INTPYTHON-447 Ensure index is ready for retriever tests (#25)

Author: blink1073

libs/mongodb/langchain_mongodb/index.py

@@ -170,7 +170,7 @@ def _is_index_ready(collection: Collection, index_name: str) -> bool:
         bool : True if the index is present and READY false otherwise
     """
     for index in collection.list_search_indexes(index_name):
-        if index["type"] == "vectorSearch" and index["status"] == "READY":
+        if index["status"] == "READY":
             return True
     return False
 

libs/mongodb/tests/integration_tests/test_retrievers.py

@@ -1,5 +1,6 @@
 import os
 from typing import Generator, List
+from time import sleep, time
 
 import pytest
 from langchain_core.documents import Document
@@ -10,7 +11,7 @@
 from langchain_mongodb import MongoDBAtlasVectorSearch
 from langchain_mongodb.index import (
     create_fulltext_search_index,
-    create_vector_search_index,
+    create_vector_search_index
 )
 from langchain_mongodb.retrievers import (
     MongoDBAtlasFullTextSearchRetriever,
@@ -160,6 +161,18 @@ def test_fulltext_retriever(
         search_field=PAGE_CONTENT_FIELD,
     )
 
+    # Wait for the search index to complete.
+    search_content = dict(index=SEARCH_INDEX_NAME, wildcard=dict(query="*", path=PAGE_CONTENT_FIELD, allowAnalyzedField=True))
+    n_docs = collection.count_documents({})
+    t0 = time()
+    while True:
+        if (time() - t0) > TIMEOUT:
+            raise TimeoutError(f'Search index {SEARCH_INDEX_NAME} did not complete in {TIMEOUT}')
+        results = collection.aggregate([{ "$search": search_content }])
+        if len(list(results)) == n_docs:
+            break
+        sleep(INTERVAL)
+
     query = "When was the last time I visited new orleans?"
     results = retriever.invoke(query)
     assert "New Orleans" in results[0].page_content