going through changes

app.py

@@ -155,53 +155,53 @@ def format_comment(
     num_failed = sum(1 for r in validation_results if not r.valid)
     total = len(validation_results)
 
+    failed = [r for r in validation_results if not r.valid]
+
+    if num_failed == 0:
+        status_line = "## ✅ EEE Validation — Ready to Merge"
+    else:
+        status_line = "## ❌ EEE Validation — Changes Requested"
+
     lines = [
-        "## EEE Validation Report",
+        status_line,
         f"**PR:** #{pr_num} | **Run:** {now}",
         "",
-        "### Validation Results",
-        "| File | Status | Details |",
-        "|------|--------|---------|",
+        f"**{num_passed}/{total} files passed**",
     ]
 
-    for r in validation_results:
-        if r.valid:
-            type_label = "EvaluationLog" if r.file_type == "json" else "InstanceLevelEvaluationLog"
-            lines.append(f"| `{r.file_path}` | PASS | Validated as {type_label} |")
-        else:
-            # Show first few errors to avoid extremely long comments
+    if num_failed > 0:
+        lines.append("")
+        lines.append("### Failures")
+        lines.append("| File | Details |")
+        lines.append("|------|---------|")
+        for r in failed:
             error_summary = "; ".join(r.errors[:5])
             if len(r.errors) > 5:
                 error_summary += f" ... and {len(r.errors) - 5} more error(s)"
-            lines.append(f"| `{r.file_path}` | FAIL | {error_summary} |")
+            lines.append(f"| `{r.file_path}` | {error_summary} |")
 
     # Dedup section
-    lines.append("")
-    lines.append("### Duplicate Check")
-
     has_any_dupes = False
+    dedup_lines: list[str] = []
     for dr in dedup_report.results:
         if dr.exact_duplicate_of:
-            lines.append(
+            dedup_lines.append(
                 f"- **Exact duplicate:** `{dr.file_path}` is identical to "
                 f"existing `{dr.exact_duplicate_of}`"
             )
             has_any_dupes = True
         if dr.near_duplicate_of:
-            lines.append(
+            dedup_lines.append(
                 f"- **Potential near-duplicate:** `{dr.file_path}` shares fingerprint "
                 f"with existing `{dr.near_duplicate_of}` "
                 f"(identical content minus timestamps/UUIDs)"
             )
             has_any_dupes = True
 
-    if not has_any_dupes:
-        lines.append("- No exact or near duplicates found.")
-
-    # Summary
-    lines.append("")
-    lines.append("### Summary")
-    lines.append(f"{total} file(s) checked: {num_passed} passed, {num_failed} failed")
+    if has_any_dupes:
+        lines.append("")
+        lines.append("### Duplicate Check")
+        lines.extend(dedup_lines)
 
     return "\n".join(lines)
 
@@ -264,16 +264,27 @@ def process_pr(pr_num: int) -> dict:
 # Startup sweep — catch PRs missed while the Space was asleep
 # ---------------------------------------------------------------------------
 
-def pr_has_validation_comment(pr_num: int) -> bool:
-    """Check if a PR already has an EEE Validation Report comment."""
+def pr_needs_validation(pr_num: int) -> bool:
+    """Check if a PR has commits newer than the last validation report."""
     details = api.get_discussion_details(
         repo_id=DATASET_REPO_ID,
         discussion_num=pr_num,
         repo_type="dataset",
     )
+    last_report_time = None
+    last_commit_time = None
     for event in details.events:
         if event.type == "comment" and event.content and event.content.startswith(REPORT_HEADER):
-            return True
+            last_report_time = event.created_at
+        if event.type == "commit":
+            last_commit_time = event.created_at
+
+    # No report yet — needs validation
+    if last_report_time is None:
+        return True
+    # Has commits after the last report — needs re-validation
+    if last_commit_time is not None and last_commit_time > last_report_time:
+        return True
     return False
 
 
@@ -288,10 +299,10 @@ def startup_sweep() -> None:
         for disc in discussions:
             if not disc.is_pull_request or disc.status != "open":
                 continue
-            if pr_has_validation_comment(disc.num):
-                logger.info("PR #%d already has a validation report, skipping", disc.num)
+            if not pr_needs_validation(disc.num):
+                logger.info("PR #%d is up to date, skipping", disc.num)
                 continue
-            logger.info("PR #%d has no validation report, processing", disc.num)
+            logger.info("PR #%d needs validation, processing", disc.num)
             try:
                 process_pr(disc.num)
             except Exception: