Update

Author: lukasmasuch

app/spec_renderer.py

@@ -1,5 +1,7 @@
+import operator
 import re
 
+import requests
 import streamlit as st
 
 from app.utils.github_utils import (
@@ -8,6 +10,7 @@
     fetch_pull_request_payload,
     fetch_repo_file_text_at_ref,
     get_all_github_prs,
+    get_headers,
 )
 
 st.set_page_config(page_title="Spec renderer", page_icon="🔧")
@@ -28,6 +31,47 @@ def filter_spec_prs(prs: list[dict]) -> list[dict]:
     return spec_prs
 
 
+@st.cache_data(ttl=600)
+def fetch_merged_specs() -> list[dict]:
+    """Fetch list of merged spec folders from the specs directory on develop branch."""
+    url = "https://api.github.com/repos/streamlit/streamlit/contents/specs"
+    try:
+        response = requests.get(url, headers=get_headers(), params={"ref": "develop"}, timeout=30)
+        if response.status_code != 200:
+            return []
+        contents = response.json()
+    except requests.RequestException:
+        return []
+
+    specs = []
+    date_pattern = re.compile(r"^(\d{4}-\d{2}-\d{2})-(.+)$")
+
+    for item in contents:
+        if item.get("type") != "dir":
+            continue
+        name = item.get("name", "")
+        match = date_pattern.match(name)
+        if match:
+            date_str = match.group(1)
+            feature_name = match.group(2).replace("-", " ").title()
+            specs.append(
+                {
+                    "folder": name,
+                    "date": date_str,
+                    "title": feature_name,
+                    "path": f"specs/{name}/product-spec.md",
+                }
+            )
+
+    specs.sort(key=operator.itemgetter("date"), reverse=True)
+    return specs
+
+
+def fetch_merged_spec_content(spec_path: str) -> tuple[str | None, str | None]:
+    """Fetch the content of a merged spec file from the develop branch."""
+    return fetch_repo_file_text_at_ref("streamlit/streamlit", spec_path, "develop")
+
+
 @st.cache_data(ttl=300)
 def fetch_pr_files(pr_number: int) -> tuple[list[dict], str | None]:
     """Fetch files changed in a PR."""
@@ -141,11 +185,13 @@ def fetch_issue_details(issue_number: int) -> tuple[dict | None, str | None]:
 
 def replace_issue_references_with_previews(markdown_content: str) -> str:
     """Replace issue references with styled previews."""
-    # Pattern for standalone issue numbers like #1234
-    standalone_pattern = r"(?<!\[)#(\d+)(?!\])"
-
     # Pattern for issue links like [#12331](https://github.com/streamlit/streamlit/issues/12331)
-    link_pattern = r"\[#(\d+)\]\(https://github\.com/streamlit/streamlit/issues/\d+\)"
+    # Also handles optional trailing content in URL (query params, anchors)
+    link_pattern = r"\[#(\d+)\]\(https://github\.com/streamlit/streamlit/issues/\d+[^)]*\)"
+
+    # Pattern for standalone issue numbers like #1234
+    # Must not be preceded by '[' and must not be followed by ']' or '('
+    standalone_pattern = r"(?<!\[)#(\d+)(?![\]\(])"
 
     def create_issue_preview(issue_number: int) -> str:
         """Create a styled issue preview using only markdown features."""
@@ -183,11 +229,11 @@ def replace_issue_link(match: re.Match) -> str:
         issue_number = int(match.group(1))
         return create_issue_preview(issue_number)
 
-    # Replace standalone issue references first
-    updated_content = re.sub(standalone_pattern, replace_standalone_issue, markdown_content)
+    # Replace issue links FIRST (before standalone), so we don't corrupt the link syntax
+    updated_content = re.sub(link_pattern, replace_issue_link, markdown_content)
 
-    # Then replace issue links
-    updated_content = re.sub(link_pattern, replace_issue_link, updated_content)
+    # Then replace standalone issue references
+    updated_content = re.sub(standalone_pattern, replace_standalone_issue, updated_content)
 
     return updated_content
 
@@ -231,20 +277,33 @@ def extract_frontmatter(markdown_content: str) -> tuple[dict | None, str]:
     return frontmatter, content_without_frontmatter
 
 
+def is_likely_github_username(name: str) -> bool:
+    """Check if a string looks like a GitHub username."""
+    if not name or len(name) > 39:
+        return False
+    # GitHub usernames: alphanumeric and hyphens, can't start/end with hyphen
+    github_username_pattern = r"^[a-zA-Z0-9](?:[a-zA-Z0-9-]*[a-zA-Z0-9])?$"
+    return bool(re.match(github_username_pattern, name))
+
+
 def render_frontmatter_caption(frontmatter: dict) -> None:
     """Render frontmatter as caption."""
     if not frontmatter:
         return
 
     info_parts = []
 
-    # Add author (with link if it's a GitHub mention)
+    # Add author (with link if it's a GitHub mention or looks like a username)
     if "author" in frontmatter:
         author = frontmatter["author"].strip("\"'")  # Remove surrounding quotes
         if author.startswith("@"):
             username = author[1:]  # Remove @
             author_link = f"[@{username}](https://github.com/{username})"
             info_parts.append(f"By {author_link}")
+        elif is_likely_github_username(author):
+            # Treat as GitHub username even without @
+            author_link = f"[@{author}](https://github.com/{author})"
+            info_parts.append(f"By {author_link}")
         else:
             info_parts.append(f"By {author}")
 
@@ -291,18 +350,73 @@ def extract_title_and_content(markdown_content: str) -> tuple[str | None, str]:
     return title, remaining_content
 
 
-def main() -> None:
-    title_row = st.container(horizontal=True, horizontal_alignment="distribute", vertical_alignment="center")
-    with title_row:
-        st.title("🔧 Spec renderer")
-        if st.button(":material/refresh: Refresh Data", type="tertiary"):
-            get_all_github_prs.clear()
-            fetch_pr_files.clear()
-            fetch_pull_request_payload.clear()
-            fetch_repo_file_text_at_ref.clear()
-            fetch_issue_payload.clear()
-    st.markdown("Read product specs from the Streamlit repo. So far only supports PRs, not merged specs.")
+def replace_local_images_with_github_urls_for_merged_spec(markdown_content: str, spec_file_path: str) -> str:
+    """Replace local image references with GitHub raw URLs for merged specs on develop branch."""
+    # Get the directory of the spec file to resolve relative paths
+    spec_dir = "/".join(spec_file_path.split("/")[:-1])  # Remove filename, keep directory
+
+    # Pattern to match markdown images with local paths
+    image_pattern = r"!\[([^\]]*)\]\((?!https?://)([^)]+)\)"
+
+    def replace_image(match: re.Match) -> str:
+        alt_text = match.group(1)
+        image_path = match.group(2)
+
+        # Handle relative paths
+        if image_path.startswith("./"):
+            clean_path = image_path[2:]
+            full_path = f"{spec_dir}/{clean_path}" if spec_dir else clean_path
+        elif image_path.startswith("../"):
+            dir_parts = spec_dir.split("/") if spec_dir else []
+            path_parts = image_path.split("/")
+            up_count = sum(1 for part in path_parts if part == "..")
+            remaining_parts = [part for part in path_parts if part != ".."]
+            final_dir_parts = dir_parts[:-up_count] if up_count <= len(dir_parts) else []
+            full_path = "/".join(final_dir_parts + remaining_parts)
+        else:
+            full_path = f"{spec_dir}/{image_path}" if spec_dir else image_path
+
+        # Create GitHub raw URL for develop branch
+        github_url = f"https://raw.githubusercontent.com/streamlit/streamlit/develop/{full_path}"
+        return f"![{alt_text}]({github_url})"
 
+    return re.sub(image_pattern, replace_image, markdown_content)
+
+
+def render_spec_content(
+    markdown_content: str,
+    spec_file_path: str,
+    pr_number: int | None = None,
+) -> None:
+    """Render spec content with all transformations applied."""
+    # Extract frontmatter
+    frontmatter, content_without_frontmatter = extract_frontmatter(markdown_content)
+
+    # Extract title and remaining content
+    title, remaining_content = extract_title_and_content(content_without_frontmatter)
+
+    # Render title if found
+    if title:
+        st.markdown(f"# {title}")
+
+    # Render frontmatter as caption below title
+    if frontmatter:
+        render_frontmatter_caption(frontmatter)
+
+    # Replace local image references with GitHub URLs
+    if pr_number is not None:
+        processed_content = replace_local_images_with_github_urls(remaining_content, pr_number, spec_file_path)
+    else:
+        processed_content = replace_local_images_with_github_urls_for_merged_spec(remaining_content, spec_file_path)
+
+    # Replace issue references with styled previews
+    processed_content = replace_issue_references_with_previews(processed_content)
+
+    st.markdown(processed_content)
+
+
+def render_open_spec_prs() -> None:
+    """Render the open spec PRs tab."""
     # Get query parameters
     query_params = st.query_params
     pr_param = query_params.get("pr", None)
@@ -393,29 +507,88 @@ def main() -> None:
             return
 
         if markdown_content:
-            # Extract frontmatter
-            frontmatter, content_without_frontmatter = extract_frontmatter(markdown_content)
+            render_spec_content(markdown_content, spec_file, pr_number)
+        else:
+            st.error("Failed to fetch markdown content.")
+
 
-            # Extract title and remaining content
-            title, remaining_content = extract_title_and_content(content_without_frontmatter)
+def render_merged_specs() -> None:
+    """Render the merged specs tab."""
+    query_params = st.query_params
+    spec_param = query_params.get("spec", None)
 
-            # Render title if found
-            if title:
-                st.markdown(f"# {title}")
+    merged_specs = fetch_merged_specs()
+
+    if not merged_specs:
+        st.info("No merged specs found in the repository.")
+        return
 
-            # Render frontmatter as caption below title
-            if frontmatter:
-                render_frontmatter_caption(frontmatter)
+    # Create selectbox options
+    spec_options = {}
+    for spec in merged_specs:
+        label = f"{spec['date']} - {spec['title']}"
+        spec_options[label] = spec
 
-            # Replace local image references with GitHub URLs
-            processed_content = replace_local_images_with_github_urls(remaining_content, pr_number, spec_file)
+    # Determine default index based on query parameter
+    default_index = None
+    if spec_param:
+        for i, spec in enumerate(merged_specs):
+            if spec["folder"] == spec_param:
+                default_index = i
+                break
 
-            # Replace issue references with styled previews
-            processed_content = replace_issue_references_with_previews(processed_content)
+    selected_option = st.selectbox(
+        "Select a merged spec:",
+        options=list(spec_options.keys()),
+        index=default_index,
+        key="merged_spec_selector",
+    )
 
-            st.markdown(processed_content)
+    if selected_option:
+        selected_spec = spec_options[selected_option]
+
+        # Update query parameter
+        if (
+            "last_selected_spec" not in st.session_state
+            or st.session_state.last_selected_spec != selected_spec["folder"]
+        ):
+            st.query_params["spec"] = selected_spec["folder"]
+            st.session_state.last_selected_spec = selected_spec["folder"]
+
+        # Fetch and render spec content
+        with st.spinner("Fetching spec content..."):
+            markdown_content, error = fetch_merged_spec_content(selected_spec["path"])
+
+        if error:
+            st.error(error)
+            return
+
+        if markdown_content:
+            render_spec_content(markdown_content, selected_spec["path"])
         else:
-            st.error("Failed to fetch markdown content.")
+            st.warning(f"Could not find `product-spec.md` in `{selected_spec['folder']}`.")
+
+
+def main() -> None:
+    title_row = st.container(horizontal=True, horizontal_alignment="distribute", vertical_alignment="center")
+    with title_row:
+        st.title("🔧 Spec renderer")
+        if st.button(":material/refresh: Refresh Data", type="tertiary"):
+            get_all_github_prs.clear()
+            fetch_pr_files.clear()
+            fetch_pull_request_payload.clear()
+            fetch_repo_file_text_at_ref.clear()
+            fetch_issue_payload.clear()
+            fetch_merged_specs.clear()
+    st.markdown("Read product specs from the Streamlit repo.")
+
+    tab_open, tab_merged = st.tabs(["Open PRs", "Merged Specs"])
+
+    with tab_open:
+        render_open_spec_prs()
+
+    with tab_merged:
+        render_merged_specs()
 
 
 main()