config.py

import json
import os
from dotenv import load_dotenv

load_dotenv()

# ── Calendly ──────────────────────────────────────────────────────────────────
CALENDLY_WEBHOOK_SECRET: str = os.getenv("CALENDLY_WEBHOOK_SECRET", "")
CALENDLY_TOKEN: str = os.getenv("CALENDLY_TOKEN", "")

# ── Feishu / Lark ─────────────────────────────────────────────────────────────
FEISHU_APP_ID: str = os.getenv("FEISHU_APP_ID", "")
FEISHU_APP_SECRET: str = os.getenv("FEISHU_APP_SECRET", "")
FEISHU_GROUP_CHAT_ID: str = os.getenv("FEISHU_GROUP_CHAT_ID", "")

# Calendar: the bot will create events on this calendar
FEISHU_CALENDAR_ID: str = os.getenv("FEISHU_CALENDAR_ID", "")

# Default attendee always invited to every interview event (e.g. the research lead)
FEISHU_DEFAULT_ATTENDEE_OPEN_ID: str = os.getenv("FEISHU_DEFAULT_ATTENDEE_OPEN_ID", "")
FEISHU_DEFAULT_ATTENDEE_NAME: str = os.getenv("FEISHU_DEFAULT_ATTENDEE_NAME", "Research Lead")

# Fallback user lookup: when Calendly email not found in DB, look up by name in a Feishu sheet
FEISHU_FALLBACK_SPREADSHEET_TOKEN: str = os.getenv("FEISHU_FALLBACK_SPREADSHEET_TOKEN", "")
FEISHU_FALLBACK_SHEET_ID: str = os.getenv("FEISHU_FALLBACK_SHEET_ID", "")

# ── Bot persona ────────────────────────────────────────────────────────────────
BOT_NAME: str = os.getenv("BOT_NAME", "Interview Assistant")
PRODUCT_NAME: str = os.getenv("PRODUCT_NAME", "our product")
# Footer shown at the bottom of every Feishu card
BOT_FOOTER: str = os.getenv("BOT_FOOTER", "_⚡ Powered by ZeroPrep-Meet_")
# Text template for announcing the assigned interviewer ("{mention}" is replaced with @user)
ASSIGNEE_ANNOUNCE_TEXT: str = os.getenv(
    "ASSIGNEE_ANNOUNCE_TEXT",
    "🎯 Assigned interviewer: {mention}",
)

# ── Test mode ─────────────────────────────────────────────────────────────────
# TEST_MODE=true → send card as a DM to TEST_USER_OPEN_ID instead of the group
TEST_MODE: bool = os.getenv("TEST_MODE", "false").lower() == "true"
TEST_USER_OPEN_ID: str = os.getenv("TEST_USER_OPEN_ID", "")

# ── Interviewers rotation list ─────────────────────────────────────────────────
# JSON array, each item: {"name": "Alice", "open_id": "ou_xxx"}
_interviewers_raw = os.getenv("INTERVIEWERS", "[]")
try:
    INTERVIEWERS: list[dict] = json.loads(_interviewers_raw)
except Exception:
    INTERVIEWERS = []

# ── AI providers ──────────────────────────────────────────────────────────────
# Claude (Anthropic direct or Vertex AI)
ANTHROPIC_API_KEY: str = os.getenv("ANTHROPIC_API_KEY", "")
ANTHROPIC_VERTEX_PROJECT_ID: str = os.getenv("ANTHROPIC_VERTEX_PROJECT_ID", "")
VERTEX_REGION: str = os.getenv("VERTEX_REGION", "us-east5")

# Groq — used for free AI insight generation (Llama 3.3 70B)
GROQ_API_KEY: str = os.getenv("GROQ_API_KEY", "")

# ── BigQuery ──────────────────────────────────────────────────────────────────
# GCP project that owns the BigQuery datasets
BIGQUERY_PROJECT_ID: str = os.getenv("BIGQUERY_PROJECT_ID", "")

# Dataset containing user/subscription/order tables (MySQL-migrated or equivalent)
BQ_MYSQL_DATASET: str = os.getenv("BQ_MYSQL_DATASET", "")

# Firebase / Analytics dataset for behavioral event data
FIREBASE_PROJECT_ID: str = os.getenv("FIREBASE_PROJECT_ID", "")
FIREBASE_DATASET: str = os.getenv("FIREBASE_DATASET", "")

# Credentials: prefer JSON string (cloud deploy) over file path (local dev)
BQ_CREDENTIALS_JSON: str = os.getenv("BQ_CREDENTIALS_JSON", "")
BQ_CREDENTIALS_PATH: str = os.getenv("BQ_CREDENTIALS_PATH", "")

# ── Server ────────────────────────────────────────────────────────────────────
PORT: int = int(os.getenv("PORT", "8000"))
LOG_LEVEL: str = os.getenv("LOG_LEVEL", "INFO")

# ── Model config ──────────────────────────────────────────────────────────────
CLAUDE_MODEL: str = os.getenv("CLAUDE_MODEL", "claude-opus-4-5")
USE_VERTEX: bool = bool(ANTHROPIC_VERTEX_PROJECT_ID)