Improve readability in full-text extraction paths.

Wrap long extraction logic for readability and return typed API responses so the on-demand extraction endpoint passes strict type checks.

Made-with: Cursor

Author: LeslieLeung

backend/apps/api/glean_api/routers/entries.py

@@ -5,7 +5,7 @@
 """
 
 from contextlib import suppress
-from typing import Annotated
+from typing import Annotated, cast
 
 from arq.connections import ArqRedis
 from fastapi import APIRouter, Depends, HTTPException, Query, status
@@ -191,15 +191,19 @@ async def extract_entry_fulltext(
 
     try:
         job_result = await job.result(timeout=30)
-        status_text = str(job_result.get("status", "unknown")) if isinstance(job_result, dict) else "unknown"
+        if isinstance(job_result, dict):
+            job_result_dict = cast(dict[str, object], job_result)
+            status_text = str(job_result_dict.get("status", "unknown"))
+        else:
+            status_text = "unknown"
     except TimeoutError:
         status_text = "queued"
 
     if status_text == "updated":
         updated_entry = await entry_service.get_entry(entry_id, current_user.id)
-        return {"status": status_text, "entry": updated_entry}
+        return FullTextExtractionResponse(status=status_text, entry=updated_entry)
 
-    return {"status": status_text, "entry": None}
+    return FullTextExtractionResponse(status=status_text, entry=None)
 
 
 # M3: Preference signal endpoints

backend/apps/worker/glean_worker/tasks/entry_fulltext.py

@@ -14,7 +14,9 @@
 logger = get_logger(__name__)
 
 
-async def extract_entry_fulltext(ctx: dict[str, Any], user_id: str, entry_id: str) -> dict[str, str]:
+async def extract_entry_fulltext(
+    ctx: dict[str, Any], user_id: str, entry_id: str
+) -> dict[str, str]:
     """
     Extract full article content for a specific entry and persist it.
 
@@ -36,17 +38,24 @@ async def extract_entry_fulltext(ctx: dict[str, Any], user_id: str, entry_id: st
         entry = result.scalar_one_or_none()
 
         if not entry:
-            logger.warning("Entry not found or inaccessible", extra={"entry_id": entry_id, "user_id": user_id})
+            logger.warning(
+                "Entry not found or inaccessible", extra={"entry_id": entry_id, "user_id": user_id}
+            )
             return {"status": "not_found"}
 
         if not entry.url:
             logger.warning("Entry has no URL for extraction", extra={"entry_id": entry_id})
             return {"status": "no_url"}
 
-        logger.info("Starting on-demand full-text extraction", extra={"entry_id": entry_id, "url": entry.url})
+        logger.info(
+            "Starting on-demand full-text extraction",
+            extra={"entry_id": entry_id, "url": entry.url},
+        )
         extracted_content = await fetch_and_extract_fulltext(entry.url)
         if not extracted_content:
-            logger.warning("Full-text extraction returned empty content", extra={"entry_id": entry_id})
+            logger.warning(
+                "Full-text extraction returned empty content", extra={"entry_id": entry_id}
+            )
             return {"status": "empty"}
 
         entry.readability_content = extracted_content

backend/packages/rss/glean_rss/extractor.py

@@ -194,7 +194,9 @@ async def extract_fulltext(html: str, url: str | None = None) -> str | None:
     # pick the largest content-like container and sanitize it.
     try:
         soup = await asyncio.to_thread(BeautifulSoup, html, "html.parser")
-        candidates = soup.select("article, main, [role='main'], .post-content, .entry-content, .article-body")
+        candidates = soup.select(
+            "article, main, [role='main'], .post-content, .entry-content, .article-body"
+        )
         best_html = ""
         best_len = 0
         for candidate in candidates: