18592817487

Committed 17 Oct 2025 12:33PM UTC coverage: 92.2% (+0.1%) from 92.062%

Build # 18592817487

Build Type

Pull #9859

github

Committed by

web-flow

Commit Message

Merge f20ff2b98 into a43c47b63

Pull Request Pull Request #9859: feat: Add FallbackChatGenerator

Run Details

13346 of 14475 relevant lines covered (92.2%)

0.92 hits per line

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

91.84

haystack/components/generators/azure.py

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

import os
from typing import Any, Optional

from openai.lib.azure import AzureADTokenProvider, AzureOpenAI

from haystack import component, default_from_dict, default_to_dict
from haystack.components.generators import OpenAIGenerator
from haystack.dataclasses import StreamingCallbackT
from haystack.utils import Secret, deserialize_callable, deserialize_secrets_inplace, serialize_callable
from haystack.utils.http_client import init_http_client


@component
class AzureOpenAIGenerator(OpenAIGenerator):
    """
    Generates text using OpenAI's large language models (LLMs).

    It works with the gpt-4 - type models and supports streaming responses
    from OpenAI API.

    You can customize how the text is generated by passing parameters to the
    OpenAI API. Use the `**generation_kwargs` argument when you initialize
    the component or when you run it. Any parameter that works with
    `openai.ChatCompletion.create` will work here too.


    For details on OpenAI API parameters, see
    [OpenAI documentation](https://platform.openai.com/docs/api-reference/chat).


    ### Usage example

    ```python
    from haystack.components.generators import AzureOpenAIGenerator
    from haystack.utils import Secret
    client = AzureOpenAIGenerator(
        azure_endpoint="<Your Azure endpoint e.g. `https://your-company.azure.openai.com/>",
        api_key=Secret.from_token("<your-api-key>"),
        azure_deployment="<this a model name, e.g.  gpt-4o-mini>")
    response = client.run("What's Natural Language Processing? Be brief.")
    print(response)
    ```

    ```
    >> {'replies': ['Natural Language Processing (NLP) is a branch of artificial intelligence that focuses on
    >> the interaction between computers and human language. It involves enabling computers to understand, interpret,
    >> and respond to natural human language in a way that is both meaningful and useful.'], 'meta': [{'model':
    >> 'gpt-4o-mini', 'index': 0, 'finish_reason': 'stop', 'usage': {'prompt_tokens': 16,
    >> 'completion_tokens': 49, 'total_tokens': 65}}]}
    ```
    """

    # pylint: disable=super-init-not-called
    def __init__(  # pylint: disable=too-many-positional-arguments  # noqa: PLR0913
        self,
        azure_endpoint: Optional[str] = None,
        api_version: Optional[str] = "2023-05-15",
        azure_deployment: Optional[str] = "gpt-4o-mini",
        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,
        streaming_callback: Optional[StreamingCallbackT] = None,
        system_prompt: Optional[str] = None,
        timeout: Optional[float] = None,
        max_retries: Optional[int] = None,
        http_client_kwargs: Optional[dict[str, Any]] = None,
        generation_kwargs: Optional[dict[str, Any]] = None,
        default_headers: Optional[dict[str, str]] = None,
        *,
        azure_ad_token_provider: Optional[AzureADTokenProvider] = None,
    ):
        """
        Initialize the Azure OpenAI Generator.

        :param azure_endpoint: The endpoint of the deployed model, for example `https://example-resource.azure.openai.com/`.
        :param api_version: The version of the API to use. Defaults to 2023-05-15.
        :param azure_deployment: The deployment of the model, usually the model name.
        :param api_key: The API key to use for authentication.
        :param azure_ad_token: [Azure Active Directory token](https://www.microsoft.com/en-us/security/business/identity-access/microsoft-entra-id).
        :param organization: Your organization ID, defaults to `None`. For help, see
        [Setting up your organization](https://platform.openai.com/docs/guides/production-best-practices/setting-up-your-organization).
        :param streaming_callback: A callback function called when a new token is received from the stream.
            It accepts [StreamingChunk](https://docs.haystack.deepset.ai/docs/data-classes#streamingchunk)
            as an argument.
        :param system_prompt: The system prompt to use for text generation. If not provided, the Generator
        omits the system prompt and uses the default system prompt.
        :param timeout: Timeout for AzureOpenAI client. If not set, it is inferred from the
            `OPENAI_TIMEOUT` environment variable or set to 30.
        :param max_retries: Maximum retries to establish contact with AzureOpenAI if it returns an internal error.
            If not set, it is inferred from the `OPENAI_MAX_RETRIES` environment variable or set to 5.
        :param http_client_kwargs:
            A dictionary of keyword arguments to configure a custom `httpx.Client`or `httpx.AsyncClient`.
            For more information, see the [HTTPX documentation](https://www.python-httpx.org/api/#client).
        :param generation_kwargs: Other parameters to use for the model, sent directly to
            the OpenAI endpoint. See [OpenAI documentation](https://platform.openai.com/docs/api-reference/chat) for
            more details.
            Some of the supported parameters:
            - `max_completion_tokens`: An upper bound for the number of tokens that can be generated for a completion,
                including visible output tokens and reasoning tokens.
            - `temperature`: The sampling temperature to use. Higher values mean the model takes more risks.
                Try 0.9 for more creative applications and 0 (argmax sampling) for ones with a well-defined answer.
            - `top_p`: An alternative to sampling with temperature, called nucleus sampling, where the model
                considers the results of the tokens with top_p probability mass. For example, 0.1 means only the tokens
                comprising the top 10% probability mass are considered.
            - `n`: The number of completions to generate for each prompt. For example, with 3 prompts and n=2,
                the LLM will generate two completions per prompt, resulting in 6 completions total.
            - `stop`: One or more sequences after which the LLM should stop generating tokens.
            - `presence_penalty`: The penalty applied if a token is already present.
                Higher values make the model less likely to repeat the token.
            - `frequency_penalty`: Penalty applied if a token has already been generated.
                Higher values make the model less likely to repeat the token.
            - `logit_bias`: Adds a logit bias to specific tokens. The keys of the dictionary are tokens, and the
                values are the bias to add to that token.
        :param default_headers: Default headers to use for the AzureOpenAI client.
        :param azure_ad_token_provider: A function that returns an Azure Active Directory token, will be invoked on
            every request.
        """
        # We intentionally do not call super().__init__ here because we only need to instantiate the client to interact
        # with the API.

        # 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.")

        # The check above makes mypy incorrectly infer that api_key is never None,
        # which propagates the incorrect type.
        self.api_key = api_key  # type: ignore
        self.azure_ad_token = azure_ad_token
        self.generation_kwargs = generation_kwargs or {}
        self.system_prompt = system_prompt
        self.streaming_callback = streaming_callback
        self.api_version = api_version
        self.azure_endpoint = azure_endpoint
        self.azure_deployment = azure_deployment
        self.organization = organization
        self.model: str = azure_deployment or "gpt-4o-mini"
        self.timeout = timeout if timeout is not None else float(os.environ.get("OPENAI_TIMEOUT", "30.0"))
        self.max_retries = max_retries if max_retries is not None else int(os.environ.get("OPENAI_MAX_RETRIES", "5"))
        self.http_client_kwargs = http_client_kwargs
        self.default_headers = default_headers or {}
        self.azure_ad_token_provider = azure_ad_token_provider

        self.client = AzureOpenAI(
            api_version=api_version,
            azure_endpoint=azure_endpoint,
            azure_deployment=azure_deployment,
            azure_ad_token_provider=azure_ad_token_provider,
            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,
            http_client=init_http_client(self.http_client_kwargs, async_client=False),
            default_headers=self.default_headers,
        )

    def to_dict(self) -> dict[str, Any]:
        """
        Serialize this component to a dictionary.

        :returns:
            The serialized component as a dictionary.
        """
        callback_name = serialize_callable(self.streaming_callback) if self.streaming_callback else None
        azure_ad_token_provider_name = None
        if self.azure_ad_token_provider:
            azure_ad_token_provider_name = serialize_callable(self.azure_ad_token_provider)
        return default_to_dict(
            self,
            azure_endpoint=self.azure_endpoint,
            azure_deployment=self.azure_deployment,
            organization=self.organization,
            api_version=self.api_version,
            streaming_callback=callback_name,
            generation_kwargs=self.generation_kwargs,
            system_prompt=self.system_prompt,
            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,
            http_client_kwargs=self.http_client_kwargs,
            default_headers=self.default_headers,
            azure_ad_token_provider=azure_ad_token_provider_name,
        )

    @classmethod
    def from_dict(cls, data: dict[str, Any]) -> "AzureOpenAIGenerator":
        """
        Deserialize this component from a dictionary.

        :param data:
            The dictionary representation of this component.
        :returns:
            The deserialized component instance.
        """
        deserialize_secrets_inplace(data["init_parameters"], keys=["api_key", "azure_ad_token"])
        init_params = data.get("init_parameters", {})
        serialized_callback_handler = init_params.get("streaming_callback")
        if serialized_callback_handler:
            data["init_parameters"]["streaming_callback"] = deserialize_callable(serialized_callback_handler)
        serialized_azure_ad_token_provider = init_params.get("azure_ad_token_provider")
        if serialized_azure_ad_token_provider:
            data["init_parameters"]["azure_ad_token_provider"] = deserialize_callable(
                serialized_azure_ad_token_provider
            )
        return default_from_dict(cls, data)

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, Optional	1✔
7
8	from openai.lib.azure import AzureADTokenProvider, AzureOpenAI	1✔
9
10	from haystack import component, default_from_dict, default_to_dict	1✔
11	from haystack.components.generators import OpenAIGenerator	1✔
12	from haystack.dataclasses import StreamingCallbackT	1✔
13	from haystack.utils import Secret, deserialize_callable, deserialize_secrets_inplace, serialize_callable	1✔
14	from haystack.utils.http_client import init_http_client	1✔
15
16
17	@component	1✔
18	class AzureOpenAIGenerator(OpenAIGenerator):	1✔
19	"""
20	Generates text using OpenAI's large language models (LLMs).
21
22	It works with the gpt-4 - type models and supports streaming responses
23	from OpenAI API.
24
25	You can customize how the text is generated by passing parameters to the
26	OpenAI API. Use the `**generation_kwargs` argument when you initialize
27	the component or when you run it. Any parameter that works with
28	`openai.ChatCompletion.create` will work here too.
29
30
31	For details on OpenAI API parameters, see
32	[OpenAI documentation](https://platform.openai.com/docs/api-reference/chat).
33
34
35	### Usage example
36
37	```python
38	from haystack.components.generators import AzureOpenAIGenerator
39	from haystack.utils import Secret
40	client = AzureOpenAIGenerator(
41	azure_endpoint="<Your Azure endpoint e.g. `https://your-company.azure.openai.com/>",
42	api_key=Secret.from_token("<your-api-key>"),
43	azure_deployment="<this a model name, e.g. gpt-4o-mini>")
44	response = client.run("What's Natural Language Processing? Be brief.")
45	print(response)
46	```
47
48	```
49	>> {'replies': ['Natural Language Processing (NLP) is a branch of artificial intelligence that focuses on
50	>> the interaction between computers and human language. It involves enabling computers to understand, interpret,
51	>> and respond to natural human language in a way that is both meaningful and useful.'], 'meta': [{'model':
52	>> 'gpt-4o-mini', 'index': 0, 'finish_reason': 'stop', 'usage': {'prompt_tokens': 16,
53	>> 'completion_tokens': 49, 'total_tokens': 65}}]}
54	```
55	"""
56
57	# pylint: disable=super-init-not-called
58	def __init__( # pylint: disable=too-many-positional-arguments # noqa: PLR0913	1✔
59	self,
60	azure_endpoint: Optional[str] = None,
61	api_version: Optional[str] = "2023-05-15",
62	azure_deployment: Optional[str] = "gpt-4o-mini",
63	api_key: Optional[Secret] = Secret.from_env_var("AZURE_OPENAI_API_KEY", strict=False),
64	azure_ad_token: Optional[Secret] = Secret.from_env_var("AZURE_OPENAI_AD_TOKEN", strict=False),
65	organization: Optional[str] = None,
66	streaming_callback: Optional[StreamingCallbackT] = None,
67	system_prompt: Optional[str] = None,
68	timeout: Optional[float] = None,
69	max_retries: Optional[int] = None,
70	http_client_kwargs: Optional[dict[str, Any]] = None,
71	generation_kwargs: Optional[dict[str, Any]] = None,
72	default_headers: Optional[dict[str, str]] = None,
73	*,
74	azure_ad_token_provider: Optional[AzureADTokenProvider] = None,
75	):
76	"""
77	Initialize the Azure OpenAI Generator.
78
79	:param azure_endpoint: The endpoint of the deployed model, for example `https://example-resource.azure.openai.com/`.
80	:param api_version: The version of the API to use. Defaults to 2023-05-15.
81	:param azure_deployment: The deployment of the model, usually the model name.
82	:param api_key: The API key to use for authentication.
83	:param azure_ad_token: [Azure Active Directory token](https://www.microsoft.com/en-us/security/business/identity-access/microsoft-entra-id).
84	:param organization: Your organization ID, defaults to `None`. For help, see
85	[Setting up your organization](https://platform.openai.com/docs/guides/production-best-practices/setting-up-your-organization).
86	:param streaming_callback: A callback function called when a new token is received from the stream.
87	It accepts [StreamingChunk](https://docs.haystack.deepset.ai/docs/data-classes#streamingchunk)
88	as an argument.
89	:param system_prompt: The system prompt to use for text generation. If not provided, the Generator
90	omits the system prompt and uses the default system prompt.
91	:param timeout: Timeout for AzureOpenAI client. If not set, it is inferred from the
92	`OPENAI_TIMEOUT` environment variable or set to 30.
93	:param max_retries: Maximum retries to establish contact with AzureOpenAI if it returns an internal error.
94	If not set, it is inferred from the `OPENAI_MAX_RETRIES` environment variable or set to 5.
95	:param http_client_kwargs:
96	A dictionary of keyword arguments to configure a custom `httpx.Client`or `httpx.AsyncClient`.
97	For more information, see the [HTTPX documentation](https://www.python-httpx.org/api/#client).
98	:param generation_kwargs: Other parameters to use for the model, sent directly to
99	the OpenAI endpoint. See [OpenAI documentation](https://platform.openai.com/docs/api-reference/chat) for
100	more details.
101	Some of the supported parameters:
102	- `max_completion_tokens`: An upper bound for the number of tokens that can be generated for a completion,
103	including visible output tokens and reasoning tokens.
104	- `temperature`: The sampling temperature to use. Higher values mean the model takes more risks.
105	Try 0.9 for more creative applications and 0 (argmax sampling) for ones with a well-defined answer.
106	- `top_p`: An alternative to sampling with temperature, called nucleus sampling, where the model
107	considers the results of the tokens with top_p probability mass. For example, 0.1 means only the tokens
108	comprising the top 10% probability mass are considered.
109	- `n`: The number of completions to generate for each prompt. For example, with 3 prompts and n=2,
110	the LLM will generate two completions per prompt, resulting in 6 completions total.
111	- `stop`: One or more sequences after which the LLM should stop generating tokens.
112	- `presence_penalty`: The penalty applied if a token is already present.
113	Higher values make the model less likely to repeat the token.
114	- `frequency_penalty`: Penalty applied if a token has already been generated.
115	Higher values make the model less likely to repeat the token.
116	- `logit_bias`: Adds a logit bias to specific tokens. The keys of the dictionary are tokens, and the
117	values are the bias to add to that token.
118	:param default_headers: Default headers to use for the AzureOpenAI client.
119	:param azure_ad_token_provider: A function that returns an Azure Active Directory token, will be invoked on
120	every request.
121	"""
122	# We intentionally do not call super().__init__ here because we only need to instantiate the client to interact
123	# with the API.
124
125	# Why is this here?
126	# AzureOpenAI init is forcing us to use an init method that takes either base_url or azure_endpoint as not
127	# None init parameters. This way we accommodate the use case where env var AZURE_OPENAI_ENDPOINT is set instead
128	# of passing it as a parameter.
129	azure_endpoint = azure_endpoint or os.environ.get("AZURE_OPENAI_ENDPOINT")	1✔
130	if not azure_endpoint:	1✔
131	raise ValueError("Please provide an Azure endpoint or set the environment variable AZURE_OPENAI_ENDPOINT.")	×
132
133	if api_key is None and azure_ad_token is None:	1✔
134	raise ValueError("Please provide an API key or an Azure Active Directory token.")	×
135
136	# The check above makes mypy incorrectly infer that api_key is never None,
137	# which propagates the incorrect type.
138	self.api_key = api_key # type: ignore	1✔
139	self.azure_ad_token = azure_ad_token	1✔
140	self.generation_kwargs = generation_kwargs or {}	1✔
141	self.system_prompt = system_prompt	1✔
142	self.streaming_callback = streaming_callback	1✔
143	self.api_version = api_version	1✔
144	self.azure_endpoint = azure_endpoint	1✔
145	self.azure_deployment = azure_deployment	1✔
146	self.organization = organization	1✔
147	self.model: str = azure_deployment or "gpt-4o-mini"	1✔
148	self.timeout = timeout if timeout is not None else float(os.environ.get("OPENAI_TIMEOUT", "30.0"))	1✔
149	self.max_retries = max_retries if max_retries is not None else int(os.environ.get("OPENAI_MAX_RETRIES", "5"))	1✔
150	self.http_client_kwargs = http_client_kwargs	1✔
151	self.default_headers = default_headers or {}	1✔
152	self.azure_ad_token_provider = azure_ad_token_provider	1✔
153
154	self.client = AzureOpenAI(	1✔
155	api_version=api_version,
156	azure_endpoint=azure_endpoint,
157	azure_deployment=azure_deployment,
158	azure_ad_token_provider=azure_ad_token_provider,
159	api_key=api_key.resolve_value() if api_key is not None else None,
160	azure_ad_token=azure_ad_token.resolve_value() if azure_ad_token is not None else None,
161	organization=organization,
162	timeout=self.timeout,
163	max_retries=self.max_retries,
164	http_client=init_http_client(self.http_client_kwargs, async_client=False),
165	default_headers=self.default_headers,
166	)
167
168	def to_dict(self) -> dict[str, Any]:	1✔
169	"""
170	Serialize this component to a dictionary.
171
172	:returns:
173	The serialized component as a dictionary.
174	"""
175	callback_name = serialize_callable(self.streaming_callback) if self.streaming_callback else None	1✔
176	azure_ad_token_provider_name = None	1✔
177	if self.azure_ad_token_provider:	1✔
178	azure_ad_token_provider_name = serialize_callable(self.azure_ad_token_provider)	1✔
179	return default_to_dict(	1✔
180	self,
181	azure_endpoint=self.azure_endpoint,
182	azure_deployment=self.azure_deployment,
183	organization=self.organization,
184	api_version=self.api_version,
185	streaming_callback=callback_name,
186	generation_kwargs=self.generation_kwargs,
187	system_prompt=self.system_prompt,
188	api_key=self.api_key.to_dict() if self.api_key is not None else None,
189	azure_ad_token=self.azure_ad_token.to_dict() if self.azure_ad_token is not None else None,
190	timeout=self.timeout,
191	max_retries=self.max_retries,
192	http_client_kwargs=self.http_client_kwargs,
193	default_headers=self.default_headers,
194	azure_ad_token_provider=azure_ad_token_provider_name,
195	)
196
197	@classmethod	1✔
198	def from_dict(cls, data: dict[str, Any]) -> "AzureOpenAIGenerator":	1✔
199	"""
200	Deserialize this component from a dictionary.
201
202	:param data:
203	The dictionary representation of this component.
204	:returns:
205	The deserialized component instance.
206	"""
207	deserialize_secrets_inplace(data["init_parameters"], keys=["api_key", "azure_ad_token"])	1✔
208	init_params = data.get("init_parameters", {})	1✔
209	serialized_callback_handler = init_params.get("streaming_callback")	1✔
210	if serialized_callback_handler:	1✔
211	data["init_parameters"]["streaming_callback"] = deserialize_callable(serialized_callback_handler)	×
212	serialized_azure_ad_token_provider = init_params.get("azure_ad_token_provider")	1✔
213	if serialized_azure_ad_token_provider:	1✔
214	data["init_parameters"]["azure_ad_token_provider"] = deserialize_callable(	×
215	serialized_azure_ad_token_provider
216	)
217	return default_from_dict(cls, data)	1✔

deepset-ai / haystack / 18592817487

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