12709332184

Committed 10 Jan 2025 12:11PM UTC coverage: 91.1% (+0.001%) from 91.099%

Build # 12709332184

Build Type

Pull #8702

github

Committed by

web-flow

Commit Message

Merge 3e4d6bfca into 08cf09f83

Pull Request Pull Request #8702: fix: `OpenAIChatGenerator` - do not pass tools to the OpenAI client when none are provided

Run Details

8660 of 9506 relevant lines covered (91.1%)

0.91 hits per line

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

66.67

haystack/components/embedders/azure_text_embedder.py

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

import os
from typing import Any, Dict, List, Optional

from openai.lib.azure import AzureOpenAI

from haystack import Document, component, default_from_dict, default_to_dict
from haystack.utils import Secret, deserialize_secrets_inplace


@component
class AzureOpenAITextEmbedder:
    """
    Embeds strings using OpenAI models deployed on Azure.

    ### Usage example

    ```python
    from haystack.components.embedders import AzureOpenAITextEmbedder

    text_to_embed = "I love pizza!"

    text_embedder = AzureOpenAITextEmbedder()

    print(text_embedder.run(text_to_embed))

    # {'embedding': [0.017020374536514282, -0.023255806416273117, ...],
    # 'meta': {'model': 'text-embedding-ada-002-v2',
    #          'usage': {'prompt_tokens': 4, 'total_tokens': 4}}}
    ```
    """

    def __init__(  # pylint: disable=too-many-positional-arguments
        self,
        azure_endpoint: Optional[str] = None,
        api_version: Optional[str] = "2023-05-15",
        azure_deployment: str = "text-embedding-ada-002",
        dimensions: Optional[int] = None,
        api_key: Optional[Secret] = Secret.from_env_var("AZURE_OPENAI_API_KEY", strict=False),
        azure_ad_token: Optional[Secret] = Secret.from_env_var("AZURE_OPENAI_AD_TOKEN", strict=False),
        organization: Optional[str] = None,
        timeout: Optional[float] = None,
        max_retries: Optional[int] = None,
        prefix: str = "",
        suffix: str = "",
    ):
        """
        Creates an AzureOpenAITextEmbedder component.

        :param azure_endpoint:
            The endpoint of the model deployed on Azure.
        :param api_version:
            The version of the API to use.
        :param azure_deployment:
            The name of the model deployed on Azure. The default model is text-embedding-ada-002.
        :param dimensions:
            The number of dimensions the resulting output embeddings should have. Only supported in text-embedding-3
            and later models.
        :param api_key:
            The Azure OpenAI API key.
            You can set it with an environment variable `AZURE_OPENAI_API_KEY`, or pass with this
            parameter during initialization.
        :param azure_ad_token:
            Microsoft Entra ID token, see Microsoft's
            [Entra ID](https://www.microsoft.com/en-us/security/business/identity-access/microsoft-entra-id)
            documentation for more information. You can set it with an environment variable
            `AZURE_OPENAI_AD_TOKEN`, or pass with this parameter during initialization.
            Previously called Azure Active Directory.
        :param organization:
            Your organization ID. See OpenAI's
            [Setting Up Your Organization](https://platform.openai.com/docs/guides/production-best-practices/setting-up-your-organization)
            for more information.
        :param timeout: The timeout for `AzureOpenAI` client calls, in seconds.
            If not set, defaults to either the
            `OPENAI_TIMEOUT` environment variable, or 30 seconds.
        :param max_retries: Maximum number of retries to contact AzureOpenAI after an internal error.
            If not set, defaults to either the `OPENAI_MAX_RETRIES` environment variable, or to 5 retries.
        :param prefix:
            A string to add at the beginning of each text.
        :param suffix:
            A string to add at the end of each text.
        """
        # Why is this here?
        # AzureOpenAI init is forcing us to use an init method that takes either base_url or azure_endpoint as not
        # None init parameters. This way we accommodate the use case where env var AZURE_OPENAI_ENDPOINT is set instead
        # of passing it as a parameter.
        azure_endpoint = azure_endpoint or os.environ.get("AZURE_OPENAI_ENDPOINT")
        if not azure_endpoint:
            raise ValueError("Please provide an Azure endpoint or set the environment variable AZURE_OPENAI_ENDPOINT.")

        if api_key is None and azure_ad_token is None:
            raise ValueError("Please provide an API key or an Azure Active Directory token.")

        self.api_key = api_key
        self.azure_ad_token = azure_ad_token
        self.api_version = api_version
        self.azure_endpoint = azure_endpoint
        self.azure_deployment = azure_deployment
        self.dimensions = dimensions
        self.organization = organization
        self.timeout = timeout or float(os.environ.get("OPENAI_TIMEOUT", 30.0))
        self.max_retries = max_retries or int(os.environ.get("OPENAI_MAX_RETRIES", 5))
        self.prefix = prefix
        self.suffix = suffix

        self._client = AzureOpenAI(
            api_version=api_version,
            azure_endpoint=azure_endpoint,
            azure_deployment=azure_deployment,
            api_key=api_key.resolve_value() if api_key is not None else None,
            azure_ad_token=azure_ad_token.resolve_value() if azure_ad_token is not None else None,
            organization=organization,
            timeout=self.timeout,
            max_retries=self.max_retries,
        )

    def _get_telemetry_data(self) -> Dict[str, Any]:
        """
        Data that is sent to Posthog for usage analytics.
        """
        return {"model": self.azure_deployment}

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

        :returns:
            Dictionary with serialized data.
        """
        return default_to_dict(
            self,
            azure_endpoint=self.azure_endpoint,
            azure_deployment=self.azure_deployment,
            dimensions=self.dimensions,
            organization=self.organization,
            api_version=self.api_version,
            prefix=self.prefix,
            suffix=self.suffix,
            api_key=self.api_key.to_dict() if self.api_key is not None else None,
            azure_ad_token=self.azure_ad_token.to_dict() if self.azure_ad_token is not None else None,
            timeout=self.timeout,
            max_retries=self.max_retries,
        )

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

        :param data:
            Dictionary to deserialize from.
        :returns:
            Deserialized component.
        """
        deserialize_secrets_inplace(data["init_parameters"], keys=["api_key", "azure_ad_token"])
        return default_from_dict(cls, data)

    @component.output_types(embedding=List[float], meta=Dict[str, Any])
    def run(self, text: str):
        """
        Embeds a single string.

        :param text:
            Text to embed.

        :returns:
            A dictionary with the following keys:
            - `embedding`: The embedding of the input text.
            - `meta`: Information about the usage of the model.
        """
        if not isinstance(text, str):
            # Check if input is a list and all elements are instances of Document
            if isinstance(text, list) and all(isinstance(elem, Document) for elem in text):
                error_message = "Input must be a string. Use AzureOpenAIDocumentEmbedder for a list of Documents."
            else:
                error_message = "Input must be a string."
            raise TypeError(error_message)

        # Preprocess the text by adding prefixes/suffixes
        # finally, replace newlines as recommended by OpenAI docs
        processed_text = f"{self.prefix}{text}{self.suffix}".replace("\n", " ")

        if self.dimensions is not None:
            response = self._client.embeddings.create(
                model=self.azure_deployment, dimensions=self.dimensions, input=processed_text
            )
        else:
            response = self._client.embeddings.create(model=self.azure_deployment, input=processed_text)

        return {
            "embedding": response.data[0].embedding,
            "meta": {"model": response.model, "usage": dict(response.usage)},
        }

1	# SPDX-FileCopyrightText: 2022-present deepset GmbH <info@deepset.ai>
2	#
3	# SPDX-License-Identifier: Apache-2.0
4
5	import os	1✔
6	from typing import Any, Dict, List, Optional	1✔
7
8	from openai.lib.azure import AzureOpenAI	1✔
9
10	from haystack import Document, component, default_from_dict, default_to_dict	1✔
11	from haystack.utils import Secret, deserialize_secrets_inplace	1✔
12
13
14	@component	1✔
15	class AzureOpenAITextEmbedder:	1✔
16	"""
17	Embeds strings using OpenAI models deployed on Azure.
18
19	### Usage example
20
21	```python
22	from haystack.components.embedders import AzureOpenAITextEmbedder
23
24	text_to_embed = "I love pizza!"
25
26	text_embedder = AzureOpenAITextEmbedder()
27
28	print(text_embedder.run(text_to_embed))
29
30	# {'embedding': [0.017020374536514282, -0.023255806416273117, ...],
31	# 'meta': {'model': 'text-embedding-ada-002-v2',
32	# 'usage': {'prompt_tokens': 4, 'total_tokens': 4}}}
33	```
34	"""
35
36	def __init__( # pylint: disable=too-many-positional-arguments	1✔
37	self,
38	azure_endpoint: Optional[str] = None,
39	api_version: Optional[str] = "2023-05-15",
40	azure_deployment: str = "text-embedding-ada-002",
41	dimensions: Optional[int] = None,
42	api_key: Optional[Secret] = Secret.from_env_var("AZURE_OPENAI_API_KEY", strict=False),
43	azure_ad_token: Optional[Secret] = Secret.from_env_var("AZURE_OPENAI_AD_TOKEN", strict=False),
44	organization: Optional[str] = None,
45	timeout: Optional[float] = None,
46	max_retries: Optional[int] = None,
47	prefix: str = "",
48	suffix: str = "",
49	):
50	"""
51	Creates an AzureOpenAITextEmbedder component.
52
53	:param azure_endpoint:
54	The endpoint of the model deployed on Azure.
55	:param api_version:
56	The version of the API to use.
57	:param azure_deployment:
58	The name of the model deployed on Azure. The default model is text-embedding-ada-002.
59	:param dimensions:
60	The number of dimensions the resulting output embeddings should have. Only supported in text-embedding-3
61	and later models.
62	:param api_key:
63	The Azure OpenAI API key.
64	You can set it with an environment variable `AZURE_OPENAI_API_KEY`, or pass with this
65	parameter during initialization.
66	:param azure_ad_token:
67	Microsoft Entra ID token, see Microsoft's
68	[Entra ID](https://www.microsoft.com/en-us/security/business/identity-access/microsoft-entra-id)
69	documentation for more information. You can set it with an environment variable
70	`AZURE_OPENAI_AD_TOKEN`, or pass with this parameter during initialization.
71	Previously called Azure Active Directory.
72	:param organization:
73	Your organization ID. See OpenAI's
74	[Setting Up Your Organization](https://platform.openai.com/docs/guides/production-best-practices/setting-up-your-organization)
75	for more information.
76	:param timeout: The timeout for `AzureOpenAI` client calls, in seconds.
77	If not set, defaults to either the
78	`OPENAI_TIMEOUT` environment variable, or 30 seconds.
79	:param max_retries: Maximum number of retries to contact AzureOpenAI after an internal error.
80	If not set, defaults to either the `OPENAI_MAX_RETRIES` environment variable, or to 5 retries.
81	:param prefix:
82	A string to add at the beginning of each text.
83	:param suffix:
84	A string to add at the end of each text.
85	"""
86	# Why is this here?
87	# AzureOpenAI init is forcing us to use an init method that takes either base_url or azure_endpoint as not
88	# None init parameters. This way we accommodate the use case where env var AZURE_OPENAI_ENDPOINT is set instead
89	# of passing it as a parameter.
90	azure_endpoint = azure_endpoint or os.environ.get("AZURE_OPENAI_ENDPOINT")	1✔
91	if not azure_endpoint:	1✔
92	raise ValueError("Please provide an Azure endpoint or set the environment variable AZURE_OPENAI_ENDPOINT.")	×
93
94	if api_key is None and azure_ad_token is None:	1✔
95	raise ValueError("Please provide an API key or an Azure Active Directory token.")	×
96
97	self.api_key = api_key	1✔
98	self.azure_ad_token = azure_ad_token	1✔
99	self.api_version = api_version	1✔
100	self.azure_endpoint = azure_endpoint	1✔
101	self.azure_deployment = azure_deployment	1✔
102	self.dimensions = dimensions	1✔
103	self.organization = organization	1✔
104	self.timeout = timeout or float(os.environ.get("OPENAI_TIMEOUT", 30.0))	1✔
105	self.max_retries = max_retries or int(os.environ.get("OPENAI_MAX_RETRIES", 5))	1✔
106	self.prefix = prefix	1✔
107	self.suffix = suffix	1✔
108
109	self._client = AzureOpenAI(	1✔
110	api_version=api_version,
111	azure_endpoint=azure_endpoint,
112	azure_deployment=azure_deployment,
113	api_key=api_key.resolve_value() if api_key is not None else None,
114	azure_ad_token=azure_ad_token.resolve_value() if azure_ad_token is not None else None,
115	organization=organization,
116	timeout=self.timeout,
117	max_retries=self.max_retries,
118	)
119
120	def _get_telemetry_data(self) -> Dict[str, Any]:	1✔
121	"""
122	Data that is sent to Posthog for usage analytics.
123	"""
124	return {"model": self.azure_deployment}	×
125
126	def to_dict(self) -> Dict[str, Any]:	1✔
127	"""
128	Serializes the component to a dictionary.
129
130	:returns:
131	Dictionary with serialized data.
132	"""
133	return default_to_dict(	1✔
134	self,
135	azure_endpoint=self.azure_endpoint,
136	azure_deployment=self.azure_deployment,
137	dimensions=self.dimensions,
138	organization=self.organization,
139	api_version=self.api_version,
140	prefix=self.prefix,
141	suffix=self.suffix,
142	api_key=self.api_key.to_dict() if self.api_key is not None else None,
143	azure_ad_token=self.azure_ad_token.to_dict() if self.azure_ad_token is not None else None,
144	timeout=self.timeout,
145	max_retries=self.max_retries,
146	)
147
148	@classmethod	1✔
149	def from_dict(cls, data: Dict[str, Any]) -> "AzureOpenAITextEmbedder":	1✔
150	"""
151	Deserializes the component from a dictionary.
152
153	:param data:
154	Dictionary to deserialize from.
155	:returns:
156	Deserialized component.
157	"""
158	deserialize_secrets_inplace(data["init_parameters"], keys=["api_key", "azure_ad_token"])	×
159	return default_from_dict(cls, data)	×
160
161	@component.output_types(embedding=List[float], meta=Dict[str, Any])	1✔
162	def run(self, text: str):	1✔
163	"""
164	Embeds a single string.
165
166	:param text:
167	Text to embed.
168
169	:returns:
170	A dictionary with the following keys:
171	- `embedding`: The embedding of the input text.
172	- `meta`: Information about the usage of the model.
173	"""
174	if not isinstance(text, str):	×
175	# Check if input is a list and all elements are instances of Document
176	if isinstance(text, list) and all(isinstance(elem, Document) for elem in text):	×
177	error_message = "Input must be a string. Use AzureOpenAIDocumentEmbedder for a list of Documents."	×
178	else:
179	error_message = "Input must be a string."	×
180	raise TypeError(error_message)	×
181
182	# Preprocess the text by adding prefixes/suffixes
183	# finally, replace newlines as recommended by OpenAI docs
184	processed_text = f"{self.prefix}{text}{self.suffix}".replace("\n", " ")	×
185
186	if self.dimensions is not None:	×
187	response = self._client.embeddings.create(	×
188	model=self.azure_deployment, dimensions=self.dimensions, input=processed_text
189	)
190	else:
191	response = self._client.embeddings.create(model=self.azure_deployment, input=processed_text)	×
192
193	return {	×
194	"embedding": response.data[0].embedding,
195	"meta": {"model": response.model, "usage": dict(response.usage)},
196	}

deepset-ai / haystack / 12709332184

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