feat: add concurrency safety, caption detection, admin enhancements, and performance improvements

app/services/jobs.py

@@ -148,7 +148,7 @@ async def run_job(db: Session, job_id: int) -> dict:
 
 async def _dispatch_job(db: Session, job: Job, payload: dict) -> dict:
     from app.services.cleaner import cleanup_tmp, delete_papers_by_date_range
-    from app.services.crawler import refresh_upvotes
+    from app.services.crawler import recrawl_single, refresh_upvotes
     from app.services.derived import reindex_chroma, reindex_fts
     from app.services.pipeline import run_crawl, run_pipeline
     from app.services.summarizer import summarize_batch, summarize_single
@@ -193,6 +193,20 @@ async def _dispatch_job(db: Session, job: Job, payload: dict) -> dict:
         return reindex_fts(db)
     if job.type == "reindex_chroma":
         return reindex_chroma(db)
+    if job.type == "recrawl_one":
+        return await recrawl_single(db, payload["arxiv_id"])
+    if job.type == "recrawl_batch":
+        updated = 0
+        skipped = 0
+        results = []
+        for arxiv_id in payload.get("arxiv_ids", []):
+            res = await recrawl_single(db, arxiv_id)
+            results.append(res)
+            if res.get("updated"):
+                updated += 1
+            else:
+                skipped += 1
+        return {"updated": updated, "skipped": skipped, "results": results}
 
     raise ValueError(f"Unsupported job type: {job.type}")