Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Paddle-Pipelines] update faiss #7793

Merged
merged 3 commits into from
Jan 8, 2024
Merged

[Paddle-Pipelines] update faiss #7793

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
8b2ce6a
update faiss
qingzhong1 Jan 8, 2024
66a0c03
update faiss
qingzhong1 Jan 8, 2024
a7abd6d
update faiss
qingzhong1 Jan 8, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 1 addition & 5 deletions pipelines/pipelines/document_stores/faiss.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -391,7 +391,7 @@ def update_embeddings(

vector_id_map = {}
for doc in document_batch:
vector_id_map[str(doc.id)] = str(vector_id)
vector_id_map[str(doc.id)] = str(vector_id) + "_" + index
vector_id += 1
self.update_vector_ids(vector_id_map, index=index)
progress_bar.set_description_str("Documents Processed")
Expand Down Expand Up @@ -443,7 +443,6 @@ def get_all_documents_generator(
)
if return_embedding is None:
return_embedding = self.return_embedding

for doc in documents:
if return_embedding:
if doc.meta and doc.meta.get("vector_id") is not None:
Expand Down Expand Up @@ -588,7 +587,6 @@ def query_by_embedding(

if filters:
logger.warning("Query filters are not implemented for the FAISSDocumentStore.")

index = index or self.index
if not self.faiss_indexes.get(index):
raise Exception(f"Index named '{index}' does not exists. Use 'update_embeddings()' to create an index.")
Expand All @@ -599,11 +597,9 @@ def query_by_embedding(
query_emb = query_emb.reshape(1, -1).astype(np.float32)
if self.similarity == "cosine":
self.normalize_embedding(query_emb)

score_matrix, vector_id_matrix = self.faiss_indexes[index].search(query_emb, top_k)
vector_ids_for_query = [str(vector_id) + "_" + index for vector_id in vector_id_matrix[0] if vector_id != -1]
documents = self.get_documents_by_vector_ids(vector_ids_for_query, index=index)

# assign query score to each document
scores_for_vector_ids: Dict[str, float] = {
str(v_id): s for v_id, s in zip(vector_id_matrix[0], score_matrix[0])
Expand Down
3 changes: 0 additions & 3 deletions pipelines/pipelines/document_stores/sql.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -216,15 +216,13 @@ def get_documents_by_vector_ids(
):
"""Fetch documents by specifying a list of text vector id strings"""
index = index or self.index

documents = []
for i in range(0, len(vector_ids), batch_size):
query = self.session.query(DocumentORM).filter(
DocumentORM.vector_id.in_(vector_ids[i : i + batch_size]), DocumentORM.index == index
)
for row in query.all():
documents.append(self._convert_sql_row_to_document(row))

sorted_documents = sorted(documents, key=lambda doc: vector_ids.index(doc.meta["vector_id"]))
return sorted_documents

Expand Down Expand Up @@ -405,7 +403,6 @@ def write_documents(
document_objects = [Document.from_dict(d) if isinstance(d, dict) else d for d in documents]
else:
document_objects = documents

document_objects = self._handle_duplicate_documents(
documents=document_objects, index=index, duplicate_documents=duplicate_documents
)
Expand Down