added modification to listen to content changes

app.py

@@ -6,6 +6,7 @@ files with Pydantic, checks for duplicates, and comments results on the PR.
 
 import logging
 import os
+import re
 import tempfile
 import threading
 from datetime import datetime, timezone
@@ -329,12 +330,27 @@ threading.Thread(target=startup_sweep, daemon=True).start()
 # Webhook endpoint
 # ---------------------------------------------------------------------------
 
+PR_REF_RE = re.compile(r"^refs/pr/(\d+)$")
+
+
+def _extract_pr_nums_from_refs(payload) -> list[int]:
+    """Extract PR numbers from updatedRefs in repo.content events."""
+    if not payload.updatedRefs:
+        return []
+    pr_nums = []
+    for ref in payload.updatedRefs:
+        m = PR_REF_RE.match(ref.ref)
+        if m:
+            pr_nums.append(int(m.group(1)))
+    return pr_nums
+
+
 @webhook_endpoint
 async def validate(payload):
     """Handle incoming webhook events from HuggingFace."""
     logger.info("Received webhook event: %s", payload.event)
 
-    # Filter: only dataset PRs, ignore comments
+    # Filter: ignore comments
     if payload.event.scope == "discussion.comment":
         logger.info("Skipping comment event")
         return {"status": "skipped", "reason": "comment event"}
@@ -343,6 +359,22 @@ async def validate(payload):
         logger.info("Skipping non-dataset event (type=%s)", payload.repo.type)
         return {"status": "skipped", "reason": "not a dataset repo"}
 
+    # Route 1: repo.content events (new commits pushed to PR branches)
+    if payload.event.scope == "repo.content":
+        pr_nums = _extract_pr_nums_from_refs(payload)
+        if not pr_nums:
+            logger.info("No PR refs in repo.content event")
+            return {"status": "skipped", "reason": "no PR refs"}
+        results = []
+        for pr_num in pr_nums:
+            try:
+                results.append(process_pr(pr_num))
+            except Exception:
+                logger.exception("Failed to process PR #%d", pr_num)
+                results.append({"status": "error", "pr": pr_num})
+        return {"status": "ok", "results": results}
+
+    # Route 2: discussion events (status changes, merges)
     if not payload.discussion or not payload.discussion.isPullRequest:
         logger.info("Skipping non-PR event")
         return {"status": "skipped", "reason": "not a pull request"}