq4Kfn3/+WZ183bt3V7mApifVpk2bVF7d2bNnpU6dOmr4rHXr1ja1odm7d68MGDDAWJikQa4h8giRs4c/EOw3ipKqcGEzIiIiKiG4ao86gs6dO6v6APSX0JdBITICC082atQoFfSg+BkjDBhNwegE0o4cFRkZKbVq1VKpWJb6muyf+VgRNQphVq9eLVOnTpVPPvlEfcgY4io4lKUZPny4nD59Wp188+bNk59++kkmT55sfByFMqNHj1aV/1988YUKHFBkdPz4cavbMA0ennnmGWPkbBqAIC8PQ2EoMlq6dKmKqrHfts5aQERERGStv/76S/U10LepUaOGCh5ACx4KXsn3JOhXYeYo1DwgeABczIWCfS1boZ4DheBId8IMTNrP9u3bVcqXVkROPhBAIEjA7AHo0GOKLgzBzZ07V+Lj41VhTEH79++XXbt2qanGMESF4ACFQCicQR4goDOPqByV+nXr1lWjD9h25cqVVreB3DlU9qOav2rVqoX2A0Nv6enpKqfu+uuvV4VBGI1AkGJP0RERERGRNbR6BfQ70DlG7QNmOsIMR4D+iadCsfOGDRtUnwtF04sWLVLZGwiCtDoJe913331qlAEXeHEBGReUUTyNC8VIXdLpdE57H6Wd2wMIVMujkMY0vSgqKkoaNWoku3fvLrQ9piRDgQwCA9OIEyceolpEr+jAm7YHiHS19oprQ/vjw/bvvfeeSl8qCO1j5MS0GEfL80ORDhEREVFJQN0A1oDARc+nn35apX0DMiLQl0E/x1NhligUQiPrBGlYmHJ1ypQpauYnR/cbdRxI48L6FgiunnjiCXUxGlPY4mIy+VANBEYaIDY21ux+RIraY6bwx1JwWwyBYbYADFuh847Of8FZB0zbK64NLYhZt26d+rd2a6patWrqxxQWTkFAgVoIIiIiopKCxdfwY+nC7LUcOXKk0H1bt24tdJ89a0BYeg4u4Jq+JrI6Fi9eXGi7nj17FvkcQJ/L0r4XvA/pSyguL0hbh4J8JIDQhqu0PDhNcHCwJCcnW9y+4Lba9sgHxEqKRbWn1SYU14Y98EezatUqNSdzuXLl7GoDOYueMOyofSaODiVaQ/u8srL14mfjscdztDb8PeC4lXauPG/Id/C8IVvg/9Fafj9u2SksedYeZ9QXaLUYnojnjajPEZk6+N61VG+CY2TtZ+j2AEJLAUIthGk6EL4kLM2IhG0sFVdjewxdIQjQ2iv4uNZecW3YAgcbuXXI4cMwIhYrsRcWcSk405M7YU7pkpYXlz8qFB8XJ37ZtgUQhoREdYt5pP0z2WktTecN+R6eN2Qtba0CTljiGl999ZXFSWYKQhaGNh2rJyvN501WVpaq8cUsWEWxdIHdIwMILZXowoULaiYBDX6vX79+oe2RmoQCZlMIBjD3L9KUkIaEIADPN4XftcVMimvDlg4/lo7/+uuv1e2jjz4qjkBxDxZgcTdEpvifOaZCu9Z8y86QHhIqGHysHBsrIbGFF7u5lsygYDkropa3D6tdq8T2kTzvvCHfwfOGbO0A/ffff+rfuGDoyVe8fQVmTEJdanHw/2JPXqQNF3xx/pT28yYwMFD1t7UL7qawSrjV7YibYdYlLBSCSnwtgEAdw8GDBy0WL6O+APP7YgpWFBEBZlQCFM3gpGjWrJm6r2/fvsbnoX0tMi6uDWth3Yjvv/9eFTJhHQlHYd+dtZCLM6ATWNL7k3f1yyY4SCchFk7mazEE5c+mgC8sTzpupZ0rzhvyPTxvyBqYrETr/OGW03KWPNQUaKtAezMtDas0nzcBAQHqbwjft5aCPVsCK7cHEBgqQaCADj1qB1Bcg8p5jBIg6sUHjvUVsDgI3iwq9xEgjBgxQg2poWYAsw/06tXLOMKA6bswNzIi5vbt28vatWtVatD06dPV49a0URwUVm/cuFEFEZjBCSsfarR9JSIiIiLyNW6fxhWwBsT999+vCpAfeughFSFhwQ+k9GBWJEzthc66Fh1h1WpU42ONhhdeeEEFCab5edgeFfgff/yx9O7dW80DjIp/bdpWa9ooDtKWAOtA4PVMf7R9JSIiIiLyNW4fgQAEDFg5Gj8FWZq2C0Np8+fPv2abGE3AT1GsacN0YRL8mMLid0REREREpY1HjEAQEREREZF3YABBRERERETelcJERERE5IgLl9MlJa3wGk+uEBUeJBXLcvY5Kj0YQBAREZHXBw/PzNoqWfriV0wuCcG6AFk49g6nBBFYA2vGjBlqvYsvvvhCtm7dKt5i7969ar0FVy4oN27cOHWsPvzwQ5e9JjGAICIiIi+HkQcEDw90uk4quHgk4OLldPlsy1G1D84chXjsscfk4YcfFm/Sv39/Ffx4w4rU5BiOQBAREZFPQPBQtUKE+ILw8HD1Q+SJWERNRERE5Abx8fHy9NNPS9OmTdV6VBs2bDA+tmDBArnjjjuMv69fv166d+8ujRs3lnbt2qnFcbOz/7/m448//pBHH31UtXXbbbfJK6+8IhkZGeoxLMq7YsUKueuuu9TzcYu1sjQ7d+5UqVNnz54t8r6BAweqRX9feuklNcKABXlffPFFuXLlinoc28L48eNVWpG18L569uwpN910k3q/CxcuNK4ajddGu99995307dtXbrzxRrXNp59+arGtZ599VgYNGmR234kTJ1QbR48etXqfqHgMIIiIiIhcLCcnR4YMGSKXL1+WVatWybx589QiupYcPnxYLbY7bNgw1ZnGYrlffvmlvPfee+rxM2fOqIVxK1asqDrXCD62b98uU6ZMUY/PnDlTdcyfe+45FaQgNQoBCIIKW2D78uXLy5o1a2T27NmyZcsWYxu//vqrukWAMWHCBKvbe/nll+XBBx+Ur776Sp5//nl1DLC/ppAW9dRTT8mmTZukY8eOauFfvOeCsGbXrl271CLEpgEKgqbrrrvOpvdK18YUJiIiIiIX++2339RV8e+//15q1Khh7ChbWgQXV+L9/PykatWqUqVKFfWDjnZERH661meffSbR0dEqsAgMzO/aTZs2Tfbv369GCDDagFGBe+65Rz1Wq1Yt1eaSJUtU4GGtevXqyciRI41ttGnTRr0GVKhQQd1GRkaqn+Kg2Hrp0qUyYMAAY60H2kxKSlLByfDhw43bYmSlU6dO6t8jRoyQjz76SA4cOCDVq1c3a7NDhw4qwEEw8uSTT0peXp4KtIYOHWr1eyTrcASCiIiIyMX++ecfKVOmjDF4gIYNG0pISEihbZGyhNSk+++/X3WkJ02aJImJiarDrbV1ww03GIMHuPXWW1V6FFJ49Hq9NG/e3KzNli1bSkJCgvqxVp06dcx+R6BgmkZlC+z/pUuXLO4X9hf7ralbt67ZawK2KQjvH+lQCBpgx44d6nV69Ohh1z5S0RhAEBEREbkYRhRwhbwg0yBAExwcLB988IGa1hXpPqdOnVIpPUgXKuo5plf6LdFeu6jnanUIpoKCgsRZbNkvS69b1PP79Okjx48fl7/++kuNRCDgQqBGzsUAgoiIiMjFMNqQmppqVtyLwEArSjb1008/ydtvvy2NGjVS6TgIJpDis3HjRmNq0cGDB806/UiNQsExrt7rdDq1RoOpPXv2qLQjdK7xOJi+NvalJCHVCD+W9gv7YzoyYwu8X4zWoF4CNRqoiyDnYwBBRERE5GKtWrWSJk2ayJgxY+T333+XP//8U/3b379w1wwd6nfeeUcVHaN4GFfXt23bpjrK2voLKMbGzEu4+r579255/fXXVRoT6iQwajF//nz5+uuv5fTp06qGYPXq1WqtCYyEXH/99RIWFqZqIv7991/55Zdf5P3337f5PaENvD72xRqPP/64KiDHvmC/UOCNQAn7a00dRVEwCoF2kQ6GOg1yPhZRExERkU/Aom7e8poIFN59911V7IyOPDq7KPzFqsoFYVpWzJq0fPlymTt3rtoWBcPadKmVKlVSj6H4GEXYGFXo1q2bseAZU6uWLVtWTcOKugPUTqCO4oEHHlCPI8jAc/E4ntegQQMZO3asmhbVFngfmBkKQcTixYut2h7pSStXrlQF4JUrV5YnnnhCBRaOuPvuu2Xq1KnqWAQEBDjUFlnmZygqiYxcClceAFONuVt6erocOnRIDa/iakJJunL8hBwYOVrqPvOkhFapYtNzM86dk+ML35Umc2ZLRF3zwi7y7fOGfAfPG7JFZmam6pzGxsaqTrLWObxwOV2embVVrUbtDsG6AFk49g6nrkRN9sMoTZcuXVQak1ZojvQunD8IvkprUJGZmSknT56U2rVrWyzWt6UvyhEIIiIi8mrouKMDn5Jm34xAjooKD2Lw4AGw/gMW1ENKFGau0oIHcj4GEEREROQTQQRHADwDiruLW0xu8ODBZms9OANqL5DWhcABtRRUchhAEBEREZHToD4DK0BfS1RUlNOPOGap0ha2o5LFAIKIiIiInCY8PFz9kO/iNK5ERERERGQ1BhBERERERGQ1BhBERERERGQ1BhBERERERGQ1BhBERERERGQ1zsJEREREXi8n+aLkpqe65bUDwiIlsEwFt7w2kTswgCAiIiKvDx7OLB4uhhz3rETtFxgk1Z+a75Qgon79+jJjxgz577//5IsvvpCtW7eKt9i7d68YDAZp0aKFU9o7e/asdOrUST744ANp1aqVU9r0ZefOnVPrYHTv3r3EX4sBBBEREXk1jDwgeIhu00cCo8q79LVzUi5J0va1ah+cOQrx2GOPycMPPyzepH///ir4cVYAQbYZO3asVK1alQEEERERkbUQPOhiqvjEAeNibOTJWERNRERE5Abx8fHy9NNPS9OmTaV9+/ayYcMG42MLFiyQO+64w/j7+vXr1ZXlxo0bS7t27WT69OmSnf3/KVt//PGHPProo6qt2267TV555RXJyMhQj+Xm5sqKFSvkrrvuUs/H7ccff2x87s6dO1XqFFKGirpv4MCB8sYbb8hLL72kRhiaNWsmL774oly5ckU9jm1h/PjxMm7cOKveP7YbOXKkvPrqq6q91q1by8yZM83eV8HtsR/Xuq+441QcHEeMpOA43nLLLTJs2DCVGgQ4FnifO3fuNG5f8L7i3pO2PfazR48ectNNN8kDDzyg0r9M4fGePXuqx3EeLFy4UH2Opm28++670qZNG5Xm1bt3b9m1a5dKezM9b0oKAwgiIiIiF8vJyZEhQ4bI5cuXZdWqVTJv3jxZtmyZxW0PHz4sEydOVJ3Z7777Tl577TX58ssv5b333lOPnzlzRh555BGpWLGifPrppyr42L59u0yZMkU9jg4sOqDPPfecClKQGoWONYIKW2D78uXLy5o1a2T27NmyZcsWYxu//vqrukWAMWHCBKvb3Lx5s1y4cEE++eQTmTZtmuo4Y9/sUdxxKg466E8++aQKHL766iv13hA84D3Zwpr3hM/kqaeeUh3+OnXqqJQ1fI6A13355ZflwQcfVPvx/PPPq3MDzzGF565cuVLeeustef/991XQc/fdd6vPp6SxBoKIiIjIxX777Tc5evSofP/991KjRg11H+oHevXqVWhbXHH28/NT+e1VqlRRP+hQRkREqMc/++wziY6OVh3mwMD8rh06riioxQgBRhtwZfyee+5Rj9WqVUu1uWTJEhV4WKtevXrq6rrWBq5+4zWgQoX8+o/IyEj1Y62oqCgVjISGhsr111+vOt7obI8ePVpsVdxxKg6OFQI6BGJoo3r16qpznpCQYNN+RFnxnoYOHapGIGDq1KmyY8cO9Tni+C5dulQGDBhgrIHBsU5KSlJtDh8+3NgGRkrwmWh0Op2EhIRIuXLlpKRxBIKIiIjIxf755x8pU6aMMXiAhg0bqg5gQUjFwdXl+++/X6WrTJo0SRITE1XHUmvrhhtuMAYPcOutt6r0qBMnToher5fmzZubtdmyZUvVMbalc4wr5aYQKNiSHmQJUnTQ0dbgfWJ/T548aXNbxR2n4uDzwKgQOvRIPcKV/927dxvTs5z5nlqZzCqFjv+NN96oPkfs76VLlyx+XmgDn6emZs2a4i4MIIiIiIhcDFfK8/LyCt1vGgRogoOD1VSmSFlBWsupU6dU+ouWWmPpORpMq2qJ9tpFPVfLtzcVFBQkzobOs6X9CggIsDoVzNrjZI1Ro0apqXNfeOEFdewQTPTp06fIQCnXwnGy5j0VPO5ox9/f36bPy1KwWWoCCByQ+fPnq6jx5ptvlieeeMKYA2YJhpZQtIP8NERjyO/TioQ0mzZtkm7duqkIEEOBGCa0tQ0NilpwRcCe/SDXSD9zVq4cP2HzT8bVoigiIiJXQ98iNTVVpTFp0OHVipJN/fTTT/L2229Lo0aNVOoLOslIZdm4caN6HGksBw8eNOvMIjUKxbR169ZVHdqCRbp79uxRaUe46q51eE1fG/viCn///bfZfiMlClfva9euXWhb7GfB43P69Gmrj1NxcHUfxecxMTHy0EMPqf4p6ieOHz+u6iusPU5/W/Ge/vzzT+O/EZzgORhFQo0Jfix9Xnh90xErd3J7DQSKelavXq0KQypXrqzyuzB8hCIfS5EuTgR01FFgkpKSogp10tPTZdasWepx5JAhx2zMmDEqNw+FJDiJUMCCPyJr2tDgw3vmmWcsXiGwtg0qOf5Bwer26Nx5drfRbNECCa3iG1P+ERGR90AKS5MmTVR/BZ1WXJ3G1W5chS4IHcd33nlH5fIjNSc5OVm2bdumUmO0XHh0ltHO4MGDVRrM66+/rtKY8BxcjUdnGHUSmJ0IBc/oeyHfHiMhyNMPCwtTNRG48o5OOYpybYU20NnGRdayZcta9RwsmIeLsKjFwHOxn8j/N00B0uBCM/p1KCzGe8ct0n5wwdia41Qc7PM333wjmZmZqu+IzwKjGQiykL6FqXVRG7Fy5UpjXQKK33EMbX1PqK1AoFCtWjVZvHix6lNiNiZ4/PHHZe7cuaoGA31ZzAyFwAifI9LG8L4swf7htTG7F/rUPhtAIOJavny5Gi7q2LGjug8HDKMRqGDXiktMIzhMUYVIUgsGME0WAg78EVSqVEkVnnTu3FkGDRpkXFQDz8OHjW2taQPDYQhkPvroI/VHhRPE1v2gkhdcPkaue2G45GVn2fzcrAuX5OyatZKbkVki+0ZERK6HRd285TXROcU0nCh2xgw8SEfBDEDoABaEaVlRhIs+E/pJ2LZDhw7G6VLR78Bj6Lsg8wIdXmRiaAXPmFoVnWNMw4r8enR+UR+gdVjR4cZz8Tie16BBA9V/evbZZ216T3gf2hV7dIqtgaAAxwJ1C+gco/+G2g1LMK3poUOH1DFDXw0zDqGTrhVyF3ecioNjhH7km2++qY4NRhGwfwimtEJsBGavvfaa3HvvvaoGAccWwYat7wkjHLjojFmeEEh++OGHqnhbO464iI6+K14LwQAydBBYXEu/fv3U54bjhOwba9PA7OFnKCrZygUQUfXt21e+/fZbs2EdHFR03LXpxzT4UHEwtanCtCAEBx4fdteuXVXRCU4URGkanEQISJDaVFwb+MPBiALmUsZVAXywODmOHDli9X6gDVtpQ1m4MuBuGEnBHyiGV3E1oSQhlejAyNFS95knXToSgPSl4wvflSZzZktEXfOiMPL884Z8B88bsgWuDKNzGhsbqzrJWgcpJ/minFk8XK1G7Q5+gUFS/an5Tl2JujRAfw0BEzrPJQ3BAM4fBBUl2bEeV8x7wkxRGB3BiJFpIbUr4P2jkBt9bkv1E7b0Rd06AoEhFsAXgSlEYNpjps6fP19oW0RoGJKLi4tTHX/8z6jgsI1pe8W1oU2/tW7dOvVv7daW/SAiIiLXQccdHfjc9FS3HPaAsEgGD1SquDWA0IqOC9Y6oIreUn4XtrdUF4Hts7KyVGRVVHt43Jo2rN1vR9uwBINBCIDcTftcXFEUrn1mWdl68XPg2NkKr6e9vr8HHHNf4MrzhnwHzxuyBf4fqyVO4Na0UNUvopwERpT8/PdFsTQbT2mFjA8s6HYtyPTAZ1jwcywpqC9BPey1YAE+pEE5wlDMe9LqanHr6nMGr4fXxfeupfpe7HfBeg6PDCC04ROk/5gOpeALwlLxDLaxNI0WtkfKBDrwWnsFH9faK64Na/fb0TYswfy+SAHxFK6YgSEvLn9kKD4uTvzsqGWwlyEhUd1iKM8/kx1eZ3LVzB3kW3jekLW0aSwduWBHJQupOVi87lqQ7YHMDdOLiSUJdaqm6e2WYFYqR/dl0qRJ6raodjDD0759+665TUnB3wxqR0zXkrB3ql63BhBaGhBW6DOdlgq/W1q0A6lJP/zwg9l96MijyBlpSjgR0YHH803hd62wubg2rOGMNizB7AGmKwq6CyJT/M8cRVaWAjlnSg8JFVSXVI6NlZDYkp0xwFRmULCcFVF5gGG1rVtghjznvCHfwfOGbO0AaUXGuGho7dVSci1XrYZsC/TdUCRd2s+bwMBA1efWLrqbOnbsmPXtiBuhyh9V7Tt37jQGEKhjwFzGmO6qIKy5gBkCML2YtvoeZkMCFE/jhGjWrJm6D8XZGrTfokULq9qwhjPasAT770nFp+gElvT+5F0deQoO0kmIhZO5pBiCdMYvOU865r7AFecN+R6eN2QNzGyjdf5wW5LFsORbtHSh0nzeBAQEqL8hfN9aKqK2JbBy60JyGCZBoIDO+JYtW9QiHSNGjFBRYpcuXdSHffHiReMQD2Y5QoCAbTCDE9Z8wFARpizTRhgw/zHm8MWUW5ipAdNtIS0I03xZ20ZxnNEGEREREZE3cvtK1FiQDfPkotgG07ciOlq2bJlK58GMRm3btjWuIIjICAtpYNENBARY7KR9+/YyefJkY3vYHnPmIveud+/eqnOPuYi19RqsaaM4zmiDiIiIiMgbuX0lagQMWDkaPwWhg266/oJWfIIV/a4FIwH4KYo1bWjuu+8+9eNIG0REREREvsLtIxBEREREROQ93D4CQUREROSoS2mJkpJ1xS0HMio4QsqHe9asQ0QliQEEEREReX3w8MKmKZKdW3iNJlcICgiSt+5+xaYg4ty5c7J//37p3r27w68/cOBAqVq1qsycOVNKo8uXL6vp9U1n4CxpO3fulEGDBqlJgJByX9owgCAiIiKvhpEHBA+9G3aVCi4eCbiYlihfHPpW7YMtAcTYsWNVp98ZAURphxk3z54969IAorRjAEFEREQ+AcFDbCSnUy9tDAaDu3eh1GERNREREZELIeUIC9B+8cUXcscdd0h2drbMnj1b2rVrJ02bNpUHHnhAfv31V7PnYN2pRx99VD1+2223ySuvvKJWctekpaXJ+PHj1cK5WNR23Lhxkp6ebvU+nTp1Sh5//HH1XLwG/m06E2b9+vVl3bp1Zs8xvW/BggVqOv533nlHWrVqpfYD+3PlyhWz7T/66CP1/ho3biz33HOPSgEytW3bNvU49gFT88+YMcO4HpjWBmbBvP3229XjL774ojqOOJ54zFpYK+ypp55S+4r3jGUFtFXOtc8I65S99NJL6r1g/S+8lun70SB9Cosjmz4fHnzwQZk1a5b4IgYQRERERC6EzjY6yHfffbesWbNGdbS3b9+uOqzoDON+dG7RmYYzZ86odacqVqwon376qXo+tp8yZYqxzc2bN6vH0aFHSg/W0Fq6dKnV+zRy5Ei1GO7atWvl888/VysWP/fccza9rz///FMFPsuXL1eBxO7du9VaWabwHu+991758ssvpUOHDuo19u3bpx77/vvv5emnn5aOHTuq94H3h/eBfTO1evVqFURgTS6swYXjheNZMOgqCjr66NxjQeOVK1eq/cXCxVjc2DRAWLFihZQvX159RgjwEOzgvoKwv+XKlVPvSXPy5En5/fffpU+fPuKLmMJERERE5ELR0dFqwdyQkBBJTU2Vr7/+WtavXy8NGzZUjw8ePFgOHz6sFtZF5/Szzz5Tz8FCuYGB+V23adOmqSJszU033SQjRoxQ/65Ro4a0adNG/vrrL6v36d9//1UjG6jLwL7htU6cOCF5eXkqmLB2od233npLBSIwadIkeeKJJ1Q7derUUfdhba2HH35Y/XvUqFFq5GDVqlXqCv+SJUvkzjvvlGeeeUY9Xrt2bZWe9Oyzz8qxY8ekXr166n4EIBjB0OA4Yp8rVKhg1X4iAAkLC1PBDIIIQEDSqVMnFQRo+4fX04KXWrVqqWNqesw1+Ey0oEjbd3ye2Edtn30NRyCIiIiI3OTgwYPqtn///uoquvbzzTffqDQb+Oeff+SGG24wBg9w6623qqv1GnRwTZUpU8Ys9ac4CD7ef/99ldKD0Q+MaCAtx9rgQdsHLXgABAXa/mvQvim8V+1x3GrP0bRs2bJQGzVr1hRHoK0bb7zRGDwAgg8ELKavowU9msjISJVuZglGGpAGduDAARX0fPXVVxYXIvYVHIEgIiIicnMBMGoDwsPDzR7TOu+mgUNRAgICHNoPXHXv2rWr/PTTT/Lbb7+pK/KLFi1SV9KRxlNQTk5OofswCmAqNze30L4VfC/YRnufloqhMQJS8HkYcSiJomu8lul7MA0wilOvXj1p0qSJChwQuF26dEl69OghvoojEERERERuct1116lb5ODjyrr2gxoArUAZnVOMVGgdcq1eAAXYWVlZDu9DQkKCvPrqq6LX69VVc+T7oyOMfUKKEaBjbVofcPr06ULtIO8fKVkaLd2nUaNGZnUSprANRlcARdBaPYRmz5496rZu3brXTJ2yBV4H+2E6moAOP97TtV6nOH369FEF1d9++6107txZoqKixFcxgCAiIiJyMYw2oJgXaTGYUQizKm3dulUVTKP4+d1331W1DFp6ExZLwzZIa0JxMgqlkcYUHBzs8L4g3QkF2xMnTpRDhw6pffjkk09U0IBUH7j55ptVcTUeRzCD4uWCV+gx69OYMWNUGtD//vc/FZR069ZN1VVoULS8YcMGFWxghiLM9IQCcRgyZIhKnVq4cKF6/Mcff5SpU6eq43Otjj3qGS5cuKD22xqYLQqzVo0ePVrVmmCGq+eff17Kli3r0Loc3bt3l+TkZBX49e7dW3wZU5iIiIjIJ2BRN295zX79+qnF5Hr27Kk67yg+RtExOqAIHKZPn27shKKuADMFYWSgV69eqsOPjnnB2YnshfQgBC3o0GOqWEwPi4JuFDVrQQwCBvxgilXM9oQOd3x8vFk7sbGx6nlIh0LaEqZpRaF0wfeNmYwQZKDGAoXiuIW77rpL5syZo1KnEERgZiOkAWGK1WvBMcGIDLZFAGJah2EJVo5G4TaOpzYbEwqk8bsjowYRERFq5AGjNmjPl/kZuPqGR9CG9ExnFXAXXEHAFQZ8CSCqL0lXjp+QAyNHS91nnpTQKlXEVTLOnZPjC9+VJnNmS0Rd8yIp8vzzhnwHzxuyBXLLcQUeHVV0orXc+ktpifLCpilqNWp3CAoIkrfufsWmlah9DaaWxRS0GEW5VuoQ1nVwR3Ex0r9w/qB+wtF6kWvB+hEoBNdmxPIkeP8Y2UGxuKU6Elv6ohyBICIiIq+Gjjs68ClZhRf5coWo4IhSHTyQqNoHXETD2g9IL/N1DCCIiIjI66EDz058YUhNQjrQtWC15b59+4ovwKrRpsXmGi3hBjNKobPvbO+99566uo+aDYyQ+ToGEEREREQ+CjULXbp0ueY2MTExDr/OsGHD1M+1oGC6pKGA2VJ2PoIKzFhVcKpcZ/nkk0+kNGEAQUREROSjUCuCn9JCK/q+Vg0EOY7TuBIRERERkdUYQBARERERkdUYQBARERERkdUYQBARERERkdUYQBARERERkdU4CxMRERF5vayLF0WfkuqW19ZFRUpwhQpueW0id2AAQURERF4fPOx7drjkZWW75fX9g4Ok2TvzGURQqcEAgoiIiLwaRh4QPFS7v48EVyzv0tfOunBJzq5Zq/aBoxBUWjCAICIiIp+A4CG0ShXxBvXr15fp06fL119/Lfv27ZOoqCjp16+fPPfcc8Zttm3bJgsXLpSjR4+qFZS7d+8uI0aMMC6GZk0bP/74oyxYsECOHTsmlSpVUm0888wzEhQU5Jb3Tb6BRdREREREbjBr1izp3bu3fPPNNzJgwADV0d+9e7d67Pvvv5enn35aOnbsKOvWrZMpU6bIxo0bZeTIkVa38fPPP8sLL7wgDzzwgAoyXnnlFdm0aZOMHj2anzc5hAEEERERkRv06tVL7r33Xqlevbo89dRTagQBIwmwZMkSufPOO9VoQe3ataVTp04qANiyZYsaTbCmjcWLF6vgAaMSNWrUkLZt26pA5Ntvv5WzZ8/yMye7MYWJiIiIyA3q1q1r9ntkZKTo9Xr173/++UelG5lq2bKl8bF69eoV28bBgwfljz/+kDVr1hgfNxgM6vb48eNSrVq1Enlf5PsYQBARERG5gaU6BK2Dr92aysvLU7eBgYFWtYHthwwZolKcCqrAaWfJAUxhIiIiIvIwKJDWUpE0e/bssTjqUJTrrrtOTp48KTVr1jT+xMfHy+uvvy5paWklst9UOjCAICIiIvIwGDnYvHmzmoUJQQBmU5o6darcfvvtVgcQTzzxhHz33Xfy9ttvqzZ+++03GT9+vKSmpnIEgrw7hQnDazixP//8c3VC33LLLTJp0iRVDGTJ5cuXZdq0aWpmAT8/P5UfOGbMGAkNDTVugxkGMAsBCoTq1KkjY8eOldatWzu1jYSEBHnttddk+/btaqjwtttuk3Hjxqkp0oiIiMj1sCaDr7zmXXfdJXPmzJFFixapIKJcuXLSo0cPGT58uNVtdO3aVebOnSvvvvuuKqiOjo6WO+64Q0aNGlUi+0ylh9sDCPxRrF69WmbOnCmVK1eW2bNnq6h7w4YNFvP68IeTkZEhK1askJSUFJkwYYKkp6eracxgx44danoyBARt2rRRhUNDhw6V9evXGyN2Z7SBadFycnLk/fffVwEEZjV49tlnzQqVyPOln7F/FoqA0BCvmW+ciMiX6aIi1WrQWNDNHfDa2AdbHDlypNB9W7duNfu9W7du6seRNu6++271Q+QzAUR2drYsX75cRcKY5xgQKbdr104N2yHSNrV//37ZtWuXmgdZ68i/+uqrKuDAvMi4+r906VLp3LmzDBo0SD2OkQM8b+XKlWpbZ7SBoANt4KpAw4YN1TYIMDDVWlJSkorwybP5BwWr26Nz5znUTrNFCxhEEBG5GVaAbvbOfLUatDsgeOAq1FSauDWAOHz4sCriMU0NwvzFjRo1UougFAwgUDyEWQNMc/8wpRnSkPbu3auG6lBwhFQiU61atVIBibPawAqQWBESIxLalGpffvmlmqcZ+0+eL7h8jFz3wnDJy86ye8gaV7pyMzKdvm9ERGQ7dODZiScqBQEEZgKA2NhYs/srVqxofMzU+fPnC22LNCdc8Y+Li1MjA0hFQipUUe05ow1sj5Qr1Gq0aNFCBR94fNWqVeLvb39dOlKh8NruhvQu09uSlJmZ3wHPytaLX5Z9nXm7RUaIn0TY99xsvXH//T3gM/MErjxvyHfwvCFbZGVlmU1zmpubywNIVuF5I+rvBbXH+N7VpgQueIzQp/X4AEL7H0fBWofg4GBJTk62uL2lughsjy8VrTNqqT087qw2cIAPHTokTZs2ValP+ECQeoUUpo8//lgiIuzrlGLhF7TrKU6dOlXir5EXlx+UxcfFiZ+dowHuYEhIVLeY1cI/kx1mV5835Ht43pC1tDUQtP8nE9miNJ83WVlZqn73xIkTRW5jqY/scQEEUoG0Wgjt39obNJ0RyXR7bFsQtg8LC1OdfK29go9r7TmjDczQhNEGTKmmBQuY3QBTq6GI+tFHH7XjaIjodDrjypLuhCAL/zOvVauWxc/BmdJDQgUlYJVjYyUk1nzUx5NlBgULyq+RthZWu5a7d8cjuPK8Id/B84Zsgf8X//fff+rf+P+1tVdLiXDxF+dPaT9vAgMDpUaNGsb+rqljx45Z3464kZZKdOHCBfVmNPgdC6gUhLSiH374wew+dPRRuIwUIqQhIQjA803hd216VWe0gToKdBxNRxrKlCmj7jt9+rTdxwMnNF7bU6ATWNL7k3c1cAwO0kmIhZPZUxmCdHI5MkDO5yZLaKbtU/iF6kIkNrKi+CJXnDfke3jekDWQJoz/V2qpFgEBATxwZBUt3a00nzf+/v7qB9+3phfuNbYEVm4NIBo0aKA64Tt37jQGEKhBOHjwoAwYMKDQ9lgj4o033lCddKymCJgNCZo3b67eeLNmzdR9ffv2NT4P7aNWwVltIAj55ptvjJEsoHYBa0b07NmzxI4XeY4LuVfkg3tiRA59IGJn1tm8blN8NoggIioJGKkv7WkoRPbS6my1vyNHuDWAQJ4VAgV06LFAStWqVdU6EOigd+nSRUWLiYmJEhkZqSKlJk2aqM79iBEjZPLkyepAoJC5V69extGBwYMHqylVMZNT+/btZe3ataquYPr06epxZ7SBbZctW6bWgnj++efVfW+99ZYKJu677z63HU9ynSxDjrrtXrmVVK1m3YqgmotpifLFoW8lQ88ZnIiIbIErx+gTXLp0SXWCcBGyNKejkPXQp9QCz9I2AmG4OkkPsmmQaeOM9+/2heSwqBsKOiZOnKgKmDFCgM45vhhwRb9Tp04yY8YM1THHlwRWrcaibY888ojqsGPaVSzLrmnbtq1aIRoL1KGwGTUFqE/Qpm11RhtIdcLidwh20AaGgzA6gfvwxUalR0xQlMRGcvVxIiKXfe/GxMjFixdVZwiBBJE1MOsQ+puoAXBkxkxvFh0dXWiWUZcGEJgKVbta7yhEQVj1GT8FVatWrdAqi/jimD9//jXbxAgBforijDYQTCCoICIiItfRRhyQ+uyMVAwqPRM2YPYhnDelcaIPnU7n1JEXuwIIzDZ02223qVEBrNhs7ZRPRERERM6AzpClQlAiS7R1D5B5wvPGTQEEUoqw8vKoUaNU/mH37t1VMNG4cWMn7BJ5m4xz5+xekTn9DCZDJVvEpV5wqH7Cl2eAIiIiIg8NIO699171g1SmL774QgUTWEANtQIIJDATUfny5Z2/t+SRwcO+p4c53I5/kPdM4eru4OH5ja843A5ngCIiIiK3FFGjDuKpp55SP3///bfMnDlTFRbPmTNHpTlhlWbMekS+Sxt5qHZ/HwmuWN7u4CG4fIyT98w3aSMPvRt2lQrh5Wx+PmeAIiIiIrfPwoRF1TAC8f3336s1HNq0aSMdO3aUbdu2yUMPPSRjxoyxe2Vm8h4IHkKrVHH3bniV/1Li7X4OggfO/kREREReE0BgETYEDV999ZVaUh7rNwwcOFClL2mrS2N9B9RILFq0iAEEkYmggPxJBxbsfN/hNoiIiIi8IoC46667VBU7ZmCaOnWqtG7d2uJ2derUkVOnTjm6j0Q+JSasrDzX6lHJzs22O3hAG0REREReE0C8/PLLqlC6uEXTnnnmGfVDROYYABAREZG3smspvu+++06tAGnJ4cOH5Z577nF0v4iIiIiIyJtHIFAsbTAY1L937dolu3fvlsTExELb/fjjj3LmzBnn7iVREeJTsiVDn39e2ipU5yeVo1hLQERERFQiAcTnn3+uCqexhDx+pkyZUmgbLcDo0aOHTTtBZG/wMGqdY8HqG/dVZxBBREREVBIBxMSJE6VPnz4qSHjkkUdk0qRJauE4U/7+/hIVFSXXXXedLftAZBdt5KFn42gpH2FbOc+lKzny1Z9Jdo9eEBEREZVWVve6UDDdsmVL9e8PPvhAbrjhBgkPDy/JfSOyCoIHpiIREREReVgAsX79eunQoYOULVtWzp07p36upVevXs7YPyIiIiIi8sYAYty4cfLZZ5+pAAL/vhbUSDCAICIiIiIqxQHEli1bpEKFCsZ/E/mCc0n2LeYWl+oneVlhTt8fIiIiIp8JIKpWrWrx35qcnBy5cuWKREdHO2/viEpIUED+EiiLfrlgbwsi0l6SY3Odul9EREREPrkSNYKFxYsXS82aNdWicTt37pThw4dLSkqKKrSeP3++lClTxvl7S+Qk5cID5am2FSU7N8+u5/9zLl5+PR0o2Tn8SIiIiKh0sWslagQIixYtUgEDTJs2TY08jB8/Xv7991958803nb2fRCUSRGD2Jnt+yoRw+lciIiIqnewKIL755hsZOXKkPPzww3L8+HE5evSoPP300zJo0CAZMWKEbN261fl7SkRERERE3hlAXLhwQZo0aaL+vW3bNrWAXPv27dXvlStXltTUVOfuJREREREReW8AUbFiRTl79qz6N0YbGjZsKOXKlVO/79+/XwURRERERETke+wKIHr06CEzZsyQxx9/XPbu3St9+vRR90+fPl0WLFigCquJiIiIiMj32DUL0wsvvCBhYWGye/duefHFF6V///7q/j///FMee+wxVQ9BRERERES+x64AAitNP/nkk+rH1CeffOKs/SIiIiIiIl8JIACF0jt27JD09HQxGApPadmrVy9H942IiIiIiHwhgPjll1/UwnEZGRlFjlAwgCAiIiIi8j12BRBYKK5OnTpq4bhKlSqpaVyJiIiIiMj32RVAYPG4hQsXSosWLZy/R0RERERE5LHsGjqoUqWKXLlyxfl7Q0REREREvhdAYPald955x7iYHBERERERlQ52pTBt2LBBzp8/L3feeadagTokJKRQEfUPP/zgrH0kIiIiIiJvDiAqV66sfoiIiIiIqHSxK4CYMWOG03YgLy9P3n77bfn888/V2hK33HKLTJo0SapXr25x+8uXL8u0adPk559/ViMd3bt3lzFjxkhoaKhxm02bNsmCBQtUihVmixo7dqy0bt3aqW3o9XqZP3++rF+/Xu33jTfeKBMmTJCGDRs67dgQEREREXkah+ZfxWxMH3zwgbzxxhsqpWnPnj02F1djNqfVq1fL1KlT1UrWCCiGDBki2dnZFrfH+hOnT5+WFStWyLx58+Snn36SyZMnGx/H4najR4+Wfv36yRdffKE6/UOHDlX76sw2sP26devktddek7Vr16pUrieeeEIFE0REREREvsquEQh08jFKgI4zVqHGVfy7775bBQP//vuvrFq1yqoUJwQJy5cvl1GjRknHjh3VfXPnzpV27drJ5s2bpUePHmbb79+/X3bt2iUbN26UunXrqvteffVVFXCMHDlSrUmxdOlS6dy5swwaNEg9jpEDPG/lypVqW2e0cebMGfXeFy9erPYVMKKBxfP++usvs5EKIk/0X0q8Xc8L1YVIbGRFp+8PERER+XgAgUABhdToNKPj36ZNG3U/rto/++yzKgiYNWtWse0cPnxY0tLSzDrcUVFR0qhRI9m9e3ehAAIjHBUqVDB2/KFly5YqgNm7d6907dpV9u3bJ+PGjTN7XqtWrVRA4qw2tm/fLpGRkdK+fXuz/d66davVx5DIHYICgtTtgp3v293GvG5TGEQQERGVYnYFELj6jjSgPn36SG5urvF+5P/jfqQ0WSM+Pv8qaGxsrNn9FStWND5mCmlSBbcNCgqS6OhoiYuLk5SUFElPTy80+mHanjPaOHnypKrRQECxZMkS1SaCHgQdpoGJrTCag9d2t4yMDLPba8nMzFS3Wdl68cvKElfS0tz0OTmSrXfpS0uOOu91kpuTazwG3iDcP1SeuPkhyc61/YAlZFyWr49tkcupSVImIMKh84aI5w3Zi981xPOmZGhZRSUWQFy6dKnIYmGkAKETbsuXADrwpoKDgyU5Odni9gW31bbPysoyduQstYfHndUG6jxQQ4GRGBRfY/Rh0aJF0r9/f5UaFRMTI/ZAYfahQ4fEU5w6darYbfLi8oOq+Lg48ct2bQARp8ptAiQhIUHExXFXmqr1CZHLyUny7+nTUhroszONAXRGSKpD5w0RzxtyFL9riOeN81nqIzstgKhZs6YqPL7tttsKPYb6AjxuDW39CFxJNl1LAh110xmRTLe3VFyN7cPCwlQnX2uv4ONae85oIzAwUAURSNXSRhzw7w4dOqiia9RT2EOn00m9evXE3RBk4Yu5Vq1aFj8HU+khoXIEU/vGxkpIrIun9k3MFvnjggrYKkXadSrb7UJW/hX8smWipYaV57u3C7pyUeSSSO3ataVWdDWHzhsinjdkL37XEM+bknHs2DGrt7Wr1/XII4+oImpcMb/99tvVcAeuyO/cuVMVRResHyiKlkp04cIFqVGjhvF+/F6/fv1C2yOtqOACdejoJyUlqRQjpCEhCMDzTeF3jIw4sw0EEabpSghMkNbkyOrcOI54bU+BTmBx+5N3NfALDtJJyNXgy1W0IFkXGChBOusiZmcJDAhQtwGBAYUWUvRVwfr8Y4z3e63zwprzhojnDTmK3zXE88a5rE1fsnsa1759+8oLL7ygpjHF1KXImcIMRrgK/9hjj8lDDz1kVTsNGjSQiIgIFXhokP508OBBtR5EQbgPdQgIVkxHPKB58+bqjTdr1sx4nwbtt2jRwqlt5OTkyJ9//ml8HKlPmJ3J2tEXIiIiIiJvZHfeBwKHe+65R3W0cTUesxI1adJEXcG3Jc9qwIABquga6yhUrVpVZs+era7wd+nSRRVoJyYmqrZx1RPto3M/YsQItQ4DCo4xEoLpU7XRgcGDB6s1G1DUjFmSUPCNuoLp06erx53RBgIJpG9heldM64r3jEXlAgIC5N5777X3kBIRERER+V4A8fXXX6sF3w4cOKCuwgM69+iUY+QB6yfYArM2oZ2JEyeqq/i4ur9s2TJVD4B0oE6dOqmVr++77z41OoBVq6dMmaLSqFCvgGlXx48fb2yvbdu2anE3FDhjRAQ1BVivQUs3ckYbgFWqEfg899xzar/x/rGoHgIhIiIiIiIp7QEERgNefPFF+fbbb9WV+u7du0v58uVV+hJSgjASMWzYMHUFfubMmVbvAK7aY/0I/BRUrVo1OXIEJbr/DwWzuNp/LRhNwE9RnNEGUq8wgmG6gjURERERka+zOoBYvXq1WvdgwoQJKu2oYKEFAgyMTODKPVJ87r///pLYXyIiIiIiciOri6jXr18v/fr1k4EDB1qs0sZIwsMPPywPPPCAmsqUiIiIiIhKcQCBxaNQUFycdu3ayT///OPofhERERERkTenMGHhljJlyhS7XdmyZSUtLc3R/aJSJD4lWzL0Bpufdy6p8IKAREREROQhAQSKpZGmVBx/f3+1LZG1wcOodWccOlhBAXYtZ0JERERErlwHgsgZtJGHno2jpXxEoF3BQ7lwnsZERERErmJTzwtTlmL60mu5cuWKo/tEpRCCh8pRQe7eDSIiIiJyVgCBBd6guPSk8PBwNY0rERERERGV4gDiww8/LNk9ISIiIiIij8fkcSIHJKX6yX/xGTY/LzgoQMqXY8oWEREReR8GEER20Pnnp/Jt3eUvW3edsusYjnqiLoMIIiIi8joMIIjsEBUioqtzQG4tc6OUr1jVpucmpejlx98SJCs7l8eeiIiIvA4DCCI7+QVlSXSkcBSBiIiIShWuwEVERERERFZjAEFERERERFZjAEFERERERFZjDQSRm1xIyLbreZwCloiIiNyJAQSRi+l0fur206/P2d0Gp4AlIiIid2EAQeRiZSJ18kCPWNHr89eSsAWngCUiIiJ3YwBB5KYggoiIiMgbsYiaiIiIiIisxhEIKrUu669ItiHHrucm5mU4fX+IiIiIvAEDCCq1wcOycz843E6gP/+EiIiIqHRh74dKJW3k4dao6yUqMMzm5+uvXJEr+36XqGq2P5eIiIjImzGAoFINwUNZXbjNz9P750pulu2zKBERERF5OxZRExERERGR1RhAEBERERGR1RhAEBERERGR1RhAEBERERGR1RhAEBERERGR1RhAEBERERGR1TiNK1Ep45d8SUSfZfPzgtISJSbbvpW7iYiIyHe4PYDIy8uTt99+Wz7//HNJTU2VW265RSZNmiTVq1e3uP3ly5dl2rRp8vPPP4ufn590795dxowZI6GhocZtNm3aJAsWLJCzZ89KnTp1ZOzYsdK6dWuntmHqq6++ktGjR8uWLVukWrVqTj0+RM4OHkI+e9Ou5+LMHo1/JF0UKVeDHwwREVEp5fYAYuHChbJ69WqZOXOmVK5cWWbPni1DhgyRDRs2SFBQUKHthw8fLhkZGbJixQpJSUmRCRMmSHp6usyaNUs9vmPHDtWZR0DQpk0bWbNmjQwdOlTWr18vdevWdVobmv/++09effVVlxwrIo1f0kXxCwyw+YD4J11Qt/qbbxdDRLRNz01LOCtlD+0W0WfygyAiIirF3BpAZGdny/Lly2XUqFHSsWNHdd/cuXOlXbt2snnzZunRo4fZ9vv375ddu3bJxo0bjR15dN4RcIwcOVIqVaokS5culc6dO8ugQYPU4xg5wPNWrlyptnVGG6ajJwg0brjhBhV0EJU0v/Qr6jbox08lJDDR7nYM0RXEEF7GpufkZCbb/XpERETkO9waQBw+fFjS0tLMUoOioqKkUaNGsnv37kIBxJ49e6RChQpmowAtW7ZUaUh79+6Vrl27yr59+2TcuHFmz2vVqpUKSJzVhmbx4sWi1+vlueeeYwBBrpGXX4OQU7+FZMeG29dGoM7m4IGIiIjIIwKI+Ph4dRsbG2t2f8WKFY2PmTp//nyhbZHmFB0dLXFxcSodCalISIUqqj1ntAF//PGHGj1BehPaJHIlQ2ikGMqU40EnIiKi0hVAoA4BCtY6BAcHS3JyssXtLdVFYPusrCzJzMwssj087qw2EGAg7Qo/tWrVcloAYTAYVNvupn0u2u21aMcrK1svflePj61pbKDPyZFsvbhMTk7+i+Xk5kiOn+0zC+Xk5qpbPdq5+h5cIUd/dQQCx8uFr5v/mvnvOSsr2+J5ast5Q8TzhuzF7xrieVMy0A9FRo7HBxAhISHqFh0h7d+AjrrpjEim21vqNGH7sLAw1cnX2iv4uNaeM9rADE61a9eWfv36iTMhHerQoUPiKU6dOlXsNnlx+aMy8XFx4pdtewARp1L6AyQhIUHEhbHTZUP+i2HEySB2dMSvdqATEhPFX++6yCcjBQcsXFKSk8U/yLXFzDkpyYKxu3Pnzkm6/pBD5w0RzxtyFL9riOeN81m6yO5xAYSWSnThwgWpUeP/p4XE7/Xr1y+0PdKKfvjhB7P70NFPSkpSKUZIQ0IQgOebwu8ojnZWG2vXrlUHuGnTpur33KtXo1Gz8dRTT6kfe+h0OqlXr554wtUdfDFjdMVSIGcqPSRUjuC4xsZKSKx52pdVErNF/rggMTExUinSdaejvz5Z5GJ+zU3ZQNtrCfSBgZIkIjHlyokuxnWpRJf1mHkpR6LKlJFKlWybRclRV/zzA5YqVapIzXoNHTpviHjekL34XUM8b0rGsWPHrN7WrQFEgwYNJCIiQnbu3GkMIHBF+ODBgzJgwIBC22ONiDfeeENOnz4tNWvWVPdhRiVo3ry5GnZp1qyZuq9v377G56H9Fi1aOK2NgsXUBw4cULMxLVmyRK6//nq7jwdeG8GLp0AnsLj9ybs6chQcpJOQq6M3ttACXV1goATprIt6nSHQoMu/DQiUwEDb/wwMAflTqOoCdVZH684QqMO+5qh9duXrqte+Om1scHDQNc8La84bIp435Ch+1xDPG+eyNn3J7QEEOkAIFNChL1eunFStWlWtA4FRgi5duqgr+4mJiRIZGalSj5o0aaI69yNGjJDJkyerPGwsOterVy/j6MDgwYPVmg2Yyal9+/ZqtABpQdOnT1ePO6MNLfDQaMXVuDKLEQwiIiIiIl/l7+4dwKJu999/v0ycOFEeeughCQgIkGXLlql0HsyK1LZtW7VmgxYZYdVqrPb8yCOPyAsvvKA6+AgENNj+tddek48//lh69+6tplfFdKvatK3OaIOIiIiIqLRy+0rUCBiQ/oOfgtDJP3IEGfb/D7ny8+fPv2abGE3AT1Gc0UbBNSIK7icRERERkS9yewBB5M2yk1FKbTu/oEDRleFibkREROR9GEAQ2cHvakHxxW0/2338qvW9j0EEEREReR0GEET2/OGEh0uFDu3EcHVxNVvkXLkiSQf+EEO27QvYeYILaQkiif9aXFQwPvOShCadlZDM/1/XRROqC5HYyIou2ksiIiIqKQwgiOz94wm3ff0Ibxbon/918cmfX8m5o5uK3vBs0Q/N6zaFQQQREZGXYwBBRFYJD8oPmO5rdLdklys8kpCVla1WJMeiglgrwtTFtET5+cDXknHuqGRFqeXHbeYfHCK6clX4aREREbkZAwgiskn58HJiiMxfM8VUpi5TsoPSpXJEBbVui9kXTeplGf1vosi/c+Q/B4539acXMIggIiJyMwYQRFTi/PX6/H806yzlK9m+nkpOyiVJ2r5W8rIynb9zREREZBMGEETkOhHlRBfDNCQiIiJvxgCCiGzin3RB8izcH5CVLaGpCRKQECR+BWogglISeZSJiIh8BAMIIrLy20KnboJ+/Mziw6h6iMI/9lt+zLQNIiIi8l4MIIjIKobwMpLd8QGRnKv1DAXo9XpJTEiQcjExotOZBwrJmcmy5d/d0i8imkebiIjIyzGAICKbgoii5GVnS1aGXvKiYsQQZJ7CpNf5S8rV1buJiIjIu/m7eweIiIiIiMh7MIAgIiIiIiKrMYAgIiIiIiKrMYAgIiIiIiKrsYiayA2CAgzin3ZZ/IIMNj3PLz1VRDgVKhEREbkPAwjyapf1VyTbkGPz8xL06Ii7h78+Q+pUyBX560ebnxuQGykit4kE8E+XiIiI3IO9EPLq4GHZuR8caiPQz/VTi/rn5Qc8WVXqS0BMRZuem3vFT+QPEUNoRAntHREREdG1MYAgr6WNPNwadb1EBYbZFTxEBhrXSHa5vOAw8b/GugqWGHKR8mT7iAsRERGRszCAIK+H4KGsLtzdu0E+Sp94TvKyMu1+vn9wiOjKVXHqPhEREbkTAwgiomsED2cWDXP4+FR/egGDCCIi8hkMIIi80IVkpDLlufx1g3UiFaJKz+zP2shDdJs+EhhV3ubn56RckqTtax0awSAiIvI0DCCIvEjQ1ZrvT7a7rw5i9L06twUR2Qln3ZJGhOBBF8M0JCIiIvX/RR4GclSiLlICU/IkODjL5ueeS8rmB2CDMuF+0u+2QMnOdf1hS0ozyNa/ciVL7/rX9tMFq9uLX86zuw2mERERETkHAwhySFxSliyp2VtkB4IH+64OQ1BA6UmLcUYQUdoERsVIhZ7DxaC3PUhlGhEREZFzMYAgh2Tq8/Pwu9UOkMqVy9odPJQL56lIxQcRRERE5H7stZFTxIT4SeWoIB5NIiIiIh/HvBEiIiIiIrIaAwgiIiIiIrIaU5iIyClrUOToDZKQGii5OoME6swfT870l6wsrhZORETkCxhAENnJPytNJNeO+VSz0314DQqt0LngXK+hItJJjtRLtP/1A4MkJixavJG961c4wtG1L4iIiIrCAILIzuAh8u8fHTt2/ld75D6yBkVObo6kJidLZJkyEhhg/tVyPiVTth8KlG//+Z/8dCbZ7n147tbBXhVEOGP9Ckdw7QsiIvLZACIvL0/efvtt+fzzzyU1NVVuueUWmTRpklSvXt3i9pcvX5Zp06bJzz//LH5+ftK9e3cZM2aMhIbiKme+TZs2yYIFC+Ts2bNSp04dGTt2rLRu3dqpbRw9elRmz54tBw4cEH9/f7Xf48aNkypVeNXP510deciqXE9yg8Jse2pGhqQePS7RQf9/rvnCGhT6HJEAfa6UjRTRBRbcLgQhhrSr2VIqly2c/lScpMxU+fX0LsnOyS4161c4whlrX+gTz9n9fI5+EBH5No8IIBYuXCirV6+WmTNnSuXKlVWnfMiQIbJhwwYJCio8Nejw4cMlIyNDVqxYISkpKTJhwgRJT0+XWbNmqcd37Ngho0ePVgFBmzZtZM2aNTJ06FBZv3691K1b1yltIAAZPHiwNGvWTD788EPJzs5W+4/9/uKLLyQ4OP/KI/k2BA+GYNty+3NzRHJs70P7hDIhURITVrrmbvDG9SsQPJxZNMyhNjj6QUTku9weQKDjvXz5chk1apR07NhR3Td37lxp166dbN68WXr06GG2/f79+2XXrl2yceNGYzDw6quvqo77yJEjpVKlSrJ06VLp3LmzDBo0SD2OkQM8b+XKlWpbZ7Txww8/qIDj9ddfl5AQXF0VFfjgPezbt89spIKIyJtoIw/RbfpIYFR5m57Llb+JiHyf2y8FHj58WNLS0sw63FFRUdKoUSPZvXt3oe337NkjFSpUMHb8oWXLlioNae/evSodylIHvlWrVsb2nNEGHsPIiRY8ANKYACMaRETeDsGDLqaKTT+2BhxEROR93D4CER8fr25jY2PN7q9YsaLxMVPnz58vtC3SnKKjoyUuLk513jEygFSootpzRhvVqlVTP6aWLFmiAgrUQhARERER+SK3BxCoQ4CCtQ6oIUhOTra4vaW6CGyflZUlmZmZRbaHx53VRkGog1i1apVMnDhRypUrJ/YwGAwqcPGUz0S7vRbteOhzciVb79oC15wcvXH2nxy/a00t6nwBKGRQtdS5knf139bCCJd6bk4uFk8QX5Gbk2N2awpvVd3q9ZKdXXQhdlH0qNC+mvKYmWlbQXJOdv4xxt91ro1/X/qr3wVZ2XrJtfF13cmR9+zo+7b1tW35viHiOUP24neNdf1QZON4RQChpQChY2CaDoSOqemMSKbbY9uCsH1YWJixeLngNqbtOaMN04M9b948WbRokTz99NMycOBAsZder5dDhw6Jpzh16lSx25w7F6duk1NSRHfetZ34y4b8zglGjAzi2uAlKDtNMJloelqa6LNtrIi+2jlLTk0Rv1zfCSA0llL4UtMxZW0ZSUhMkICrwYAtUvVp6hYjhFlBtnWI/dMSBWXuJ0+elNxE2zqpAcnxEoWR0rg4yUvxngDCkffs6Pu297Wt+b4h4jlDjuJ3zbVZusDukQGElkp04cIFqVGjhvF+/F6/fv1C2yOtCAXMptDRT0pKUilGSENCEIDnm8LvKI52Vhtah3/8+PHy9ddfq9tHH33UoWOh0+mkXr164glROv7AatWqZTGIM2W4gkj1jJSJipJKlVw724y/PlnkYn7NTNlA165yHJARIHJBJCw8XPKCbZzGNSBArqA7HRklAWUixVdg5AHBAz6PgEDzr5ZcXf5tTLkYqVTO9hEIXUaSSHL+90WliAo2PTfncrCkHhSpXiZIAsvZNnVujgThZaVybKwEljVPafRk2nuuXbu26CrVsvn5+vOhkvCbfe/b1te25fuGiOcM2YvfNcU7duyYWMvtAUSDBg0kIiJCdu7caQwg0Ak5ePCgDBgwoND2qC9444035PTp01KzZk11H2ZUgubNm6uhF0ytivv69u1rfB7ab9GihdPaAEzx+v3338ubb76p1pFwFF4XgYunwP/Mi9sfbbRGFxggQTrrolZnCTTk90qxaFlggQ5rSfO/ulBaQECA8d/WMlwttg8IDBBd4NWetQ9B8FDwfQUGGFR3PFCnk6Ag2+du0OUEGq+MhITYNkVyTniEpGLE59vFYq+Q8AgJtPF13UkfpFPvGaOtwXZ8p2RdHQ0ODtKJzsb3be9rW/N9Q8RzhhzF75qiWZu+5BEBBDoECBTQoUftQNWqVdV0qBgl6NKli8oxT0xMlMjISPU/pCZNmqjO/YgRI2Ty5MmqZgCLzvXq1cs4OoD1GbBmA2Zyat++vaxdu1alBk2fPl097ow21q1bp6aBRRCBGZwuXrxofE/avhKR9y/mhtWkvXEtB8hOOOvS59nTBuotkDKFUQ8ELlyEjojI87k9gNAWdcvJyVEFyCi8wwjBsmXLVEoPVoHu1KmTzJgxQ+677z4VHWHV6ilTpsgjjzyiroB37dpVpRBp2rZtK6+99pqaZhVrSiAtaPHixcZpW53RBtKWAOtA4MeUtq9E5Bm8NQCwF4IeuPjlPKe0U9KvjXoLpExpuAgdEZFn84gAAmkgWPUZPwVhqtQjR46Y3RcTEyPz58+/ZpsYTcBPURxtA4vfERH54qiLIyMvtr42ZnpCsTbqLQIykyVp+1rjQnZEROSZPCKAICIi3xl1seW1MU0sZnpCsXZAmu/VBBER+SIGEERuor+CuZhs5xcYKIHhrp11ioiIiEjDAILIxfz8sSaCSPKBA3a3UaFDBwYRRERE5BYMIIhc/UcXGiJlbmoihryryzPbIC8jU64cPyYGC6s9+7qsrHA5k5ArGXasCB0a5Cex0a6dZpiIiMhXMYAgcscfXqh90/z63rrV1klM9ZNjRzvJG0exCrVtK1Fr5g2qziCCiIjICRhAkJJx7pzkZuTPfIKpdPPi4iU9JFTyilnPIjP+vNceQf+sNJFc20cBICATS2WRq2Tn5C9uc2djndQuX8am515MzZEv9iRJRjYWsyMiIiJHMYAgFTzse3pYoSNhPnmuZfHB5USq91CFvd4WPET+/aPjDfnZvqoy2a9shL/ElmUqEhERkTt5V6+PSoQ28lDt/j4SXLG82bzswUHXnlYxMCVPZEeWBEZEeNenc3XkIatyPckNCrOvDT9/MQSFOne/fNyFZIwC5Nn8vITU/BEIIiIicj8GEGSE4CG0ShXxy8oSv+wsCYmtLCHB116JNjgYBa1nvfYoIngwBHNK1JIWlD/xlHyy3d7i7xCzdoiIiMh9GEAQUYkrE+4n/W4LlGz7Sk4kLfuK/HXhb4kOb+vsXSMiIiIbMYAgIpcFEfYKzjKI7nKmJGXZXryelGl7yhQREREVjQEEEXm8AP/8r6pfT+20a/0IkZslWc2cde2UPCIiIioeAwgi8njhujBpW6ul5ObZXkNx9nK2xMWJ6HNL6yoaREREzsUAgoi8JoiwR1gg1+wgIiJyJk5iT0REREREVuMIBBHRNSSkJ0l2TrbdxygoMEhiwqJ5jImIyGcwgCAiukbw8PaO9x0+Ps/dOphBBBER+QwGEERERdBGHtrWbCnRIZE2H6ekzFT59fQuh0YwiIiIPA0DCCKiYiB4iAkrx+PkItkJ7lnd3j84RHTlqtj1XH3iOcnLynTLaxMRuRoDCCIi8gh+uvx1Oi5+Oc9t+1D96QU2d+QRPJxZNMwtr01E5A4MIIiIyCMERsVIhZ7DxaDPcvlr56RckqTta+0aRdCeE92mjwRGlXfpaxMRuQMDCCIi8qggwlsheNDFcASBiHwfAwgiKhWOXkix+TmXM5MlKyu8RPaHiIjIWzGAICKfFqwLULdf7sTXXbqNz9aJSCdJrZctMfYthE1ERORzGEAQkU+LLRMmPVumS5Y+1+bnpqT7y29HgiTQj6MQpYU9M0C5a9YoZ8wAxdmfiMgeDCCIvJD+yhW7n+sXGCiB4eGlLoiwx8UUA0pcnb4/5JszQGltuJIzZoDi7E9EZCsGEOTVovS5EpSRKv42Xl0OyEwVb+Tnn5+Ok3zggEPtVOjQodQFEY64kIxAIs/m5yVn+ktGRhk5k5ArGZm2zywUGuQnsdFBNj+PXD8DFIIHdxSAOzIDFGd/IiJ7MYAgr+WXcUX6oTD2wh4HGvEXbxIYGiJlbmoihjzb03EgLyNTrhw/JoYcXlW3RlB+vCafbLf3eIWKSAd543i6HfUX+eYNqs4gwkU4AxQRkXUYQJDX8svL79QlVagpQaFl7GjAXwxB6OCJ1wUR9tI7dU98X5lwP+l3W6Bk2xevSVr2Ffkj/rC0q9VKrWZti4upOfLFniTJyMboB5UG3lh/QUSlEwMI8no5umDRBTMdh0ouiLBXcJZBgi+nScUof4kJd20qUlxStkPBB9OnXMdb6y+IqPRiAEFE5GMQPDz/wRmH22H6VOmov3BkFMNdszj5pyWK/vwpyQqxb0SWs08ROYYBBBGRB/svMdvu5/RuES0VIm3/mndn+lRpHTlxR/2FM0Y+3DGLU87leCnzy2JJcLAdzj5FZD8GEEREHihIl1/gv2DzBbvbqFouSGIivOdrniMn3jXy4egsTvauX6GPP65uw1r2lLCYWK/Zb2eMfHDND/IUbv8/S15enrz99tvy+eefS2pqqtxyyy0yadIkqV69usXtL1++LNOmTZOff/5Z/Pz8pHv37jJmzBgJDf3/YthNmzbJggUL5OzZs1KnTh0ZO3astG7d2uVteJNEXaQEpuRJcHCWZGdnSxyWGUjMlqBiLuSdS7L96mhBl/VXJNtg+yw3qTlpUtHhVy+d7F1HwpE1JHLS0hya/am0rV+Bjv9zXSpKtt726WO1AMTR4MGe0Q/Iyr0i0eHWVZ7j+yYxO1mCr1yU86kBXjty4q2cMfJhTwpUTmqCnP98pkOvG1i2iujsCCC8dd0NrvlBnsTtAcTChQtl9erVMnPmTKlcubLMnj1bhgwZIhs2bJAgC73X4cOHS0ZGhqxYsUJSUlJkwoQJkp6eLrNmzVKP79ixQ0aPHq06823atJE1a9bI0KFDZf369VK3bl2XteFN4pKyZEnN3iI7cBVK+x9BgMgf1l/5DArwtzt4WHbuB7ueWz47R+7DnnrZVKzevo6EPWtIIHi4+NNPdr+mI6/tzdw1euCM0Y96122R4OA0659wUdSaGZj2NjRYL7Fl7Vv8j7wrBapsh4ckICzKpudkZesl7sJFKRtZTkrTuhtc84M8iVsDCFx5Wr58uYwaNUo6duyo7ps7d660a9dONm/eLD169DDbfv/+/bJr1y7ZuHGjsSP/6quvqoBj5MiRUqlSJVm6dKl07txZBg0apB7HyAGet3LlSrWtq9rwJplXr3B2qx0glSuXFX1OjiQkJEhMTIzoAgOtCh7Khdt3KmkjD7dGXS9RgbZ1GLCAnFzcI8EBQcLrjSW/joS2hkR2UpLFkYTcnFwxpGdIbkCqSODVBRQKjHhE1K0n/nZMQ+vt61ckZdm3cKEuQCdRIRF2PTcl84roc+2cuNdPZEDbYLumr/0vKV22HwqQmyo1kesqFT8iq9fnSGJCgpSLiZFTiXo5cVwk5+oUzeTZ3FX8nZuZJYYU+17TlCPT5iJ40MVUcWnRujNe2xt5a9qW3kv32ysCiMOHD0taWppZalBUVJQ0atRIdu/eXSiA2LNnj1SoUMHYaYeWLVuqFKK9e/dK165dZd++fTJu3Diz57Vq1UoFJK5qo1u3buKNYkL8pHJUkGSjz5EuUikyUIJ0rilGRPBQVmfblWVbV58mx9aRyLFy9OLyNR4LCA+36/W9df2KAP/8r9hfT+20u43eje62OYhA8PDFwU3iDllZ+Du+WcqGlpGYsAirLiTpdZlSLjRaEoLsW2yP3McbF99z17S5zipaL01T9npr2pbeS/fbawKI+Ph4dRsba57DWLFiReNjps6fP19oW6Q5RUdHS1xcnEolQhoRUqGKas8VbdhDr9eLwWCQP/74Q1wtO0Avz/esLMH+OZLrl4XkJakSVVHEzyC54vhVnmuJCQuVYdF3SYD446KnTfzKRol/lQclwi9ADPZP1U82qFK/4TUeNR0HKuIDsfdzMohUrd9IxN/f/jbcIFqi5MGyPe0cITNIrsGgUvQyDbb9HeqCddLn5h4S4IeD5doDZjD4SXZeoAT550jileLfObYIiy4nqdkiFaJi5Pme+c+19T1DmTJ+6rssITNTks5k2PkOyBsYQivKwYv4jO35nANEOoww/8qyhZ9IXDKupiS59nUdeW1DmPi3Gy5xx86JHD8nXsMgar8NWPjVxu8yPzzDkCdxf58Q8TuR35xBZP++feLp+x1oyJXU0/Hid/aSuBL6orgY7vEBBGoIoGCtQ3BwsCQnJ1vc3lJdBLbPysqSzMzMItvD465qwx7aB2btB+dc+Vfys/LcdzrYn7Cgc+zLmLyLffXEXs2RZB53JgIhiHDHc53xfCLf5K3/v3Rwvw3eud9lcnIkJFAnroQ+qFcEECFXF4DBELb2b0An3NJsRtgG2xaE7cPCwlQHXmuv4ONae65owx5Nmza163lERERERK7k1ulrtDSgCxfMZ/rA75YKkZFWVHBbdOSTkpJUihFSiNCBv1Z7rmiDiIiIiMhXuTWAaNCggURERMjOnf9fYIgahIMHD6r1IArCfahDOH36tPE+zIYEzZs3V8MuzZo1M96nQfstWrRwWRtERERERL7KrQEE6ggGDBggb7zxhmzZskXNyjRixAh1hb9Lly6Sm5srFy9eNNYlNGnSRHXusQ2KjbFeAxad69Wrl3F0YPDgwfLNN9/I+++/L8ePH5fXX39dDh06JI888ohL2yAiIiIi8kV+Bkz940YIEubMmSPr1q1TgYK2EnW1atXUKtCdOnWSGTNmyH33YckwUesTTJkyRX755RdVr4BpV8ePH2+sXQAs+IYF6jBKUK9ePbUonOlUsa5qg4iIiIjI17g9gCAiIiIiIu/h1hQmIiIiIiLyLgwgiIiIiIjIagwgiIiIiIjIagwgiIiIiIjIagwgiIiIiIjIagwgiIiIiIjIagwgyExeXp7Mnz9f2rVrJzfffLM88cQTcubMGR4lH5WUlKTWXWnfvr1aHPGhhx6SPXv2GB//7bff1BosWDwRa51ggUVTWVlZaj0UrJHStGlTefHFFyUxMdFsG2e0QZ7r5MmT6nPDWj4aLLyJRULxHXLHHXfIBx98YPP3jDPaIM+C9ZW6desmjRs3lu7du8umTZuMj2HdpyeffFJ9D7Vt21beeusttU6UqY8++kitDXXTTTdJ//795eDBg2aPO6MN8iw5OTkyb948uf3229X3zMMPPyy///678XF+17gR1oEg0ixYsMDQqlUrw48//mg4dOiQ4bHHHjN06dLFkJWVxYPkgwYPHmzo0aOHYffu3YYTJ04YpkyZYrjpppsMx48fNxw7dszQuHFjw5w5c9S/33vvPUOjRo0M//vf/4zPHzdunKFz587q+QcOHDD06tXL8PDDDxsfd0Yb5Lmys7MN9913n+H66683rF27Vt2XmJiovkPGjx+vPvM1a9aocwC31n7POKMN8izr169Xf/urVq0ynD592rBw4UJDgwYNDPv27VPnET67oUOHGo4cOWL4/vvvDS1btjTMmzfP+Px169ap76Yvv/zScPToUcPo0aPVNgkJCepxZ7RBnmf+/PmGNm3aGH755RfDqVOnDBMmTDA0b97ccP78eX7XuBkDCDLC/3ibNm1q+Oijj4z3JScnqy/cDRs28Ej5GHwZo+O3Z88e4315eXmqM//WW28ZXn75ZcP9999v9pyRI0eqjhrEx8erDsC2bduMjyMIQZvoFIAz2iDP9eabbxoGDRpkFkAsXrzY0LZtW4NerzfbDp07a79nnNEGeQ58r9x+++2GmTNnmt2P7wF81vjMbrzxRkNSUpLxsU8++cTQrFkzY0CIz/711183Po5zo0OHDur54Iw2yPP07NnTMGPGDOPvqamp6vvmu+++43eNmzGFiYwOHz4saWlpKpVEExUVJY0aNZLdu3fzSPmYsmXLypIlS1Q6gcbPz0/9pKSkqFQm03MBbr31Vtm7dy8uPKhb7T5N7dq1pVKlSsbzxRltkGfC5/Ppp5/KzJkzze7HZ96yZUsJDAw03ofP99SpU3Lp0iWrvmec0QZ5Vprbf//9J/fcc4/Z/cuWLVMpR/i8b7jhBilTpozZ533lyhWVopKQkKA+e9PPG+dGixYtzM4ZR9sgzxMTEyM//vijSk9DOhq+c4KCgqRBgwb8rnEzBhBkFB8fr25jY2PNjkrFihWNj5HvQIerQ4cO6stY891338np06dVXjk+88qVKxc6FzIyMuTy5cty/vx5FYQEBwcXeb44ow3yPAgwx4wZIxMnTiz0fVHUZw5xcXFWfc84ow3yrAAC0tPT5fHHH1ed+L59+8rWrVvV/TxnqCgTJkwQnU6n6lZwsWvu3Lmq9qlGjRo8b9yMAQQZoVMHph1KQOcOha7k2/bt2yfjx4+XLl26SMeOHSUzM7PQuaD9np2drc6Xgo8XPF+c0QZ5nsmTJ6uCxoJXlIv6zLUAEZ+pNd8zzmiDPAdGAWDs2LHSo0cPWb58ubRp00aeeeYZNckCzxkqyrFjxyQyMlLeeecdNfqACTlGjRqlRpV43rjX/48PU6kXEhJi7Nhp/wb8Dzk0NLTUHx9f9sMPP6gvZcxe8sYbbxg7YzgXTGm/43zAOVLw8YLnizPaIM+bSQfpIhs2bLD4uKXPVOvUh4WFWfU944w2yHPgCjJg9KF3797q3w0bNlQzIL3//vs2fd4Ft7HnnCmqDfIsGG3ErHwrVqxQqWaAUQgEFQsWLOB3jZtxBIKMtHSACxcumB0V/I6cdPJNq1atkmHDhqlp8hYvXmy80ovzwdK5gP8Z44oQUkwwDWzB/yGbni/OaIM8y9q1a1U+OUapMAqBH3jllVdkyJAh6jO19JkDPlNrvmec0QZ5Du0zuf76683ur1evnspt5zlDlhw4cED0er1ZnR5gSnCk2vK8cS8GEGSEoqSIiAjZuXOnWa4zrhLdcsstPFI+aPXq1TJ16lQ1t/acOXPMUkJwxWfXrl1m2+/YsUONUvj7+0vz5s3VXPxaIbSW64y6Bu18cUYb5FkwQrVx40Y1EqH9wPDhw2X69Onqc8PnaTr/Pj5zFMejINKa7xlntEGeA8XN4eHhqkNo6p9//lG57PjM8NlpqU7a543n4LPGZ47P3vTzxvoAGAkzPWccbYM8i1YHdeTIkULnTa1atfhd427ungaKPAvm68e82D/88IPZ3OqYY5t8C6ZLveGGGwzPPvus4cKFC2Y/KSkphn/++Uc9Pnv2bDUX/7Jlywqt4YApWe+44w7Djh07jGs4DBgwwPi4M9ogz2c6jeulS5cMt9xyi2Hs2LFqrn3cjzUcMAe/td8zzmiDPMs777yjpt7FdKum60Dg7z4zM1NNH/3444+rz1JbwwFrfWg+/fRTNU0vzgFtDQesA6Kt4eCMNsiz5ObmGh566CFD165dDb/99pvh5MmThrlz5xoaNmxo+P333/ld42YMIMhMTk6Omif71ltvNdx8882GJ554wnDmzBkeJR+0aNEi1fGz9IOOG/z0009qoTnMr44v8W+++casjbS0NLWwT4sWLdQPggEsAmbKGW2Q9wQQgEDwgQceUJ855v//8MMPbf6ecUYb5FmWL1+uLhbgogLm90cn33RdGixsiUARa4BgLRp0IE1hIcr27durIKB///6GgwcPmj3ujDbIs2Bdj8mTJxs6duyoAtAHH3zQsHPnTuPj/K5xHz/8x92jIERERERE5B1YA0FERERERFZjAEFERERERFZjAEFERERERFZjAEFERERERFZjAEFERERERFZjAEFERERERFZjAEFERERERFZjAEFERERERFYLtH5TIiKiaxs4cKDs2rXL7D4/Pz8JCwuTWrVqySOPPCL33nsvDyMRkRdjAEFERE7VqFEjeeWVV4y/5+bmSnx8vKxYsULGjBkj0dHR0qFDBx51IiIvxQCCiIicKiIiQm6++eZC97dv315at24t69atYwBBROTFWANBREQuERwcLEFBQSqlCfLy8mTJkiVy5513yo033ih33XWXfPjhh8btFy9erO5PTk42awcjGTfccIMkJCSo38+dOycjR46Uli1bSpMmTVSa1MGDB43bnz17VurXry+bNm2S4cOHS9OmTdW2EydOlPT0dON22GbBggVmr4Xfcb+pPXv2yIABA9RroZ2xY8dKYmKik48WEZHnYgBBREROZTAYJCcnx/iTlZUlJ06ckPHjx0taWpqxBmLy5Mkyf/586dmzpwoWunbtKq+99pq888476vF77rlHPX/z5s1m7X/zzTfStm1biYmJUR33fv36yd9//y0vv/yyvPnmmyowefjhh+X48eNmz0NaVdWqVWXhwoXy+OOPy5o1a2TRokU2vbfdu3fLo48+KiEhIfLWW2/JSy+9pGo+Bg0aJJmZmQ4fOyIib8AUJiIicip0sjFCYAqjDtdff73MmzdPbr/9djl58qR89tlnauRg6NChahsEBdju3Xfflf79+6vO/i233CJff/219O3bV23z77//yh9//CFz585Vv69cuVKSkpLk448/VttrqVLdunVTr4UARYO6C4wWAFKptm/fLtu2bZMXX3zR6veGAKV27dpqHwMCAtR9GIno3r27rF27VgUuRES+jiMQRETkVAgecHUfP7jaj8ABMzDhij1GGWDHjh1qpOKOO+4wG63A7xix2Lt3r9oOoxMISC5evGgcfUCNBbaD3377TRo2bCiVKlUytuHv76+CiP/9739m+1WwLqNy5cpmKUzFycjIkAMHDqhAxHSUpXr16lK3bl0VkBARlQYcgSAiIqcKDw+Xxo0bG3/HFXoEAo899pgqoC5XrpwaNQBcubfk/Pnz6hYBx9SpU1X9AtKEEECgVgIpRIB2Tp8+XWjEw7TTrwkNDTV7DIEGAgFrpaSkqPSopUuXqh9LNR5ERKUBAwgiIipR5cuXl0mTJsnzzz8v06dPV2lAUVFRxhQkBBwFValSRd1GRkaq0QYEELfeeqscPXpU1Tpo8DgKmTE9rCUo2rYFppw1ZTpCgf1EihVqICwFPgUDFCIiX8UUJiIiKnEYSWjXrp2qZ0DRcYsWLdT9ly9fVqMV2g+KolG7oI1QAIquf//9d1XngMACAYMG/0Y9BeoSTNv58ssvVQqVVqdgDaRGaSMfmn379pk9jjUuUBBu+lrXXXedmq1p586dDh4lIiLvwACCiIhcAjMW6XQ6mTZtmtSrV0+lNWE04b333lM1EQgQRo8erYII1ExoEHhg8blPP/1UzcykTQMLGA1AWhFuN27cqGoi0Camg0VQYYuOHTuqFKlPPvlEtYN9QXqUKRR9//rrr6rw+qeffpKtW7fKkCFD1PZFpVEREfkaBhBEROQSderUkYEDB8qRI0dUsDBjxgwZPHiw6rCjE46pXDF70vLly81GDgIDA1XKENKLEHSYQvE0no8ZmDAt7FNPPaVmaUKqFIIKW2CaWcwQNWvWLLVeRFhYWKEZmjBT1LJly9TK2tgGqVPY1/fff9/i4nlERL7Iz2BLBRkREREREZVqHIEgIiIiIiKrMYAgIiIiIiKrMYAgIiIiIiKrMYAgIiIiIiIGEERERERE5HwcgSAiIiIiIqsxgCAiIiIiIqsxgCAiIiIiIqsxgCAiIiIiIqsxgCAiIiIiIqsxgCAiIiIiIqsxgCAiIiIiIrHW/wE4yK77NX5rMgAAAABJRU5ErkJggg==", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "arm_revenue = (\n", + " policy_df\n", + " .groupby(['arm', 'arm_name'])[outcome]\n", + " .agg(['count', 'mean', 'median', 'std', 'min', 'max'])\n", + " .reset_index()\n", + ")\n", + "display(arm_revenue)\n", + "\n", + "plt.figure(figsize=(8, 4))\n", + "\n", + "sns.histplot(\n", + " data=policy_df,\n", + " x=outcome,\n", + " hue='arm_name',\n", + " bins=40,\n", + " element='step',\n", + " stat='density',\n", + " common_norm=False\n", + ")\n", + "\n", + "plt.title('Revenue density by arm')\n", + "plt.tight_layout()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "639d2437", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:48.195732Z", + "iopub.status.busy": "2026-05-09T16:49:48.195645Z", + "iopub.status.idle": "2026-05-09T16:49:48.262584Z", + "shell.execute_reply": "2026-05-09T16:49:48.262041Z" + } + }, + "source": [ + "Revenue 평균은 `none` 약 6,586, `discount_only` 약 12,248, `tech_support_only` 약 15,104, `discount_plus_support` 약 26,784로 나타납니다.\n", + "\n", + "하지만 이 차이를 그대로 처치 효과로 해석해서는 안 됩니다. 앞서 확인했듯이 규모가 큰 고객일수록 더 강한 개입을 받는 경향이 있기 때문입니다. 즉, Revenue 차이에는 고객 특성의 영향과 실제 처치 효과가 함께 섞여 있습니다.\n", + "\n", + "따라서 단순 평균 비교만으로 정책을 판단하기보다는, 이후 AIPW 기반 정책 평가를 통해 비교해야 합니다." + ] + }, + { + "cell_type": "markdown", + "id": "61b7329f", + "metadata": {}, + "source": [ + "다음으로 positivity(overlap)를 확인합니다. \n", + "\n", + "이번 데이터는 무작위 실험이 아닌 관찰 데이터이므로, 이후 AIPW 기반 정책학습 및 평가를 위해 propensity score를 추정해야 합니다.\n", + "\n", + "이때 중요한 가정이 positivity입니다. 즉, 고객 특성 $X$의 여러 영역에서 각 처치가 모두 어느 정도 관측되어야 합니다.\n", + "\n", + "$$\n", + "e_a(x) = P(A_i = a \\mid X_i=x)\n", + "$$\n", + "\n", + "또한 propensity score가 극단적으로 치우쳐 있지 않은지도 확인합니다. 만약 특정 고객군에서 어떤 처치의 $e_a(x)$ 값이 0에 매우 가까우면, 일부 관측치에 지나치게 큰 가중치가 부여되어 추정이 불안정해질 수 있습니다." + ] + }, + { + "cell_type": "code", + "execution_count": 96, + "id": "9be31313", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
armmean_propensitymin_propensityp01_propensitypropensity_below_0_05_rate
0none0.2639410.0576000.0668310.0
1tech_support_only0.2298630.1265820.1562830.0
2discount_only0.2379340.1256380.1413220.0
3discount_plus_support0.2682630.0872990.1013040.0
\n", + "
" + ], + "text/plain": [ + " arm mean_propensity min_propensity p01_propensity \\\n", + "0 none 0.263941 0.057600 0.066831 \n", + "1 tech_support_only 0.229863 0.126582 0.156283 \n", + "2 discount_only 0.237934 0.125638 0.141322 \n", + "3 discount_plus_support 0.268263 0.087299 0.101304 \n", + "\n", + " propensity_below_0_05_rate \n", + "0 0.0 \n", + "1 0.0 \n", + "2 0.0 \n", + "3 0.0 " + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "idx = np.arange(len(policy_df))\n", + "train_idx, test_idx = train_test_split(\n", + " idx,\n", + " test_size=0.5,\n", + " stratify=A,\n", + " random_state=SEED,\n", + ")\n", + "\n", + "X_train = X[train_idx]\n", + "X_test = X[test_idx]\n", + "A_train = A[train_idx]\n", + "A_test = A[test_idx]\n", + "Y_train = Y[train_idx]\n", + "Y_test = Y[test_idx]\n", + "\n", + "multi_propensity = RandomForestClassifier(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + ")\n", + "multi_propensity.fit(X_train, A_train)\n", + "e_hat_raw = multi_propensity.predict_proba(X_test)\n", + "\n", + "propensity_summary = pd.DataFrame({\n", + " 'arm': ARM_LABELS,\n", + " 'mean_propensity': e_hat_raw.mean(axis=0),\n", + " 'min_propensity': e_hat_raw.min(axis=0),\n", + " 'p01_propensity': np.quantile(e_hat_raw, 0.01, axis=0),\n", + " 'propensity_below_0_05_rate': (e_hat_raw < 0.05).mean(axis=0),\n", + "})\n", + "display(propensity_summary)" + ] + }, + { + "cell_type": "code", + "execution_count": 97, + "id": "fl52yekx049", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "prop_df = pd.DataFrame(e_hat_raw, columns=ARM_LABELS)\n", + "prop_long = prop_df.melt(var_name='treatment', value_name='propensity')\n", + "\n", + "fig, ax = plt.subplots(figsize=(8, 4))\n", + "sns.histplot(\n", + " data=prop_long, x='propensity', hue='treatment',\n", + " bins=30, element='step', stat='density', common_norm=False, ax=ax,\n", + ")\n", + "\n", + "ax.axvline(0.05, color='tomato', lw=1.5, ls='--', alpha=0.8, label='threshold = 0.05')\n", + "ax.legend(title='Treatment', fontsize=8)\n", + "ax.set_xlabel('Propensity score')\n", + "ax.set_ylabel('Density')\n", + "ax.set_title('Propensity score distribution by treatment')\n", + "plt.tight_layout()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "0b42eca5", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:48.264007Z", + "iopub.status.busy": "2026-05-09T16:49:48.263913Z", + "iopub.status.idle": "2026-05-09T16:49:48.571081Z", + "shell.execute_reply": "2026-05-09T16:49:48.570638Z" + } + }, + "source": [ + "이 데이터에서는 $\\hat e_a(X_i)$의 최솟값이 약 0.058이며, `e_hat < 0.05`인 경우도 관측되지 않았습니다. 즉, 극단적으로 작은 propensity score로 인해 일부 관측치에 과도한 가중치가 부여되는 문제는 크지 않아 보입니다.\n", + "\n", + "또한 그래프에서도 네 가지 처치의 propensity 분포가 전반적으로 잘 겹쳐 나타나므로, 뚜렷한 positivity 위반은 관찰되지 않습니다." + ] + }, + { + "cell_type": "markdown", + "id": "f4edd2d8", + "metadata": {}, + "source": [ + "## 정책학습 방법론\n", + "\n", + "세 가지 정책학습 방법을 비교합니다.\n", + "\n", + "- Plug-in Policy (DRLearner)\n", + "\n", + " 먼저 고객별 CATE를 추정한 뒤, 기대 순이익(net benefit)이 가장 큰 처치를 선택합니다.\n", + "\n", + " 정책 형태에 제약이 없는 유연한 접근이지만, 최종 성능은 CATE 추정 품질에 크게 영향을 받습니다.\n", + "\n", + "- DRPolicyTree\n", + "\n", + " DRPolicyTree는 AIPW score를 직접 최대화하도록 정책을 학습합니다.\n", + "\n", + " 정책 구조를 얕은 decision tree로 제한하기 때문에, 결과를 if-then 규칙 형태로 해석할 수 있습니다. 따라서 정책 해석과 설명이 중요한 상황에서 유용합니다.\n", + "\n", + "- DRPolicyForest\n", + "\n", + " DRPolicyForest 역시 AIPW score를 직접 최대화하지만, 단일 트리 대신 forest 구조를 사용합니다.\n", + "\n", + " 단일 tree는 구조가 단순하지만 분산이 커질 수 있습니다. 반면 forest는 여러 트리의 결과를 평균적으로 활용하기 때문에 분산을 줄이면서 더 안정적으로 이질적인 처치 효과를 학습할 수 있습니다.\n", + "\n", + " 대신 최종 정책을 하나의 명확한 if-then 규칙 형태로 해석하기는 더 어렵습니다.\n", + "\n", + "세 정책은 모두 train set에서 학습하고, 동일한 test set의 AIPW score로 정책 가치를 비교합니다." + ] + }, + { + "cell_type": "markdown", + "id": "d7d4bc26", + "metadata": {}, + "source": [ + "### 평가 지표 구성\n", + "\n", + "AIPW(Augmented Inverse Probability Weighting) score는 다음과 같이 정의합니다.\n", + "\n", + "$$\n", + "\\hat\\Gamma_{i,a} = \\hat\\mu_a(X_i) + \\frac{\\mathbf{1}[A_i=a]}{\\hat e_a(X_i)}\\bigl(Y_i^{net} - \\hat\\mu_a(X_i)\\bigr)\n", + "$$\n", + "\n", + "- $\\hat\\mu_a(X_i)$: outcome model이 예측한 $E[Y^{net}(a)\\mid X_i]$\n", + "- $\\hat e_a(X_i)$: propensity model이 예측한 $P(A_i=a\\mid X_i)$\n", + "- 두 번째 항은 실제 관측값과 예측값의 차이를 IPW로 보정하는 역할을 합니다.\n", + "\n", + "이 구조 덕분에 outcome model과 propensity model 중 하나만 올바르게 추정되어도 불편성이 유지됩니다. 이를 doubly robust(이중 견고성)라고 부릅니다." + ] + }, + { + "cell_type": "code", + "execution_count": 98, + "id": "e3156595", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:48.578060Z", + "iopub.status.busy": "2026-05-09T16:49:48.577946Z", + "iopub.status.idle": "2026-05-09T16:49:49.475582Z", + "shell.execute_reply": "2026-05-09T16:49:49.474994Z" + } + }, + "outputs": [], + "source": [ + "Y_net_train = Y_net[train_idx]\n", + "Y_net_test = Y_net[test_idx]\n", + "\n", + "prop_model_net = RandomForestClassifier(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + ")\n", + "\n", + "prop_model_net.fit(X_train, A_train)\n", + "e_hat_net = prop_model_net.predict_proba(X_test)\n", + "e_hat_net = np.clip(e_hat_net, 0.02, 0.98)\n", + "e_hat_net = e_hat_net / e_hat_net.sum(axis=1, keepdims=True)\n", + "\n", + "gamma_net = np.zeros((len(X_test), 4))\n", + "mu_hat_net = np.zeros((len(X_test), 4))\n", + "\n", + "for arm_id in range(4):\n", + " outcome_model = RandomForestRegressor(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + " )\n", + " outcome_model.fit(X_train[A_train == arm_id], Y_net_train[A_train == arm_id])\n", + " mu_a = outcome_model.predict(X_test)\n", + " observed_a = (A_test == arm_id).astype(float)\n", + "\n", + " gamma_net[:, arm_id] = mu_a + observed_a / e_hat_net[:, arm_id] * (Y_net_test - mu_a)\n", + " mu_hat_net[:, arm_id] = mu_a\n" + ] + }, + { + "cell_type": "code", + "execution_count": 99, + "id": "e8533283", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:49.477219Z", + "iopub.status.busy": "2026-05-09T16:49:49.477080Z", + "iopub.status.idle": "2026-05-09T16:49:49.483510Z", + "shell.execute_reply": "2026-05-09T16:49:49.483027Z" + } + }, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
samplee_hatmu_hatgamma
sample_idactual_armobserved_net_outcomenonetech_support_onlydiscount_onlydiscount_plus_supportnonetech_support_onlydiscount_onlydiscount_plus_supportnonetech_support_onlydiscount_onlydiscount_plus_support
00discount_plus_support9254.5585410.2486330.2673380.2723660.2116625370.0992819559.7777634547.4845526984.7316095370.0992819559.7777634547.48455217708.540116
11tech_support_only7702.8616620.2110870.3072040.2433940.23831510468.9425226728.9884198373.31663311121.82660910468.9425229899.1085088373.31663311121.826609
22tech_support_only5062.1844200.3765700.2419010.2562610.1252682879.7417865449.409859937.5151032959.5143472879.7417863848.652052937.5151032959.514347
33none8930.4950220.3396530.1947010.2465050.2191428458.7792178254.3720696054.5989736560.0280699847.5980528254.3720696054.5989736560.028069
44none8156.5755020.3122650.2500210.3121230.12559010137.4698213378.7480355743.8029392988.2173143793.8450543378.7480355743.8029392988.217314
\n", + "
" + ], + "text/plain": [ + " sample e_hat \\\n", + " sample_id actual_arm observed_net_outcome none \n", + "0 0 discount_plus_support 9254.558541 0.248633 \n", + "1 1 tech_support_only 7702.861662 0.211087 \n", + "2 2 tech_support_only 5062.184420 0.376570 \n", + "3 3 none 8930.495022 0.339653 \n", + "4 4 none 8156.575502 0.312265 \n", + "\n", + " mu_hat \\\n", + " tech_support_only discount_only discount_plus_support none \n", + "0 0.267338 0.272366 0.211662 5370.099281 \n", + "1 0.307204 0.243394 0.238315 10468.942522 \n", + "2 0.241901 0.256261 0.125268 2879.741786 \n", + "3 0.194701 0.246505 0.219142 8458.779217 \n", + "4 0.250021 0.312123 0.125590 10137.469821 \n", + "\n", + " gamma \\\n", + " tech_support_only discount_only discount_plus_support none \n", + "0 9559.777763 4547.484552 6984.731609 5370.099281 \n", + "1 6728.988419 8373.316633 11121.826609 10468.942522 \n", + "2 5449.409859 937.515103 2959.514347 2879.741786 \n", + "3 8254.372069 6054.598973 6560.028069 9847.598052 \n", + "4 3378.748035 5743.802939 2988.217314 3793.845054 \n", + "\n", + " \n", + " tech_support_only discount_only discount_plus_support \n", + "0 9559.777763 4547.484552 17708.540116 \n", + "1 9899.108508 8373.316633 11121.826609 \n", + "2 3848.652052 937.515103 2959.514347 \n", + "3 8254.372069 6054.598973 6560.028069 \n", + "4 3378.748035 5743.802939 2988.217314 " + ] + }, + "execution_count": 99, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "sample_ids = np.arange(5)\n", + "\n", + "sample_prediction_table = pd.concat(\n", + " {\n", + " 'sample': pd.DataFrame({\n", + " 'sample_id': sample_ids,\n", + " 'actual_arm': pd.Series(A_test[sample_ids]).map(ARM_NAMES).to_numpy(),\n", + " 'observed_net_outcome': Y_net_test[sample_ids],\n", + " }),\n", + " 'e_hat': pd.DataFrame(e_hat_net[sample_ids], columns=ARM_LABELS),\n", + " 'mu_hat': pd.DataFrame(mu_hat_net[sample_ids], columns=ARM_LABELS),\n", + " 'gamma': pd.DataFrame(gamma_net[sample_ids], columns=ARM_LABELS),\n", + " },\n", + " axis=1,\n", + ")\n", + "\n", + "sample_prediction_table" + ] + }, + { + "cell_type": "markdown", + "id": "3c967473", + "metadata": {}, + "source": [ + "### Plug-in Policy\n", + "\n", + "Plug-in policy는 먼저 고객별 처치 효과(CATE)를 추정한 뒤, 그 값을 이용해 정책을 만드는 방식입니다.\n", + "\n", + "Binary treatment에서는 일반적으로 $\\hat\\tau(x) > 0$이면 처치하고, 비용이 있으면 $\\hat\\tau(x) > c(x)$일 때 처치합니다.\n", + "\n", + "여러 처치가 있는 경우에는 baseline 처치와의 상대효과를 비교합니다. 여기서는 `none`(처치 없음)을 baseline으로 두고, `DRLearner`로 다음 값을 추정합니다.\n", + "\n", + "$$\n", + "\\hat\\tau_a(x) = \\widehat{\\mathbb{E}}[Y^{net}(a) - Y^{net}(0) \\mid X=x]\n", + "\\qquad a \\in \\{1,2,3\\}\n", + "$$\n", + "\n", + "baseline 처치의 상대효과는 0이므로, 고객별로 다음 네 값을 비교합니다.\n", + "\n", + "$$\n", + "[0, \\hat\\tau_1(x), \\hat\\tau_2(x), \\hat\\tau_3(x)]\n", + "$$\n", + "\n", + "그리고 가장 큰 값을 갖는 처치를 선택합니다.\n", + "\n", + "$$\n", + "\\hat\\pi_{plugin}(x) = \\arg\\max_{a \\in \\{0,1,2,3\\}} \\widehat{\\mathbb{E}}[Y^{net}(a) - Y^{net}(0) \\mid X=x]\n", + "$$\n", + "\n", + "이 방식은 직관적이고 유연하지만, CATE 추정 품질에 크게 영향을 받습니다. 따라서 학습된 정책은 별도의 test set에서 AIPW value로 평가합니다." + ] + }, + { + "cell_type": "code", + "execution_count": 100, + "id": "debd92a9", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:49.484723Z", + "iopub.status.busy": "2026-05-09T16:49:49.484610Z", + "iopub.status.idle": "2026-05-09T16:49:52.385416Z", + "shell.execute_reply": "2026-05-09T16:49:52.384824Z" + } + }, + "outputs": [], + "source": [ + "dr_cate = DRLearner(\n", + " model_regression=RandomForestRegressor(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + " ),\n", + " model_propensity=RandomForestClassifier(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + " ),\n", + " model_final=RandomForestRegressor(\n", + " n_estimators=300,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + " ),\n", + " categories=[0, 1, 2, 3],\n", + " min_propensity=0.02,\n", + " cv=3,\n", + " random_state=SEED,\n", + ")\n", + "dr_cate.fit(Y_net_train, A_train, X=X_train)\n", + "\n", + "cate_vs_none = dr_cate.const_marginal_effect(X_test)\n", + "plugin_arm_values = np.column_stack([\n", + " np.zeros(len(X_test)),\n", + " cate_vs_none,\n", + "])\n", + "pi_plugin = np.argmax(plugin_arm_values, axis=1)\n" + ] + }, + { + "cell_type": "code", + "execution_count": 101, + "id": "33437c2c", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:52.389559Z", + "iopub.status.busy": "2026-05-09T16:49:52.389447Z", + "iopub.status.idle": "2026-05-09T16:49:52.392977Z", + "shell.execute_reply": "2026-05-09T16:49:52.392636Z" + } + }, + "outputs": [ + { + "data": { + "text/plain": [ + "none 0.092\n", + "tech_support_only 0.500\n", + "discount_only 0.035\n", + "discount_plus_support 0.373\n", + "Name: proportion, dtype: float64" + ] + }, + "execution_count": 101, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "pd.Series(pi_plugin).map(ARM_NAMES).value_counts(normalize=True).reindex(ARM_LABELS).fillna(0)" + ] + }, + { + "cell_type": "code", + "execution_count": 102, + "id": "33ktqy8zrqm", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
SizeEmployee CountIT Spendassigned_armτ_techτ_discτ_both
09620414217655none-6140-1419-1719
1253901787971none-14556-4252-7232
26655527919899none-17629-2896-16582
3708682026263tech_support_only4265-404155
424152184610tech_support_only4174-2316-771
5548556314772tech_support_only2042-698553
639922836133052discount_plus_support7849-128111819
71817042459529discount_plus_support9158521610248
829437737112485discount_plus_support71348589048
\n", + "
" + ], + "text/plain": [ + " Size Employee Count IT Spend assigned_arm τ_tech τ_disc \\\n", + "0 96204 142 17655 none -6140 -1419 \n", + "1 25390 178 7971 none -14556 -4252 \n", + "2 66555 279 19899 none -17629 -2896 \n", + "3 70868 20 26263 tech_support_only 4265 -40 \n", + "4 24152 18 4610 tech_support_only 4174 -2316 \n", + "5 54855 63 14772 tech_support_only 2042 -698 \n", + "6 399228 36 133052 discount_plus_support 7849 -1281 \n", + "7 181704 24 59529 discount_plus_support 9158 5216 \n", + "8 294377 37 112485 discount_plus_support 7134 858 \n", + "\n", + " τ_both \n", + "0 -1719 \n", + "1 -7232 \n", + "2 -16582 \n", + "3 4155 \n", + "4 -771 \n", + "5 553 \n", + "6 11819 \n", + "7 10248 \n", + "8 9048 " + ] + }, + "execution_count": 102, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "none_idx = np.where(pi_plugin == 0)[0][:3]\n", + "tech_idx = np.where(pi_plugin == 1)[0][:3]\n", + "both_idx = np.where(pi_plugin == 3)[0][:3]\n", + "sample_ids = np.concatenate([none_idx, tech_idx, both_idx])\n", + "\n", + "df_plugin_sample = pd.DataFrame(X_test[sample_ids], columns=covariates)[['Size', 'Employee Count', 'IT Spend']]\n", + "df_plugin_sample['assigned_arm'] = pd.Series(pi_plugin[sample_ids]).map(ARM_NAMES).values\n", + "df_plugin_sample[['τ_tech', 'τ_disc', 'τ_both']] = np.round(cate_vs_none[sample_ids], 0).astype(int)\n", + "df_plugin_sample.index = pd.RangeIndex(len(df_plugin_sample))\n", + "df_plugin_sample" + ] + }, + { + "cell_type": "markdown", + "id": "05b0ee18", + "metadata": {}, + "source": [ + "plug-in 정책은 약 50% 고객에게 `tech_support_only`, 37% 고객에게 `discount_plus_support`, 그리고 약 9% 고객에게 `none`을 배정합니다.\n", + "\n", + "비용을 함께 고려하면서, 기대 순이익이 음수로 예상되는 일부 고객에게는 처치를 하지 않는 선택이 나타납니다." + ] + }, + { + "cell_type": "markdown", + "id": "97d958cd", + "metadata": {}, + "source": [ + "### Policy Tree\n", + "\n", + "현업에서는 성능뿐 아니라 정책의 설명 가능성도 중요한 경우가 많습니다. 얕은 decision tree 기반 정책은 다음과 같은 이유로 실무에서 자주 사용됩니다.\n", + "\n", + "- **이해관계자 설명**: if-then 규칙 형태라 정책을 쉽게 설명하고 공유할 수 있습니다.\n", + "- **공정성 검토**: 어떤 고객이 어떤 처치를 받는지 구조가 명확해 편향 여부를 점검하기 쉽습니다.\n", + "- **운영 안정성**: 복잡한 모델보다 단순한 규칙 기반 정책이 실제 운영과 관리 측면에서 안정적입니다.\n", + "\n", + "이 경우 정책을 모든 가능한 함수에서 찾는 대신, 제한된 정책 클래스 $\\Pi$ 안에서 탐색합니다.\n", + "\n", + "$$\n", + "\\hat\\pi = \\arg\\max_{\\pi \\in \\Pi}\n", + "\\frac{1}{n}\\sum_{i=1}^{n} \\widehat\\Gamma_{i,\\pi(X_i)}\n", + "$$\n", + "\n", + "Policy Tree는 $\\Pi$를 얕은 decision tree로 제한합니다. 따라서 depth를 작게 두면 사람이 읽을 수 있는 형태로 정책을 해석할 수 있습니다. 여기서는 `econml`의 `DRPolicyTree`를 사용합니다.\n", + "\n", + "또한 leaf가 지나치게 작아지면 다중 처치 환경에서 value 추정이 불안정해질 수 있으므로, `min_samples_leaf`를 사용해 너무 작은 leaf를 방지합니다." + ] + }, + { + "cell_type": "code", + "execution_count": 103, + "id": "6c987a8b", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:53.651642Z", + "iopub.status.busy": "2026-05-09T16:49:53.651575Z", + "iopub.status.idle": "2026-05-09T16:49:55.354506Z", + "shell.execute_reply": "2026-05-09T16:49:55.354139Z" + } + }, + "outputs": [ + { + "data": { + "text/plain": [ + "none 0.199\n", + "tech_support_only 0.475\n", + "discount_only 0.000\n", + "discount_plus_support 0.326\n", + "Name: proportion, dtype: float64" + ] + }, + "execution_count": 103, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "dr_policy_tree = DRPolicyTree(\n", + " max_depth=2,\n", + " min_samples_leaf=30,\n", + " model_regression=RandomForestRegressor(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + " ),\n", + " model_propensity=RandomForestClassifier(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + " ),\n", + " categories=[0, 1, 2, 3],\n", + " min_propensity=0.02,\n", + " cv=3,\n", + " random_state=SEED,\n", + ")\n", + "dr_policy_tree.fit(Y_net_train, A_train, X=X_train)\n", + "pi_tree = dr_policy_tree.predict(X_test).astype(int).ravel()\n", + "\n", + "pd.Series(pi_tree).map(ARM_NAMES).value_counts(normalize=True).reindex(ARM_LABELS).fillna(0)\n" + ] + }, + { + "cell_type": "code", + "execution_count": 104, + "id": "64529830", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:55.355867Z", + "iopub.status.busy": "2026-05-09T16:49:55.355774Z", + "iopub.status.idle": "2026-05-09T16:49:55.574472Z", + "shell.execute_reply": "2026-05-09T16:49:55.574068Z" + } + }, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(figsize=(13, 7))\n", + "dr_policy_tree.plot(feature_names=covariates, treatment_names=ARM_LABELS, ax=ax)\n", + "ax.set_title('4-arm DRPolicyTree')\n", + "plt.tight_layout()" + ] + }, + { + "cell_type": "markdown", + "id": "2a3ffb5d", + "metadata": {}, + "source": [ + "이 트리 정책은 주로 `Size`를 기준으로 고객을 나눕니다. 특징적으로 약 20%의 고객에게는 `none`을 배정합니다. 비용을 반영한 결과, 일부 고객은 처치했을 때 기대 순이익이 낮아 처치하지 않는 편이 더 낫다고 판단한 것입니다.\n", + "\n", + "전체적으로는 규모가 큰 고객에게 `discount_plus_support`를, 그 외 고객에게는 주로 `tech_support_only`를 배정합니다. 기대 순이익이 낮은 고객에게는 `none`을 선택합니다." + ] + }, + { + "cell_type": "markdown", + "id": "35dd95b1", + "metadata": {}, + "source": [ + "### Policy Forest\n", + "\n", + "Policy Forest는 여러 tree를 사용해 더 유연한 정책을 학습합니다. 단일 tree보다 복잡한 이질성을 더 잘 포착할 수 있지만, 개별 규칙 형태로 해석하기는 어렵습니다.\n", + "\n", + "여기서는 `econml`의 `DRPolicyForest`를 사용합니다.\n", + "\n", + "실무에서는 일반적으로 forest를 성능 중심의 정책으로, tree를 해석 가능한 정책으로 보는 경우가 많습니다. 두 방법 모두 학습 이후에는 동일한 test set의 AIPW score로 정책 가치를 비교합니다." + ] + }, + { + "cell_type": "code", + "execution_count": 105, + "id": "d7939447", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:55.582919Z", + "iopub.status.busy": "2026-05-09T16:49:55.582816Z", + "iopub.status.idle": "2026-05-09T16:49:57.495404Z", + "shell.execute_reply": "2026-05-09T16:49:57.494962Z" + } + }, + "outputs": [ + { + "data": { + "text/plain": [ + "none 0.113\n", + "tech_support_only 0.471\n", + "discount_only 0.000\n", + "discount_plus_support 0.416\n", + "Name: proportion, dtype: float64" + ] + }, + "execution_count": 105, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "dr_policy_forest = DRPolicyForest(\n", + " n_estimators=400,\n", + " max_depth=5,\n", + " min_samples_leaf=30,\n", + " model_regression=RandomForestRegressor(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + " ),\n", + " model_propensity=RandomForestClassifier(\n", + " n_estimators=400,\n", + " min_samples_leaf=20,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + " ),\n", + " categories=[0, 1, 2, 3],\n", + " min_propensity=0.02,\n", + " cv=3,\n", + " random_state=SEED,\n", + " n_jobs=1,\n", + ")\n", + "dr_policy_forest.fit(Y_net_train, A_train, X=X_train)\n", + "pi_forest = dr_policy_forest.predict(X_test).astype(int).ravel()\n", + "\n", + "pd.Series(pi_forest).map(ARM_NAMES).value_counts(normalize=True).reindex(ARM_LABELS).fillna(0)\n" + ] + }, + { + "cell_type": "markdown", + "id": "15d7d727", + "metadata": {}, + "source": [ + "DRPolicyForest는 전반적으로 plug-in 정책과 유사한 배정을 보입니다. `none` 11%, `tech_support_only` 47%, `discount_plus_support` 42%로, plug-in 정책보다 `discount_plus_support` 비중이 조금 더 높게 나타납니다.\n", + "\n", + "Forest는 단일 if-then 규칙 형태로 해석하기는 어렵지만, 더 복잡한 이질성을 학습할 수 있다는 장점이 있습니다. 따라서 보통 해석 가능한 기준 정책으로는 tree를, 성능 중심 정책으로는 forest를 함께 비교합니다." + ] + }, + { + "cell_type": "markdown", + "id": "99bac91c", + "metadata": {}, + "source": [ + "## 정책 평가\n", + "\n", + "이제 학습된 정책들을 동일한 test set의 AIPW score 기준으로 비교합니다.\n", + "\n", + "비교를 위해 모든 고객에게 동일한 처치를 적용하는 baseline 정책도 함께 사용합니다.\n", + "\n", + "- `all_none`\n", + "- `all_tech_support_only`\n", + "- `all_discount_only`\n", + "- `all_discount_plus_support`\n", + "\n", + "예를 들어 `all_discount_plus_support`는 모든 고객에게 기술지원과 할인을 함께 제공했을 때의 평균 순이익을 의미합니다.\n", + "\n", + "반면 학습된 정책은 고객 특성 $X_i$에 따라 서로 다른 처치를 배정합니다.\n", + "\n", + "만약 학습된 정책의 value가 가장 좋은 baseline 정책보다 높다면, 모든 고객에게 동일한 처치를 적용하는 것보다 고객별 targeting이 더 효과적이라는 의미입니다.\n", + "\n", + "여기서 비교하는 값은 단순 관측 평균이 아니라, 동일한 AIPW 평가식을 통해 추정한 counterfactual policy value입니다." + ] + }, + { + "cell_type": "code", + "execution_count": 109, + "id": "14b616ad", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-09T16:49:57.502590Z", + "iopub.status.busy": "2026-05-09T16:49:57.502507Z", + "iopub.status.idle": "2026-05-09T16:49:57.508438Z", + "shell.execute_reply": "2026-05-09T16:49:57.508058Z" + } + }, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
policyvalueci_lowerci_upperrate_nonerate_tech_support_onlyrate_discount_onlyrate_discount_plus_support
4plugin_drlearner_4arm11810.37771211283.56156312337.1938610.0920.5000.0350.373
6dr_policy_forest_4arm11625.62754011102.02258912149.2324910.1130.4710.0000.416
5dr_policy_tree_4arm11148.46236910700.60923011596.3155080.1990.4750.0000.326
1all_tech_support_only10526.5238769833.03475111220.0130010.0001.0000.0000.000
3all_discount_plus_support10136.7088829439.44594110833.9718240.0000.0000.0001.000
2all_discount_only7545.0792056993.5366218096.6217880.0000.0001.0000.000
0all_none7249.3454676955.4474327543.2435021.0000.0000.0000.000
\n", + "
" + ], + "text/plain": [ + " policy value ci_lower ci_upper \\\n", + "4 plugin_drlearner_4arm 11810.377712 11283.561563 12337.193861 \n", + "6 dr_policy_forest_4arm 11625.627540 11102.022589 12149.232491 \n", + "5 dr_policy_tree_4arm 11148.462369 10700.609230 11596.315508 \n", + "1 all_tech_support_only 10526.523876 9833.034751 11220.013001 \n", + "3 all_discount_plus_support 10136.708882 9439.445941 10833.971824 \n", + "2 all_discount_only 7545.079205 6993.536621 8096.621788 \n", + "0 all_none 7249.345467 6955.447432 7543.243502 \n", + "\n", + " rate_none rate_tech_support_only rate_discount_only \\\n", + "4 0.092 0.500 0.035 \n", + "6 0.113 0.471 0.000 \n", + "5 0.199 0.475 0.000 \n", + "1 0.000 1.000 0.000 \n", + "3 0.000 0.000 0.000 \n", + "2 0.000 0.000 1.000 \n", + "0 1.000 0.000 0.000 \n", + "\n", + " rate_discount_plus_support \n", + "4 0.373 \n", + "6 0.416 \n", + "5 0.326 \n", + "1 0.000 \n", + "3 1.000 \n", + "2 0.000 \n", + "0 0.000 " + ] + }, + "execution_count": 109, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "policy_assignments = {\n", + " **{f'all_{arm_name}': np.full(len(X_test), arm_id) for arm_id, arm_name in ARM_NAMES.items()},\n", + " 'plugin_drlearner_4arm': pi_plugin,\n", + " 'dr_policy_tree_4arm': pi_tree,\n", + " 'dr_policy_forest_4arm': pi_forest,\n", + "}\n", + "\n", + "eval_rows = []\n", + "for policy_name, policy_assignment in policy_assignments.items():\n", + " pi = np.asarray(policy_assignment).astype(int).ravel()\n", + " scores = gamma_net[np.arange(len(pi)), pi]\n", + "\n", + " row = {\n", + " 'policy': policy_name,\n", + " 'value': scores.mean(),\n", + " 'value_se': scores.std(ddof=1) / np.sqrt(len(scores)),\n", + " }\n", + " for arm_id, arm_name in ARM_NAMES.items():\n", + " row[f'rate_{arm_name}'] = np.mean(pi == arm_id)\n", + " eval_rows.append(row)\n", + "\n", + "policy_eval = pd.DataFrame(eval_rows)\n", + "policy_eval['ci_lower'] = policy_eval['value'] - 1.96 * policy_eval['value_se']\n", + "policy_eval['ci_upper'] = policy_eval['value'] + 1.96 * policy_eval['value_se']\n", + "\n", + "display_cols = ['policy', 'value', 'ci_lower', 'ci_upper'] + [f'rate_{n}' for n in ARM_LABELS]\n", + "policy_eval.sort_values('value', ascending=False)[display_cols]" + ] + }, + { + "cell_type": "code", + "execution_count": 108, + "id": "4814d1ed", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, axes = plt.subplots(1, 2, figsize=(15, 5))\n", + "plot_df = policy_eval.sort_values('value', ascending=False)\n", + "sns.barplot(data=plot_df, x='value', y='policy', ax=axes[0], color='seagreen')\n", + "axes[0].set_title('4-arm net policy value')\n", + "axes[0].set_xlabel('AIPW-estimated net value')\n", + "axes[0].set_ylabel('')\n", + "\n", + "arm_rate_cols = [f'rate_{name}' for name in ARM_LABELS]\n", + "rate_df = policy_eval.set_index('policy')[arm_rate_cols]\n", + "rate_df.columns = ARM_LABELS\n", + "rate_df.loc[['plugin_drlearner_4arm', 'dr_policy_tree_4arm', 'dr_policy_forest_4arm']].plot(kind='barh', stacked=True, ax=axes[1], colormap='tab20')\n", + "axes[1].set_title('Assigned arm share')\n", + "axes[1].set_xlabel('Share')\n", + "axes[1].set_ylabel('')\n", + "axes[1].legend(loc='center left', bbox_to_anchor=(1.0, 0.5))\n", + "plt.tight_layout()" + ] + }, + { + "cell_type": "markdown", + "id": "dec4e9f6", + "metadata": {}, + "source": [ + "DRLearner plug-in 정책(11,810)의 정책 가치가 가장 높았습니다. `tech_support_only`(50%)와 `discount_plus_support`(37%)를 중심으로 배정하고, 일부 고객에게는 `none`(9%)을 선택했습니다.\n", + "\n", + "`DRPolicyForest`(11,626)와 `DRPolicyTree`(11,148)도 모두 높은 성능을 보였지만, plug-in 정책보다는 다소 낮았습니다. 특히 tree 정책은 해석 가능한 구조를 유지하는 대신 더 보수적으로 배정하며, 약 20% 고객에게 `none`을 선택했습니다.\n", + "\n", + "모든 고객에게 동일한 처치를 적용하는 정책 중에서는 `all_tech_support_only`(10,527)가 가장 높았고, `all_discount_plus_support`(10,137)가 그 뒤를 이었습니다. 할인 비용이 고객 규모에 따라 증가하도록 설정했기 때문에, 일부 고객에서는 할인 추가 제공의 편익이 비용을 충분히 상쇄하지 못한 것으로 볼 수 있습니다.\n", + "\n", + "신뢰구간 기준으로 보면 `DRLearner plug-in`과 `DRPolicyForest`는 구간이 상당 부분 겹쳐 두 정책의 성능 차이가 크다고 보기는 어렵습니다. 반면 `DRPolicyTree`는 전체적으로 더 낮은 구간에 위치해, 해석 가능성을 유지하는 대신 일부 성능을 희생한 것으로 해석할 수 있습니다.\n", + "\n", + "전체적으로는 세 학습 정책 모두 가장 성능이 좋은 단일 처치 정책보다 높은 value를 보였습니다. 이는 모든 고객에게 동일한 처치를 적용하는 것보다 고객별 targeting이 더 효과적이라는 의미입니다." + ] + }, + { + "cell_type": "markdown", + "id": "b40f71c0", + "metadata": {}, + "source": [ + "## 결론\n", + "\n", + "고객 특성에 따라 처치를 다르게 배정한 세 정책 모두, 가장 성능이 좋은 단일 처치 정책(`all_tech_support_only`, 10,527)보다 높은 value를 기록했습니다.\n", + "\n", + "- DRLearner plug-in: 11,810\n", + "- DRPolicyForest: 11,626\n", + "- DRPolicyTree: 11,148\n", + "\n", + "또한 세 정책 모두 무처치 정책 대비 통계적으로 유의하게 높은 성능을 보였습니다.\n", + "\n", + "공통적으로는 규모가 큰 고객에게 `discount_plus_support`를, 상대적으로 작은 고객에게는 `tech_support_only`를 주로 배정했습니다. 반면 기대 순이익이 낮은 일부 고객에게는 `none`을 선택했습니다.\n", + "\n", + "즉, 비용이 고객 규모에 따라 달라지는 환경에서도, 모든 고객에게 동일한 처치를 적용하는 것보다 고객 특성에 맞춰 targeting하는 전략이 더 효과적임을 확인할 수 있었습니다.\n", + "\n", + "### 참고 자료\n", + "\n", + "- **이 노트북의 주요 참고 튜토리얼**: [Stanford ML+CI Tutorial — Policy Learning I (Binary Treatment)](https://bookdown.org/stanfordgsbsilab/ml-ci-tutorial/policy-learning-i---binary-treatment.html?utm_source=chatgpt.com)\n", + "- **데이터셋**: [Software Usage Promotion Campaign Uplift Model (Kaggle)](https://www.kaggle.com/datasets/hwwang98/software-usage-promotion-campaign-uplift-model?utm_source=chatgpt.com)\n", + "- **모수적 정책학습 이론**: [Athey and Wager (2021, Econometrica)](https://onlinelibrary.wiley.com/doi/abs/10.3982/ECTA15732?utm_source=chatgpt.com) · [ArXiv Version (1702.02896)](https://arxiv.org/abs/1702.02896?utm_source=chatgpt.com)\n", + "- **비용 곡선을 통한 정책 비교**: [Imai and Li (2019)](https://arxiv.org/pdf/1905.05389.pdf?utm_source=chatgpt.com)\n", + "- **추가 참고**: [arXiv 2604.06123](https://arxiv.org/html/2604.06123v1?utm_source=chatgpt.com)" + ] + }, + { + "cell_type": "markdown", + "id": "e0ec9c4f", + "metadata": {}, + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": ".venv", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.14.3" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/requirements-book.txt b/requirements-book.txt index 898be45..0548471 100644 --- a/requirements-book.txt +++ b/requirements-book.txt @@ -5,6 +5,9 @@ pandas scipy matplotlib seaborn -scikit-learn +scikit-learn<1.7 statsmodels graphviz +econml +scikit-uplift +kagglehub diff --git a/requirements.txt b/requirements.txt index d010b0b..c75646f 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,5 +1,6 @@ # Jupyter Book (2.x - MyST-MD 기반) jupyter-book>=2.0.0 +jupyterlab # Video Production manim>=0.18.0 @@ -9,6 +10,18 @@ numpy pandas matplotlib seaborn -scikit-learn +scikit-learn<1.7 +scikit-uplift statsmodels graphviz +econml +kagglehub +ipywidgets + +# PDF Processing +pypdf +pdfplumber +reportlab +pdf2image +pytesseract +Pillow