13972131258

Committed 20 Mar 2025 02:43PM UTC coverage: 90.021% (-0.03%) from 90.054%

Build # 13972131258

Build Type

Pull #9069

github

Committed by

web-flow

Commit Message

Merge 8371761b0 into 67ab3788e

Pull Request Pull Request #9069: refactor!: `ChatMessage` serialization-deserialization updates

Run Details

9833 of 10923 relevant lines covered (90.02%)

0.9 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

90.91

haystack/components/classifiers/zero_shot_document_classifier.py

# SPDX-FileCopyrightText: 2022-present deepset GmbH <info@deepset.ai>
#
# SPDX-License-Identifier: Apache-2.0

from typing import Any, Dict, List, Optional

from haystack import Document, component, default_from_dict, default_to_dict
from haystack.lazy_imports import LazyImport
from haystack.utils import ComponentDevice, Secret, deserialize_secrets_inplace
from haystack.utils.hf import deserialize_hf_model_kwargs, resolve_hf_pipeline_kwargs, serialize_hf_model_kwargs

with LazyImport(message="Run 'pip install transformers[torch,sentencepiece]'") as torch_and_transformers_import:
    from transformers import pipeline


@component
class TransformersZeroShotDocumentClassifier:
    """
    Performs zero-shot classification of documents based on given labels and adds the predicted label to their metadata.

    The component uses a Hugging Face pipeline for zero-shot classification.
    Provide the model and the set of labels to be used for categorization during initialization.
    Additionally, you can configure the component to allow multiple labels to be true.

    Classification is run on the document's content field by default. If you want it to run on another field, set the
    `classification_field` to one of the document's metadata fields.

    Available models for the task of zero-shot-classification include:
        - `valhalla/distilbart-mnli-12-3`
        - `cross-encoder/nli-distilroberta-base`
        - `cross-encoder/nli-deberta-v3-xsmall`

    ### Usage example

    The following is a pipeline that classifies documents based on predefined classification labels
    retrieved from a search pipeline:

    ```python
    from haystack import Document
    from haystack.components.retrievers.in_memory import InMemoryBM25Retriever
    from haystack.document_stores.in_memory import InMemoryDocumentStore
    from haystack.core.pipeline import Pipeline
    from haystack.components.classifiers import TransformersZeroShotDocumentClassifier

    documents = [Document(id="0", content="Today was a nice day!"),
                 Document(id="1", content="Yesterday was a bad day!")]

    document_store = InMemoryDocumentStore()
    retriever = InMemoryBM25Retriever(document_store=document_store)
    document_classifier = TransformersZeroShotDocumentClassifier(
        model="cross-encoder/nli-deberta-v3-xsmall",
        labels=["positive", "negative"],
    )

    document_store.write_documents(documents)

    pipeline = Pipeline()
    pipeline.add_component(instance=retriever, name="retriever")
    pipeline.add_component(instance=document_classifier, name="document_classifier")
    pipeline.connect("retriever", "document_classifier")

    queries = ["How was your day today?", "How was your day yesterday?"]
    expected_predictions = ["positive", "negative"]

    for idx, query in enumerate(queries):
        result = pipeline.run({"retriever": {"query": query, "top_k": 1}})
        assert result["document_classifier"]["documents"][0].to_dict()["id"] == str(idx)
        assert (result["document_classifier"]["documents"][0].to_dict()["classification"]["label"]
                == expected_predictions[idx])
    ```
    """

    def __init__(  # pylint: disable=too-many-positional-arguments
        self,
        model: str,
        labels: List[str],
        multi_label: bool = False,
        classification_field: Optional[str] = None,
        device: Optional[ComponentDevice] = None,
        token: Optional[Secret] = Secret.from_env_var(["HF_API_TOKEN", "HF_TOKEN"], strict=False),
        huggingface_pipeline_kwargs: Optional[Dict[str, Any]] = None,
    ):
        """
        Initializes the TransformersZeroShotDocumentClassifier.

        See the Hugging Face [website](https://huggingface.co/models?pipeline_tag=zero-shot-classification&sort=downloads&search=nli)
        for the full list of zero-shot classification models (NLI) models.

        :param model:
            The name or path of a Hugging Face model for zero shot document classification.
        :param labels:
            The set of possible class labels to classify each document into, for example,
            ["positive", "negative"]. The labels depend on the selected model.
        :param multi_label:
            Whether or not multiple candidate labels can be true.
            If `False`, the scores are normalized such that
            the sum of the label likelihoods for each sequence is 1. If `True`, the labels are considered
            independent and probabilities are normalized for each candidate by doing a softmax of the entailment
            score vs. the contradiction score.
        :param classification_field:
            Name of document's meta field to be used for classification.
            If not set, `Document.content` is used by default.
        :param device:
            The device on which the model is loaded. If `None`, the default device is automatically
            selected. If a device/device map is specified in `huggingface_pipeline_kwargs`, it overrides this parameter.
        :param token:
            The Hugging Face token to use as HTTP bearer authorization.
            Check your HF token in your [account settings](https://huggingface.co/settings/tokens).
        :param huggingface_pipeline_kwargs:
            Dictionary containing keyword arguments used to initialize the
            Hugging Face pipeline for text classification.
        """

        torch_and_transformers_import.check()

        self.classification_field = classification_field

        self.token = token
        self.labels = labels
        self.multi_label = multi_label

        huggingface_pipeline_kwargs = resolve_hf_pipeline_kwargs(
            huggingface_pipeline_kwargs=huggingface_pipeline_kwargs or {},
            model=model,
            task="zero-shot-classification",
            supported_tasks=["zero-shot-classification"],
            device=device,
            token=token,
        )

        self.huggingface_pipeline_kwargs = huggingface_pipeline_kwargs
        self.pipeline = None

    def _get_telemetry_data(self) -> Dict[str, Any]:
        """
        Data that is sent to Posthog for usage analytics.
        """
        if isinstance(self.huggingface_pipeline_kwargs["model"], str):
            return {"model": self.huggingface_pipeline_kwargs["model"]}
        return {"model": f"[object of type {type(self.huggingface_pipeline_kwargs['model'])}]"}

    def warm_up(self):
        """
        Initializes the component.
        """
        if self.pipeline is None:
            self.pipeline = pipeline(**self.huggingface_pipeline_kwargs)

    def to_dict(self) -> Dict[str, Any]:
        """
        Serializes the component to a dictionary.

        :returns:
            Dictionary with serialized data.
        """
        serialization_dict = default_to_dict(
            self,
            labels=self.labels,
            model=self.huggingface_pipeline_kwargs["model"],
            huggingface_pipeline_kwargs=self.huggingface_pipeline_kwargs,
            token=self.token.to_dict() if self.token else None,
        )

        huggingface_pipeline_kwargs = serialization_dict["init_parameters"]["huggingface_pipeline_kwargs"]
        huggingface_pipeline_kwargs.pop("token", None)

        serialize_hf_model_kwargs(huggingface_pipeline_kwargs)
        return serialization_dict

    @classmethod
    def from_dict(cls, data: Dict[str, Any]) -> "TransformersZeroShotDocumentClassifier":
        """
        Deserializes the component from a dictionary.

        :param data:
            Dictionary to deserialize from.
        :returns:
            Deserialized component.
        """
        deserialize_secrets_inplace(data["init_parameters"], keys=["token"])
        if data["init_parameters"].get("huggingface_pipeline_kwargs") is not None:
            deserialize_hf_model_kwargs(data["init_parameters"]["huggingface_pipeline_kwargs"])
        return default_from_dict(cls, data)

    @component.output_types(documents=List[Document])
    def run(self, documents: List[Document], batch_size: int = 1):
        """
        Classifies the documents based on the provided labels and adds them to their metadata.

        The classification results are stored in the `classification` dict within
        each document's metadata. If `multi_label` is set to `True`, the scores for each label are available under
        the `details` key within the `classification` dictionary.

        :param documents:
            Documents to process.
        :param batch_size:
            Batch size used for processing the content in each document.
        :returns:
            A dictionary with the following key:
            - `documents`: A list of documents with an added metadata field called `classification`.
        """

        if self.pipeline is None:
            raise RuntimeError(
                "The component TransformerZeroShotDocumentClassifier wasn't warmed up. "
                "Run 'warm_up()' before calling 'run()'."
            )

        if not isinstance(documents, list) or documents and not isinstance(documents[0], Document):
            raise TypeError(
                "TransformerZeroShotDocumentClassifier expects a list of documents as input. "
                "In case you want to classify and route a text, please use the TransformersZeroShotTextRouter."
            )

        invalid_doc_ids = []

        for doc in documents:
            if self.classification_field is not None and self.classification_field not in doc.meta:
                invalid_doc_ids.append(doc.id)

        if invalid_doc_ids:
            raise ValueError(
                f"The following documents do not have the classification field '{self.classification_field}': "
                f"{', '.join(invalid_doc_ids)}"
            )

        texts = [
            (doc.content if self.classification_field is None else doc.meta[self.classification_field])
            for doc in documents
        ]

        predictions = self.pipeline(texts, self.labels, multi_label=self.multi_label, batch_size=batch_size)

        for prediction, document in zip(predictions, documents):
            formatted_prediction = {
                "label": prediction["labels"][0],
                "score": prediction["scores"][0],
                "details": dict(zip(prediction["labels"], prediction["scores"])),
            }
            document.meta["classification"] = formatted_prediction

        return {"documents": documents}

1	# SPDX-FileCopyrightText: 2022-present deepset GmbH <info@deepset.ai>
2	#
3	# SPDX-License-Identifier: Apache-2.0
4
5	from typing import Any, Dict, List, Optional	1✔
6
7	from haystack import Document, component, default_from_dict, default_to_dict	1✔
8	from haystack.lazy_imports import LazyImport	1✔
9	from haystack.utils import ComponentDevice, Secret, deserialize_secrets_inplace	1✔
10	from haystack.utils.hf import deserialize_hf_model_kwargs, resolve_hf_pipeline_kwargs, serialize_hf_model_kwargs	1✔
11
12	with LazyImport(message="Run 'pip install transformers[torch,sentencepiece]'") as torch_and_transformers_import:	1✔
13	from transformers import pipeline	1✔
14
15
16	@component	1✔
17	class TransformersZeroShotDocumentClassifier:	1✔
18	"""
19	Performs zero-shot classification of documents based on given labels and adds the predicted label to their metadata.
20
21	The component uses a Hugging Face pipeline for zero-shot classification.
22	Provide the model and the set of labels to be used for categorization during initialization.
23	Additionally, you can configure the component to allow multiple labels to be true.
24
25	Classification is run on the document's content field by default. If you want it to run on another field, set the
26	`classification_field` to one of the document's metadata fields.
27
28	Available models for the task of zero-shot-classification include:
29	- `valhalla/distilbart-mnli-12-3`
30	- `cross-encoder/nli-distilroberta-base`
31	- `cross-encoder/nli-deberta-v3-xsmall`
32
33	### Usage example
34
35	The following is a pipeline that classifies documents based on predefined classification labels
36	retrieved from a search pipeline:
37
38	```python
39	from haystack import Document
40	from haystack.components.retrievers.in_memory import InMemoryBM25Retriever
41	from haystack.document_stores.in_memory import InMemoryDocumentStore
42	from haystack.core.pipeline import Pipeline
43	from haystack.components.classifiers import TransformersZeroShotDocumentClassifier
44
45	documents = [Document(id="0", content="Today was a nice day!"),
46	Document(id="1", content="Yesterday was a bad day!")]
47
48	document_store = InMemoryDocumentStore()
49	retriever = InMemoryBM25Retriever(document_store=document_store)
50	document_classifier = TransformersZeroShotDocumentClassifier(
51	model="cross-encoder/nli-deberta-v3-xsmall",
52	labels=["positive", "negative"],
53	)
54
55	document_store.write_documents(documents)
56
57	pipeline = Pipeline()
58	pipeline.add_component(instance=retriever, name="retriever")
59	pipeline.add_component(instance=document_classifier, name="document_classifier")
60	pipeline.connect("retriever", "document_classifier")
61
62	queries = ["How was your day today?", "How was your day yesterday?"]
63	expected_predictions = ["positive", "negative"]
64
65	for idx, query in enumerate(queries):
66	result = pipeline.run({"retriever": {"query": query, "top_k": 1}})
67	assert result["document_classifier"]["documents"][0].to_dict()["id"] == str(idx)
68	assert (result["document_classifier"]["documents"][0].to_dict()["classification"]["label"]
69	== expected_predictions[idx])
70	```
71	"""
72
73	def __init__( # pylint: disable=too-many-positional-arguments	1✔
74	self,
75	model: str,
76	labels: List[str],
77	multi_label: bool = False,
78	classification_field: Optional[str] = None,
79	device: Optional[ComponentDevice] = None,
80	token: Optional[Secret] = Secret.from_env_var(["HF_API_TOKEN", "HF_TOKEN"], strict=False),
81	huggingface_pipeline_kwargs: Optional[Dict[str, Any]] = None,
82	):
83	"""
84	Initializes the TransformersZeroShotDocumentClassifier.
85
86	See the Hugging Face [website](https://huggingface.co/models?pipeline_tag=zero-shot-classification&sort=downloads&search=nli)
87	for the full list of zero-shot classification models (NLI) models.
88
89	:param model:
90	The name or path of a Hugging Face model for zero shot document classification.
91	:param labels:
92	The set of possible class labels to classify each document into, for example,
93	["positive", "negative"]. The labels depend on the selected model.
94	:param multi_label:
95	Whether or not multiple candidate labels can be true.
96	If `False`, the scores are normalized such that
97	the sum of the label likelihoods for each sequence is 1. If `True`, the labels are considered
98	independent and probabilities are normalized for each candidate by doing a softmax of the entailment
99	score vs. the contradiction score.
100	:param classification_field:
101	Name of document's meta field to be used for classification.
102	If not set, `Document.content` is used by default.
103	:param device:
104	The device on which the model is loaded. If `None`, the default device is automatically
105	selected. If a device/device map is specified in `huggingface_pipeline_kwargs`, it overrides this parameter.
106	:param token:
107	The Hugging Face token to use as HTTP bearer authorization.
108	Check your HF token in your [account settings](https://huggingface.co/settings/tokens).
109	:param huggingface_pipeline_kwargs:
110	Dictionary containing keyword arguments used to initialize the
111	Hugging Face pipeline for text classification.
112	"""
113
114	torch_and_transformers_import.check()	1✔
115
116	self.classification_field = classification_field	1✔
117
118	self.token = token	1✔
119	self.labels = labels	1✔
120	self.multi_label = multi_label	1✔
121
122	huggingface_pipeline_kwargs = resolve_hf_pipeline_kwargs(	1✔
123	huggingface_pipeline_kwargs=huggingface_pipeline_kwargs or {},
124	model=model,
125	task="zero-shot-classification",
126	supported_tasks=["zero-shot-classification"],
127	device=device,
128	token=token,
129	)
130
131	self.huggingface_pipeline_kwargs = huggingface_pipeline_kwargs	1✔
132	self.pipeline = None	1✔
133
134	def _get_telemetry_data(self) -> Dict[str, Any]:	1✔
135	"""
136	Data that is sent to Posthog for usage analytics.
137	"""
138	if isinstance(self.huggingface_pipeline_kwargs["model"], str):	×
139	return {"model": self.huggingface_pipeline_kwargs["model"]}	×
140	return {"model": f"[object of type {type(self.huggingface_pipeline_kwargs['model'])}]"}	×
141
142	def warm_up(self):	1✔
143	"""
144	Initializes the component.
145	"""
146	if self.pipeline is None:	1✔
147	self.pipeline = pipeline(**self.huggingface_pipeline_kwargs)	1✔
148
149	def to_dict(self) -> Dict[str, Any]:	1✔
150	"""
151	Serializes the component to a dictionary.
152
153	:returns:
154	Dictionary with serialized data.
155	"""
156	serialization_dict = default_to_dict(	1✔
157	self,
158	labels=self.labels,
159	model=self.huggingface_pipeline_kwargs["model"],
160	huggingface_pipeline_kwargs=self.huggingface_pipeline_kwargs,
161	token=self.token.to_dict() if self.token else None,
162	)
163
164	huggingface_pipeline_kwargs = serialization_dict["init_parameters"]["huggingface_pipeline_kwargs"]	1✔
165	huggingface_pipeline_kwargs.pop("token", None)	1✔
166
167	serialize_hf_model_kwargs(huggingface_pipeline_kwargs)	1✔
168	return serialization_dict	1✔
169
170	@classmethod	1✔
171	def from_dict(cls, data: Dict[str, Any]) -> "TransformersZeroShotDocumentClassifier":	1✔
172	"""
173	Deserializes the component from a dictionary.
174
175	:param data:
176	Dictionary to deserialize from.
177	:returns:
178	Deserialized component.
179	"""
180	deserialize_secrets_inplace(data["init_parameters"], keys=["token"])	1✔
181	if data["init_parameters"].get("huggingface_pipeline_kwargs") is not None:	1✔
182	deserialize_hf_model_kwargs(data["init_parameters"]["huggingface_pipeline_kwargs"])	1✔
183	return default_from_dict(cls, data)	1✔
184
185	@component.output_types(documents=List[Document])	1✔
186	def run(self, documents: List[Document], batch_size: int = 1):	1✔
187	"""
188	Classifies the documents based on the provided labels and adds them to their metadata.
189
190	The classification results are stored in the `classification` dict within
191	each document's metadata. If `multi_label` is set to `True`, the scores for each label are available under
192	the `details` key within the `classification` dictionary.
193
194	:param documents:
195	Documents to process.
196	:param batch_size:
197	Batch size used for processing the content in each document.
198	:returns:
199	A dictionary with the following key:
200	- `documents`: A list of documents with an added metadata field called `classification`.
201	"""
202
203	if self.pipeline is None:	1✔
204	raise RuntimeError(	1✔
205	"The component TransformerZeroShotDocumentClassifier wasn't warmed up. "
206	"Run 'warm_up()' before calling 'run()'."
207	)
208
209	if not isinstance(documents, list) or documents and not isinstance(documents[0], Document):	1✔
210	raise TypeError(	1✔
211	"TransformerZeroShotDocumentClassifier expects a list of documents as input. "
212	"In case you want to classify and route a text, please use the TransformersZeroShotTextRouter."
213	)
214
215	invalid_doc_ids = []	1✔
216
217	for doc in documents:	1✔
218	if self.classification_field is not None and self.classification_field not in doc.meta:	1✔
219	invalid_doc_ids.append(doc.id)	×
220
221	if invalid_doc_ids:	1✔
222	raise ValueError(	×
223	f"The following documents do not have the classification field '{self.classification_field}': "
224	f"{', '.join(invalid_doc_ids)}"
225	)
226
227	texts = [	1✔
228	(doc.content if self.classification_field is None else doc.meta[self.classification_field])
229	for doc in documents
230	]
231
232	predictions = self.pipeline(texts, self.labels, multi_label=self.multi_label, batch_size=batch_size)	1✔
233
234	for prediction, document in zip(predictions, documents):	1✔
235	formatted_prediction = {	1✔
236	"label": prediction["labels"][0],
237	"score": prediction["scores"][0],
238	"details": dict(zip(prediction["labels"], prediction["scores"])),
239	}
240	document.meta["classification"] = formatted_prediction	1✔
241
242	return {"documents": documents}	1✔

deepset-ai / haystack / 13972131258

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous