highperfocused/open-webui
1
0
Fork 0
mirror of https://github.com/open-webui/open-webui.git synced 2025-04-21 22:14:47 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #12507 from Ithanil/fix_web_result_collection_source_ids

Browse Source 
fix: fix web results all getting the same source id when using embedding and retrieval
This commit is contained in:
Timothy Jaeryang Baek 2025-04-06 15:43:21 -07:00 committed by GitHub
commit 9825d03602
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 27 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
backend/open_webui/routers/retrieval.py
View File

@ -1464,11 +1464,9 @@ async def process_web_search(
log.debug(f"web_results: {web_results}")
try:
collection_name = form_data.collection_name
if collection_name == "" or collection_name is None:
collection_name = f"web-search-{calculate_sha256_string(form_data.query)}"[
:63
]
collection_basename = form_data.collection_name
if collection_basename == "" or collection_basename is None:
collection_basename = "web-search"
urls = [result.link for result in web_results]
loader = get_web_loader(
@ -1495,18 +1493,23 @@ async def process_web_search(
"loaded_count": len(docs),
}
else:
await run_in_threadpool(
save_docs_to_vector_db,
request,
docs,
collection_name,
overwrite=True,
user=user,
)
collection_names = []
for doc_idx, doc in enumerate(docs):
collection_sha = calculate_sha256_string(f"{form_data.query}-{urls[doc_idx]}")
doc_collection_name = f"{collection_basename}-{collection_sha}"[:63]
collection_names.append(doc_collection_name)
await run_in_threadpool(
save_docs_to_vector_db,
request,
[doc],
doc_collection_name,
overwrite=True,
user=user,
)
return {
"status": True,
"collection_name": collection_name,
"collection_names": collection_names,
"filenames": urls,
"loaded_count": len(docs),
}

19
backend/open_webui/utils/middleware.py
View File

@ -399,15 +399,16 @@ async def chat_web_search_handler(
all_results.append(results)
files = form_data.get("files", [])
if results.get("collection_name"):
files.append(
{
"collection_name": results["collection_name"],
"name": searchQuery,
"type": "web_search",
"urls": results["filenames"],
}
)
if results.get("collection_names"):
for col_idx, collection_name in enumerate(results.get("collection_names")):
files.append(
{
"collection_name": collection_name,
"name": searchQuery,
"type": "web_search",
"urls": [results["filenames"][col_idx]],
}
)
elif results.get("docs"):
# Invoked when bypass embedding and retrieval is set to True
docs = results["docs"]