Skip to content

fix: support torch.Tensor and other embeddings in match #158

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

Merged
hanxiao merged 4 commits into from
Feb 28, 2022
Merged

fix: support torch.Tensor and other embeddings in match #158

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
083218a
fix: implement private _find method
alaeddine-13 Feb 28, 2022
ca25d29
fix: support torch.Tensor in find
alaeddine-13 Feb 28, 2022
2b35be1
test: cover torch.Tensor and qdrant in test_match
alaeddine-13 Feb 28, 2022
47638f9
fix: align qdrant _find interface with other storage backends
alaeddine-13 Feb 28, 2022
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
12 changes: 6 additions & 6 deletions docarray/array/storage/qdrant/find.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -74,21 +74,21 @@ def _find_similar_vectors(self, q: 'QdrantArrayType', limit=10):

return DocumentArray(docs)

def find(
self, query: 'QdrantArrayType', limit: int = 10
) -> Union['DocumentArray', List['DocumentArray']]:
def _find(
self, query: 'QdrantArrayType', limit: int = 10, **kwargs
) -> List['DocumentArray']:
"""Returns approximate nearest neighbors given a batch of input queries.
:param query: input supported to be used in Qdrant.
:param limit: number of retrieved items

:return: DocumentArray containing the closest documents to the query if it is a single query, otherwise a list of DocumentArrays containing
the closest Document objects for each of the queries in `query`.

:return: a list of DocumentArrays containing the closest Document objects for each of the queries in `query`.
"""

num_rows, _ = ndarray.get_array_rows(query)

if num_rows == 1:
return self._find_similar_vectors(query, limit=limit)
return [self._find_similar_vectors(query, limit=limit)]
else:
closest_docs = []
for q in query:
Expand Down
3 changes: 2 additions & 1 deletion docarray/array/storage/weaviate/find.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@
from .... import Document, DocumentArray
from ....math import ndarray
from ....math.helper import EPSILON
from ....math.ndarray import to_numpy_array
from ....score import NamedScore

if TYPE_CHECKING:
Expand All @@ -27,7 +28,7 @@

class FindMixin:
def _find_similar_vectors(self, query: 'WeaviateArrayType', limit=10):

query = to_numpy_array(query)
is_all_zero = np.all(query == 0)
if is_all_zero:
query = query + EPSILON
Expand Down
13 changes: 11 additions & 2 deletions tests/unit/array/mixins/test_match.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,12 +71,21 @@ def doc_lists_to_doc_arrays(doc_lists, *args, **kwargs):

@pytest.mark.parametrize(
'storage, config',
[('weaviate', WeaviateConfig(3)), ('pqlite', {'n_dim': 3})],
[('weaviate', {'n_dim': 3}), ('pqlite', {'n_dim': 3}), ('qdrant', {'n_dim': 3})],
)
@pytest.mark.parametrize('limit', [1, 2, 3])
@pytest.mark.parametrize('exclude_self', [True, False])
def test_match(storage, config, doc_lists, limit, exclude_self, start_storage):
@pytest.mark.parametrize('as_tensor', [True, False])
def test_match(
storage, config, doc_lists, limit, exclude_self, start_storage, as_tensor
):
D1, D2 = doc_lists_to_doc_arrays(doc_lists)
if as_tensor:
for d in D1:
d.embedding = torch.from_numpy(d.embedding)

for d in D2:
d.embedding = torch.from_numpy(d.embedding)

if config:
da = DocumentArray(D2, storage=storage, config=config)
Expand Down