Recap screen fetch cached data on load

backend/reviews/admin.py

@@ -62,10 +62,7 @@ def with_pct(counts_dict):
         .order_by("speaker__gender")
     )
     gender_counts = with_pct(
-        {
-            item["speaker__gender"] or "unknown": item["count"]
-            for item in gender_stats
-        }
+        {item["speaker__gender"] or "unknown": item["count"] for item in gender_stats}
     )
 
     level_stats = (
@@ -87,9 +84,7 @@ def with_pct(counts_dict):
     )
 
     speaker_level_stats = (
-        qs.values("speaker_level")
-        .annotate(count=Count("id"))
-        .order_by("speaker_level")
+        qs.values("speaker_level").annotate(count=Count("id")).order_by("speaker_level")
     )
     speaker_level_counts = with_pct(
         {item["speaker_level"]: item["count"] for item in speaker_level_stats}
@@ -455,6 +450,7 @@ def review_recap_compute_analysis_view(self, request, review_session_id):
         conference = review_session.conference
         accepted_submissions = list(self._get_accepted_submissions(conference))
         force_recompute = request.GET.get("recompute") == "1"
+        check_only = request.GET.get("check") == "1"
 
         from django.core.cache import cache
 
@@ -471,6 +467,9 @@ def review_recap_compute_analysis_view(self, request, review_session_id):
             if cached_result is not None:
                 return JsonResponse(cached_result)
 
+        if check_only:
+            return JsonResponse({"status": "empty"})
+
         # Use cache.add as a lock to prevent duplicate task dispatch.
         # Short TTL so lock auto-expires if the worker is killed before cleanup.
         computing_key = f"{combined_cache_key}:computing"

backend/reviews/templates/reviews-recap.html

@@ -682,6 +682,22 @@ <h2 class="recap-section-title">🔗 Similar Talks</h2>
 
   btn.addEventListener('click', function() { fetchAnalysis(false); });
   recomputeBtn.addEventListener('click', function() { fetchAnalysis(true); });
+
+  // Auto-load cached results on page load (check only, don't trigger computation)
+  fetch(computeUrl + '?check=1', {
+    headers: { 'X-Requested-With': 'XMLHttpRequest' }
+  })
+  .then(function(response) {
+    if (!response.ok) return;
+    return response.json();
+  })
+  .then(function(data) {
+    if (!data || data.status === 'processing' || data.status === 'error' || data.status === 'empty') return;
+    handleResult(data);
+  })
+  .catch(function() {
+    // Silently ignore errors on auto-load
+  });
 })();
 </script>
 {% endblock %}

backend/reviews/tests/test_recap.py

@@ -174,10 +174,7 @@ def test_recap_view_redirects_when_shortlist_not_visible(rf, mocker):
     response = admin.review_recap_view(request, review_session.id)
 
     assert response.status_code == 302
-    assert (
-        response.url
-        == f"/admin/reviews/reviewsession/{review_session.id}/change/"
-    )
+    assert response.url == f"/admin/reviews/reviewsession/{review_session.id}/change/"
 
 
 # --- review_recap_compute_analysis_view tests ---
@@ -194,9 +191,7 @@ def cache_get_side_effect(key):
     mock_cache_get = mocker.patch(
         "django.core.cache.cache.get", side_effect=cache_get_side_effect
     )
-    mock_cache_add = mocker.patch(
-        "django.core.cache.cache.add", return_value=True
-    )
+    mock_cache_add = mocker.patch("django.core.cache.cache.add", return_value=True)
     mocker.patch("django.core.cache.cache.set")
     mocker.patch("django.core.cache.cache.delete")
     mock_task = mocker.patch("reviews.tasks.compute_recap_analysis.apply_async")
@@ -573,6 +568,56 @@ def cache_get_side_effect(key):
     mock_check.assert_not_called()
 
 
+def test_compute_analysis_view_check_only_returns_empty_on_cache_miss(rf, mocker):
+    user, conference, review_session, submissions = _create_recap_setup()
+
+    _, _, mock_task, mock_check = _mock_analysis_deps(mocker, cache_return=None)
+
+    request = rf.get("/?check=1")
+    request.user = user
+
+    admin = ReviewSessionAdmin(ReviewSession, AdminSite())
+    response = admin.review_recap_compute_analysis_view(request, review_session.id)
+
+    data = json.loads(response.content)
+    assert data == {"status": "empty"}
+
+    # Task should NOT be dispatched in check-only mode
+    mock_task.assert_not_called()
+    mock_check.assert_not_called()
+
+
+def test_compute_analysis_view_check_only_returns_cached_result(rf, mocker):
+    user, conference, review_session, submissions = _create_recap_setup()
+    sub1, sub2 = submissions
+
+    cached_data = {
+        "submissions_list": [
+            {
+                "id": sub1.id,
+                "title": str(sub1.title),
+                "type": sub1.type.name,
+                "speaker": sub1.speaker.display_name,
+                "similar": [],
+            },
+        ],
+        "topic_clusters": {"topics": [], "outliers": [], "submission_topics": {}},
+    }
+
+    _, _, mock_task, _ = _mock_analysis_deps(mocker, cache_return=cached_data)
+
+    request = rf.get("/?check=1")
+    request.user = user
+
+    admin = ReviewSessionAdmin(ReviewSession, AdminSite())
+    response = admin.review_recap_compute_analysis_view(request, review_session.id)
+
+    data = json.loads(response.content)
+    assert data["submissions_list"][0]["id"] == sub1.id
+
+    mock_task.assert_not_called()
+
+
 def test_error_cache_ttl_is_shorter_than_result_ttl():
     from reviews.tasks import ERROR_CACHE_TTL, RESULT_CACHE_TTL