Fix (#228)

_index.py

@@ -1,29 +1,14 @@
 from __future__ import annotations
 
-from typing import (
-    Callable,
-    Iterable,
-    Optional,
-    Sequence,
-    Union,
-    Literal,
-    cast,
-)
-
-from langchain_core.documents import Document
-from langchain_core.vectorstores import VectorStore
+from typing import Callable, Iterable, Literal, Optional, Sequence, Union, cast
 
 from langchain.document_loaders.base import BaseLoader
+from langchain.indexes._api import (IndexingResult, _batch,
+                                    _deduplicate_in_order,
+                                    _get_source_id_assigner, _HashedDocument)
 from langchain.indexes.base import RecordManager
-from langchain.indexes._api import (
-    IndexingResult,
-    _get_source_id_assigner,
-    _deduplicate_in_order,
-    _HashedDocument,
-    _batch,
-)
-
-
+from langchain.schema.document import Document
+from langchain.schema.vectorstore import VectorStore
 
 
 def index(

ingest.py

@@ -5,16 +5,18 @@
 from parser import langchain_docs_extractor
 
 import weaviate
-from _index import index
 from bs4 import BeautifulSoup, SoupStrainer
-from chain import get_embeddings_model
-from constants import WEAVIATE_DOCS_INDEX_NAME
 from langchain.document_loaders import RecursiveUrlLoader, SitemapLoader
 from langchain.indexes import SQLRecordManager
 from langchain.text_splitter import RecursiveCharacterTextSplitter
-from langchain.utils.html import PREFIXES_TO_IGNORE_REGEX, SUFFIXES_TO_IGNORE_REGEX
+from langchain.utils.html import (PREFIXES_TO_IGNORE_REGEX,
+                                  SUFFIXES_TO_IGNORE_REGEX)
 from langchain.vectorstores.weaviate import Weaviate
 
+from _index import index
+from chain import get_embeddings_model
+from constants import WEAVIATE_DOCS_INDEX_NAME
+
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
 
@@ -144,10 +146,10 @@ def ingest_docs():
         force_update=(os.environ.get("FORCE_UPDATE") or "false").lower() == "true",
     )
 
-    logger.info("Indexing stats: ", indexing_stats)
+    logger.info(f"Indexing stats: {indexing_stats}")
+    num_vecs = client.query.aggregate(WEAVIATE_DOCS_INDEX_NAME).with_meta_count().do()
     logger.info(
-        "LangChain now has this many vectors: ",
-        client.query.aggregate(WEAVIATE_DOCS_INDEX_NAME).with_meta_count().do(),
+        f"LangChain now has this many vectors: {num_vecs}",
     )