mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-07-13 00:26:10 +00:00
Handle attaching files at vector store creation
This wires up the `file_ids` from the vector store create request to actually attach all those files at creation time. This also required smarter handling of the file_ids and file_count metadata handling to ensure we update the in-memory cache and persistent representation of those as we attach those files. And, expand that logic to handle errors during file attachment to persist the failed status. Signed-off-by: Ben Browning <bbrownin@redhat.com>
This commit is contained in:
parent
459d50a365
commit
a2f0f608db
3 changed files with 97 additions and 42 deletions
|
@ -660,7 +660,6 @@ class SQLiteVecVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtoc
|
|||
(store_id, file_id),
|
||||
)
|
||||
row = cur.fetchone()
|
||||
print(f"!!! row is {row}")
|
||||
if row is None:
|
||||
return None
|
||||
(metadata,) = row
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue