Upload app.py

metadata_search/app.py

@@ -119,6 +119,9 @@ UI_DIR = PROJECT_ROOT / "github_pages"
 
 @app.get("/api/test-summary")
 def test_summary():
+    import re
+    from huggingface_hub import hf_hub_download
+
     file_name = request.args.get("file", "").strip()
 
     if not file_name:
@@ -127,15 +130,44 @@ def test_summary():
             "error": "Missing ?file=..."
         }), 400
 
+    # First test the normal application path.
     path = summary_path(file_name)
     text = load_summary(file_name)
 
+    # If normal resolution fails, expose the exact cloud lookup attempt.
+    year_match = re.search(r"\\s-\\s(\\d{4})\\s-\\s", file_name)
+    year = year_match.group(1) if year_match else None
+    repo_id = os.getenv(
+        "WATERSCOPE_DATASET_REPO",
+        "EnvironmentalAI/WaterScopeAI-Data",
+    ).strip()
+    repo_file = f"Summary/{year}/{Path(file_name).name}" if year else None
+
+    direct_download_path = None
+    direct_download_error = None
+
+    if path is None and repo_file:
+        try:
+            direct_download_path = hf_hub_download(
+                repo_id=repo_id,
+                repo_type="dataset",
+                filename=repo_file,
+                local_dir="/tmp/waterscope-summary-debug",
+            )
+        except Exception as exc:
+            direct_download_error = f"{type(exc).__name__}: {exc}"
+
     return jsonify({
         "ok": path is not None and text is not None,
         "file_name": file_name,
+        "detected_year": year,
+        "repo_id": repo_id,
+        "repo_file_attempted": repo_file,
         "resolved_path": str(path) if path else None,
         "text_length": len(text) if text else 0,
         "preview": text[:500] if text else None,
+        "direct_download_path": direct_download_path,
+        "direct_download_error": direct_download_error,
     })