Merge pull request #168 from ufal/add-caching

Fix output management
ufal · Dec 9, 2024 · 9ca7b44 · 9ca7b44
2 parents 794e564 + d44d43d
commit 9ca7b44
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 4 deletions.
diff --git a/factgenie/static/js/manage.js b/factgenie/static/js/manage.js
@@ -74,7 +74,7 @@ function deleteOutput(dataset, split, setup_id) {
                 alert(response.error);
             } else {
                 // reload
-                window.location.hash = "#local";
+                window.location.hash = "#outputs";
                 location.reload();
             }
         }

diff --git a/factgenie/templates/pages/manage.html b/factgenie/templates/pages/manage.html
@@ -172,7 +172,7 @@ <h3><img src="{{ host_prefix }}/static/img/manage.png" class="heading-img-inline
                   class="btn btn-outline-secondary" data-bs-toggle="tooltip" title="Export model outputs">
                   <i class="fa fa-download"></i>
                 </a>
-                <a onclick="deleteOutput('{{ dataset }}', '{{ split }}', '{{ setup_id }}')"
+                <a onclick="deleteOutput('{{ record.dataset }}', '{{ record.split }}', '{{ record.setup_id }}')"
                   class="btn btn-outline-danger" data-bs-toggle="tooltip" title="Delete the output">
                   <i class="fa fa-trash"></i>
                 </a>

diff --git a/factgenie/workflows.py b/factgenie/workflows.py
@@ -338,6 +338,7 @@ def load_outputs_from_file(file_path, cols):
 
                 # drop any keys that are not in the key set
                 j = {k: v for k, v in j.items() if k in cols}
+                j["jsonl_file"] = file_path
                 outputs.append(j)
             except Exception as e:
                 logger.error(
@@ -347,6 +348,13 @@ def load_outputs_from_file(file_path, cols):
     return outputs
 
 
+def remove_outputs(app, file_path):
+    """Remove outputs from the output index for a specific file"""
+    if app.db["output_index"] is not None:
+        # Filter out outputs from the specified file
+        app.db["output_index"] = app.db["output_index"][app.db["output_index"]["jsonl_file"] != file_path]
+
+
 def get_output_index(app, force_reload=True):
     if hasattr(app, "db") and app.db["output_index"] is not None and not force_reload:
         return app.db["output_index"]
@@ -362,14 +370,14 @@ def get_output_index(app, force_reload=True):
     # Handle modified files
     for file_path, mod_time in current_outs.items():
         if file_path not in cached_outs or cached_outs[file_path] < mod_time:
+            remove_outputs(app, file_path)
             new_outputs.extend(load_outputs_from_file(file_path, cols))
 
     # Handle deleted files
     for file_path in set(cached_outs.keys()) - set(current_outs.keys()):
         # Remove outputs for deleted files from the index
         if app.db["output_index"] is not None:
-            file_mask = app.db["output_index"]["file_path"] == file_path
-            app.db["output_index"] = app.db["output_index"][~file_mask]
+            remove_outputs(app, file_path)
 
     # Update the cache
     app.db["output_index_cache"] = current_outs