Skip to content

fix: add empty judgement to index search #1533

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
jupyterjazz merged 3 commits into from
May 15, 2023
Merged

fix: add empty judgement to index search #1533

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
aa9459c
fix: find return when index is empty
AnneYang720 May 12, 2023
ebcf109
fix: hnsw find
jupyterjazz May 15, 2023
da22587
refactor: if statement
jupyterjazz May 15, 2023
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: 6 additions & 0 deletions docarray/index/backends/hnswlib.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -280,6 +280,9 @@ def _find_batched(
limit: int,
search_field: str = '',
) -> _FindResultBatched:
if not self.num_docs():
Comment thread
jupyterjazz marked this conversation as resolved.
Outdated
return _FindResultBatched(documents=[], scores=[])

index = self._hnsw_indices[search_field]
labels, distances = index.knn_query(queries, k=limit)
result_das = [
Expand All @@ -293,6 +296,9 @@ def _find_batched(
def _find(
self, query: np.ndarray, limit: int, search_field: str = ''
) -> _FindResult:
if not self.num_docs():
return _FindResult(documents=[], scores=[]) # type: ignore

query_batched = np.expand_dims(query, axis=0)
docs, scores = self._find_batched(
queries=query_batched, limit=limit, search_field=search_field
Expand Down
8 changes: 8 additions & 0 deletions docarray/index/backends/in_memory.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -193,6 +193,10 @@ def find(
"""
self._logger.debug(f'Executing `find` for search field {search_field}')
self._validate_search_field(search_field)

if not self.num_docs():
Comment thread
jupyterjazz marked this conversation as resolved.
Outdated
return FindResult(documents=[], scores=[]) # type: ignore

config = self._column_infos[search_field].config

docs, scores = find(
Expand Down Expand Up @@ -233,6 +237,10 @@ def find_batched(
"""
self._logger.debug(f'Executing `find_batched` for search field {search_field}')
self._validate_search_field(search_field)

if not self.num_docs():
return FindResultBatched(documents=[], scores=[])

config = self._column_infos[search_field].config

find_res = find_batched(
Expand Down
9 changes: 9 additions & 0 deletions tests/index/hnswlib/test_find.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,6 +54,15 @@ class SimpleSchema(BaseDoc):
assert np.allclose(result.tens, np.zeros(10))


def test_find_empty_index(tmp_path):
empty_index = HnswDocumentIndex[SimpleDoc](work_dir=str(tmp_path))
query = SimpleDoc(tens=np.ones(10))

docs, scores = empty_index.find(query, search_field='tens', limit=5)
assert len(docs) == 0
assert len(scores) == 0


@pytest.mark.parametrize('space', ['cosine', 'l2', 'ip'])
def test_find_torch(tmp_path, space):
index = HnswDocumentIndex[TorchDoc](work_dir=str(tmp_path))
Expand Down
10 changes: 10 additions & 0 deletions tests/index/in_memory/test_in_memory.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,6 +70,11 @@ class MyDoc(BaseDoc):
assert len(scores) == 5
assert doc_index.num_docs() == 10

empty_index = InMemoryExactNNIndex[MyDoc]()
docs, scores = empty_index.find(query, search_field='tensor', limit=5)
assert len(docs) == 0
assert len(scores) == 0


@pytest.mark.parametrize('space', ['cosine_sim', 'euclidean_dist', 'sqeuclidean_dist'])
@pytest.mark.parametrize('is_query_doc', [True, False])
Expand All @@ -96,6 +101,11 @@ class MyDoc(BaseDoc):
assert len(result) == 5
assert doc_index.num_docs() == 10

empty_index = InMemoryExactNNIndex[MyDoc]()
docs, scores = empty_index.find_batched(query, search_field='tensor', limit=5)
assert len(docs) == 0
assert len(scores) == 0


def test_concatenated_queries(doc_index):
query = SchemaDoc(text='query', price=0, tensor=np.ones(10))
Expand Down