sentence-transformers-haystack 0.1.1__tar.gz → 0.2.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- sentence_transformers_haystack-0.2.0/CHANGELOG.md +33 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/PKG-INFO +5 -3
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/README.md +2 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/pyproject.toml +2 -1
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/embedding_backend/backend.py +5 -1
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/sentence_transformers_doc_image_embedder.py +4 -2
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/sentence_transformers_document_embedder.py +4 -2
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/sentence_transformers_sparse_document_embedder.py +4 -2
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/sentence_transformers_sparse_text_embedder.py +4 -2
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/sentence_transformers_text_embedder.py +4 -2
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/rankers/sentence_transformers/sentence_transformers_diversity.py +4 -2
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/rankers/sentence_transformers/sentence_transformers_similarity.py +9 -7
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_similarity.py +29 -0
- sentence_transformers_haystack-0.1.1/CHANGELOG.md +0 -9
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/.gitignore +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/LICENSE.txt +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/pydoc/config_docusaurus.yml +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/py.typed +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/__init__.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/embedding_backend/__init__.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/embedders/sentence_transformers/embedding_backend/sparse_backend.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/rankers/py.typed +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/src/haystack_integrations/components/rankers/sentence_transformers/__init__.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/__init__.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/conftest.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_files/images/apple.jpg +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_files/images/haystack-logo.png +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_files/images/image_metadata.png +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_files/pdf/sample_pdf_1.pdf +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_hybrid_rag_e2e.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_diversity.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_doc_image_embedder.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_document_embedder.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_embedding_backend.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_sparse_document_embedder.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_sparse_embedding_backend.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_sparse_text_embedder.py +0 -0
- {sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/test_sentence_transformers_text_embedder.py +0 -0
|
@@ -0,0 +1,33 @@
|
|
|
1
|
+
# Changelog
|
|
2
|
+
|
|
3
|
+
## [integrations/sentence_transformers-v0.1.2] - 2026-09-23
|
|
4
|
+
|
|
5
|
+
### 🐛 Bug Fixes
|
|
6
|
+
|
|
7
|
+
- Fix new issues raised by ruff 0.16.0 (#3670)
|
|
8
|
+
- Sentence Transformers - fix types for 6.0 (#3822)
|
|
9
|
+
- Respect falsy run() overrides in `SentenceTransformersSimilarityRanker` (#3855)
|
|
10
|
+
|
|
11
|
+
### 🧹 Chores
|
|
12
|
+
|
|
13
|
+
- Re-standardize READMEs (#3768)
|
|
14
|
+
|
|
15
|
+
|
|
16
|
+
## [integrations/sentence_transformers-v0.1.1] - 2026-07-07
|
|
17
|
+
|
|
18
|
+
### 🐛 Bug Fixes
|
|
19
|
+
|
|
20
|
+
- Support quantization ranges for int8/uint8 sentence-transformers… (#3543)
|
|
21
|
+
|
|
22
|
+
### 🧪 Testing
|
|
23
|
+
|
|
24
|
+
- Force Transformers and Sentence Transformers integration tests to run on CPU (#3550)
|
|
25
|
+
|
|
26
|
+
|
|
27
|
+
## [integrations/sentence_transformers-v0.1.0] - 2026-06-16
|
|
28
|
+
|
|
29
|
+
### 🚀 Features
|
|
30
|
+
|
|
31
|
+
- Move Sentence Transformers from Haystack (#3443)
|
|
32
|
+
|
|
33
|
+
<!-- generated by git-cliff -->
|
|
@@ -1,6 +1,6 @@
|
|
|
1
|
-
Metadata-Version: 2.
|
|
1
|
+
Metadata-Version: 2.5
|
|
2
2
|
Name: sentence-transformers-haystack
|
|
3
|
-
Version: 0.
|
|
3
|
+
Version: 0.2.0
|
|
4
4
|
Summary: Haystack integration for sentence_transformers
|
|
5
5
|
Project-URL: Documentation, https://github.com/deepset-ai/haystack-core-integrations/tree/main/integrations/sentence_transformers#readme
|
|
6
6
|
Project-URL: Issues, https://github.com/deepset-ai/haystack-core-integrations/issues
|
|
@@ -23,7 +23,7 @@ Requires-Python: >=3.10
|
|
|
23
23
|
Requires-Dist: haystack-ai>=2.24.0
|
|
24
24
|
Requires-Dist: pillow>=11.3.0
|
|
25
25
|
Requires-Dist: pypdfium2>=4.0.0
|
|
26
|
-
Requires-Dist: sentence-transformers>=5.
|
|
26
|
+
Requires-Dist: sentence-transformers>=5.6.0
|
|
27
27
|
Description-Content-Type: text/markdown
|
|
28
28
|
|
|
29
29
|
# sentence-transformers-haystack
|
|
@@ -39,3 +39,5 @@ Description-Content-Type: text/markdown
|
|
|
39
39
|
## Contributing
|
|
40
40
|
|
|
41
41
|
Refer to the general [Contribution Guidelines](https://github.com/deepset-ai/haystack-core-integrations/blob/main/CONTRIBUTING.md).
|
|
42
|
+
|
|
43
|
+
To run integration tests locally, you need to export the `OPENAI_API_KEY` environment variable.
|
{sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/pyproject.toml
RENAMED
|
@@ -23,7 +23,7 @@ classifiers = [
|
|
|
23
23
|
"Programming Language :: Python :: Implementation :: CPython",
|
|
24
24
|
"Programming Language :: Python :: Implementation :: PyPy",
|
|
25
25
|
]
|
|
26
|
-
dependencies = ["haystack-ai>=2.24.0", "sentence-transformers>=5.
|
|
26
|
+
dependencies = ["haystack-ai>=2.24.0", "sentence-transformers>=5.6.0", "pillow>=11.3.0", "pypdfium2>=4.0.0"]
|
|
27
27
|
|
|
28
28
|
[project.urls]
|
|
29
29
|
Documentation = "https://github.com/deepset-ai/haystack-core-integrations/tree/main/integrations/sentence_transformers#readme"
|
|
@@ -135,6 +135,7 @@ ignore = [
|
|
|
135
135
|
"PLR0912",
|
|
136
136
|
"PLR0913",
|
|
137
137
|
"PLR0915",
|
|
138
|
+
"PLR0917",
|
|
138
139
|
# Allow `Any` type - used legitimately for dynamic types and SDK boundaries
|
|
139
140
|
"ANN401",
|
|
140
141
|
]
|
|
@@ -112,5 +112,9 @@ class _SentenceTransformersEmbeddingBackend:
|
|
|
112
112
|
if quantization_ranges is not None and precision in ("int8", "uint8"):
|
|
113
113
|
kwargs["precision"] = "float32"
|
|
114
114
|
embeddings = self.model.encode(data, **kwargs)
|
|
115
|
-
|
|
115
|
+
quantized = quantize_embeddings(embeddings, precision=precision, ranges=np.asarray(quantization_ranges))
|
|
116
|
+
if isinstance(quantized, list):
|
|
117
|
+
msg = "Unexpected multi-vector output from quantize_embeddings"
|
|
118
|
+
raise TypeError(msg)
|
|
119
|
+
return quantized.tolist()
|
|
116
120
|
return self.model.encode(data, **kwargs).tolist()
|
|
@@ -114,8 +114,10 @@ class SentenceTransformersDocumentImageEmbedder:
|
|
|
114
114
|
Additional keyword arguments for `AutoModelForSequenceClassification.from_pretrained`
|
|
115
115
|
when loading the model. Refer to specific model documentation for available kwargs.
|
|
116
116
|
:param tokenizer_kwargs:
|
|
117
|
-
Additional keyword arguments
|
|
118
|
-
|
|
117
|
+
Additional keyword arguments forwarded as `processor_kwargs` to the Sentence Transformers model
|
|
118
|
+
when loading it. Refer to the
|
|
119
|
+
[Sentence Transformers documentation](https://sbert.net/docs/package_reference/sentence_transformer/model.html)
|
|
120
|
+
for available kwargs.
|
|
119
121
|
:param config_kwargs:
|
|
120
122
|
Additional keyword arguments for `AutoConfig.from_pretrained` when loading the model configuration.
|
|
121
123
|
:param precision:
|
|
@@ -107,8 +107,10 @@ class SentenceTransformersDocumentEmbedder:
|
|
|
107
107
|
Additional keyword arguments for `AutoModelForSequenceClassification.from_pretrained`
|
|
108
108
|
when loading the model. Refer to specific model documentation for available kwargs.
|
|
109
109
|
:param tokenizer_kwargs:
|
|
110
|
-
Additional keyword arguments
|
|
111
|
-
|
|
110
|
+
Additional keyword arguments forwarded as `processor_kwargs` to the Sentence Transformers model
|
|
111
|
+
when loading it. Refer to the
|
|
112
|
+
[Sentence Transformers documentation](https://sbert.net/docs/package_reference/sentence_transformer/model.html)
|
|
113
|
+
for available kwargs.
|
|
112
114
|
:param config_kwargs:
|
|
113
115
|
Additional keyword arguments for `AutoConfig.from_pretrained` when loading the model configuration.
|
|
114
116
|
:param precision:
|
|
@@ -95,8 +95,10 @@ class SentenceTransformersSparseDocumentEmbedder:
|
|
|
95
95
|
Additional keyword arguments for `AutoModelForSequenceClassification.from_pretrained`
|
|
96
96
|
when loading the model. Refer to specific model documentation for available kwargs.
|
|
97
97
|
:param tokenizer_kwargs:
|
|
98
|
-
Additional keyword arguments
|
|
99
|
-
|
|
98
|
+
Additional keyword arguments forwarded as `processor_kwargs` to the Sentence Transformers model
|
|
99
|
+
when loading it. Refer to the
|
|
100
|
+
[Sentence Transformers documentation](https://sbert.net/docs/package_reference/sentence_transformer/model.html)
|
|
101
|
+
for available kwargs.
|
|
100
102
|
:param config_kwargs:
|
|
101
103
|
Additional keyword arguments for `AutoConfig.from_pretrained` when loading the model configuration.
|
|
102
104
|
:param backend:
|
|
@@ -76,8 +76,10 @@ class SentenceTransformersSparseTextEmbedder:
|
|
|
76
76
|
Additional keyword arguments for `AutoModelForSequenceClassification.from_pretrained`
|
|
77
77
|
when loading the model. Refer to specific model documentation for available kwargs.
|
|
78
78
|
:param tokenizer_kwargs:
|
|
79
|
-
Additional keyword arguments
|
|
80
|
-
|
|
79
|
+
Additional keyword arguments forwarded as `processor_kwargs` to the Sentence Transformers model
|
|
80
|
+
when loading it. Refer to the
|
|
81
|
+
[Sentence Transformers documentation](https://sbert.net/docs/package_reference/sentence_transformer/model.html)
|
|
82
|
+
for available kwargs.
|
|
81
83
|
:param config_kwargs:
|
|
82
84
|
Additional keyword arguments for `AutoConfig.from_pretrained` when loading the model configuration.
|
|
83
85
|
:param backend:
|
|
@@ -97,8 +97,10 @@ class SentenceTransformersTextEmbedder:
|
|
|
97
97
|
Additional keyword arguments for `AutoModelForSequenceClassification.from_pretrained`
|
|
98
98
|
when loading the model. Refer to specific model documentation for available kwargs.
|
|
99
99
|
:param tokenizer_kwargs:
|
|
100
|
-
Additional keyword arguments
|
|
101
|
-
|
|
100
|
+
Additional keyword arguments forwarded as `processor_kwargs` to the Sentence Transformers model
|
|
101
|
+
when loading it. Refer to the
|
|
102
|
+
[Sentence Transformers documentation](https://sbert.net/docs/package_reference/sentence_transformer/model.html)
|
|
103
|
+
for available kwargs.
|
|
102
104
|
:param config_kwargs:
|
|
103
105
|
Additional keyword arguments for `AutoConfig.from_pretrained` when loading the model configuration.
|
|
104
106
|
:param precision:
|
|
@@ -163,8 +163,10 @@ class SentenceTransformersDiversityRanker:
|
|
|
163
163
|
Additional keyword arguments for `AutoModelForSequenceClassification.from_pretrained`
|
|
164
164
|
when loading the model. Refer to specific model documentation for available kwargs.
|
|
165
165
|
:param tokenizer_kwargs:
|
|
166
|
-
Additional keyword arguments
|
|
167
|
-
|
|
166
|
+
Additional keyword arguments forwarded as `processor_kwargs` to the Sentence Transformers model
|
|
167
|
+
when loading it. Refer to the
|
|
168
|
+
[Sentence Transformers documentation](https://sbert.net/docs/package_reference/sentence_transformer/model.html)
|
|
169
|
+
for available kwargs.
|
|
168
170
|
:param config_kwargs:
|
|
169
171
|
Additional keyword arguments for `AutoConfig.from_pretrained` when loading the model configuration.
|
|
170
172
|
:param backend:
|
|
@@ -98,8 +98,10 @@ class SentenceTransformersSimilarityRanker:
|
|
|
98
98
|
Additional keyword arguments for `AutoModelForSequenceClassification.from_pretrained`
|
|
99
99
|
when loading the model. Refer to specific model documentation for available kwargs.
|
|
100
100
|
:param tokenizer_kwargs:
|
|
101
|
-
Additional keyword arguments
|
|
102
|
-
|
|
101
|
+
Additional keyword arguments forwarded as `processor_kwargs` to the Sentence Transformers model
|
|
102
|
+
when loading it. Refer to the
|
|
103
|
+
[Sentence Transformers documentation](https://sbert.net/docs/package_reference/sentence_transformer/model.html)
|
|
104
|
+
for available kwargs.
|
|
103
105
|
:param config_kwargs:
|
|
104
106
|
Additional keyword arguments for `AutoConfig.from_pretrained` when loading the model configuration.
|
|
105
107
|
:param backend:
|
|
@@ -250,14 +252,14 @@ class SentenceTransformersSimilarityRanker:
|
|
|
250
252
|
if not documents:
|
|
251
253
|
return {"documents": []}
|
|
252
254
|
|
|
253
|
-
top_k
|
|
254
|
-
scale_score = scale_score or self.scale_score
|
|
255
|
-
score_threshold = score_threshold or self.score_threshold
|
|
256
|
-
|
|
257
|
-
if top_k <= 0:
|
|
255
|
+
if top_k is not None and top_k <= 0:
|
|
258
256
|
msg = f"top_k must be > 0, but got {top_k}"
|
|
259
257
|
raise ValueError(msg)
|
|
260
258
|
|
|
259
|
+
top_k = self.top_k if top_k is None else top_k
|
|
260
|
+
scale_score = self.scale_score if scale_score is None else scale_score
|
|
261
|
+
score_threshold = self.score_threshold if score_threshold is None else score_threshold
|
|
262
|
+
|
|
261
263
|
deduplicated_documents = _deduplicate_documents(documents)
|
|
262
264
|
prepared_query = self.query_prefix + query + self.query_suffix
|
|
263
265
|
prepared_documents = []
|
|
@@ -281,6 +281,35 @@ class TestSentenceTransformersSimilarityRanker:
|
|
|
281
281
|
with pytest.raises(ValueError):
|
|
282
282
|
ranker.run(query="test", documents=[Document(content="document")], top_k=-1)
|
|
283
283
|
|
|
284
|
+
def test_run_top_k_zero_raises(self):
|
|
285
|
+
ranker = SentenceTransformersSimilarityRanker()
|
|
286
|
+
ranker._cross_encoder = MagicMock()
|
|
287
|
+
|
|
288
|
+
with pytest.raises(ValueError, match="top_k must be > 0, but got 0"):
|
|
289
|
+
ranker.run(query="test", documents=[Document(content="document")], top_k=0)
|
|
290
|
+
|
|
291
|
+
def test_run_scale_score_false_overrides_init(self):
|
|
292
|
+
ranker = SentenceTransformersSimilarityRanker() # scale_score=True by default
|
|
293
|
+
mock_cross_encoder = MagicMock()
|
|
294
|
+
mock_cross_encoder.rank.return_value = [{"corpus_id": 0, "score": 1.0}]
|
|
295
|
+
ranker._cross_encoder = mock_cross_encoder
|
|
296
|
+
|
|
297
|
+
ranker.run(query="test", documents=[Document(content="document")], scale_score=False)
|
|
298
|
+
|
|
299
|
+
_, kwargs = mock_cross_encoder.rank.call_args
|
|
300
|
+
assert isinstance(kwargs["activation_fn"], torch.nn.Identity)
|
|
301
|
+
|
|
302
|
+
def test_run_score_threshold_zero_overrides_init(self):
|
|
303
|
+
ranker = SentenceTransformersSimilarityRanker(score_threshold=0.5)
|
|
304
|
+
mock_cross_encoder = MagicMock()
|
|
305
|
+
mock_cross_encoder.rank.return_value = [{"corpus_id": 0, "score": 0.2}, {"corpus_id": 1, "score": -0.4}]
|
|
306
|
+
ranker._cross_encoder = mock_cross_encoder
|
|
307
|
+
|
|
308
|
+
documents = [Document(content="first"), Document(content="second")]
|
|
309
|
+
result = ranker.run(query="test", documents=documents, score_threshold=0.0)
|
|
310
|
+
|
|
311
|
+
assert [doc.score for doc in result["documents"]] == [0.2]
|
|
312
|
+
|
|
284
313
|
def test_returns_empty_list_if_no_documents_are_provided(self):
|
|
285
314
|
ranker = SentenceTransformersSimilarityRanker()
|
|
286
315
|
ranker._cross_encoder = MagicMock()
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
{sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/__init__.py
RENAMED
|
File without changes
|
{sentence_transformers_haystack-0.1.1 → sentence_transformers_haystack-0.2.0}/tests/conftest.py
RENAMED
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|