Enable daily cron for search engine population and make delete operations non-blocking (#765)

Uncomment the cron schedule so the search engine is populated daily at 07:05 UTC.
Remove the `wait_for_task_completion` decorator from Meilisearch delete functions
(`delete_embedding_db`, `clear_embedding_db`, `delete_documents_from_db`) so they
fire-and-forget instead of blocking until the task finishes.

Made-with: Cursor

Author: mishig25

.github/workflows/populate_search_engine.yml

@@ -1,8 +1,8 @@
 name: Populate Search Engine
 
 on:
-  # schedule:
-  #   - cron: "5 7 * * *" # every day at 07:05
+  schedule:
+    - cron: "5 7 * * *" # every day at 07:05
   # to run this workflow manually from the Actions tab
   workflow_dispatch:
 

src/doc_builder/meilisearch_helper.py

@@ -161,19 +161,15 @@ def update_db_settings(client: Client, index_name: str):
     return client, task_info
 
 
-@wait_for_task_completion
 def delete_embedding_db(client: Client, index_name: str):
     index = client.index(index_name)
-    task_info = index.delete()
-    return client, task_info
+    index.delete()
 
 
-@wait_for_task_completion
 def clear_embedding_db(client: Client, index_name: str):
     """Delete all documents from an index without deleting the index itself."""
     index = client.index(index_name)
-    task_info = index.delete_all_documents()
-    return client, task_info
+    index.delete_all_documents()
 
 
 def get_all_document_ids(client: Client, index_name: str) -> set[str]:
@@ -199,12 +195,10 @@ def get_all_document_ids(client: Client, index_name: str) -> set[str]:
     return all_ids
 
 
-@wait_for_task_completion
 def delete_documents_from_db(client: Client, index_name: str, doc_ids: list[str]):
     """Delete a batch of documents by ID from a Meilisearch index."""
     index = client.index(index_name)
-    task_info = index.delete_documents(doc_ids)
-    return client, task_info
+    index.delete_documents(doc_ids)
 
 
 def sanitize_for_id(text):