14591523099

Committed 22 Apr 2025 09:32AM UTC coverage: 90.514% (+0.01%) from 90.504%

Build # 14591523099

Build Type

Pull #9270

github

Committed by

web-flow

Commit Message

Merge 42a33e27f into 114b4568b

Pull Request Pull Request #9270: feat: Allow OpenAI client config in other components

Run Details

10830 of 11965 relevant lines covered (90.51%)

0.91 hits per line

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

88.68

haystack/components/embedders/openai_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 import AsyncOpenAI, OpenAI
from openai.types import CreateEmbeddingResponse

from haystack import component, default_from_dict, default_to_dict
from haystack.utils import Secret, deserialize_secrets_inplace
from haystack.utils.http_client import init_http_client


@component
class OpenAITextEmbedder:
    """
    Embeds strings using OpenAI models.

    You can use it to embed user query and send it to an embedding Retriever.

    ### Usage example

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

    text_to_embed = "I love pizza!"

    text_embedder = OpenAITextEmbedder()

    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,
        api_key: Secret = Secret.from_env_var("OPENAI_API_KEY"),
        model: str = "text-embedding-ada-002",
        dimensions: Optional[int] = None,
        api_base_url: Optional[str] = None,
        organization: Optional[str] = None,
        prefix: str = "",
        suffix: str = "",
        timeout: Optional[float] = None,
        max_retries: Optional[int] = None,
        http_client_kwargs: Optional[Dict[str, Any]] = None,
    ):
        """
        Creates an OpenAITextEmbedder component.

        Before initializing the component, you can set the 'OPENAI_TIMEOUT' and 'OPENAI_MAX_RETRIES'
        environment variables to override the `timeout` and `max_retries` parameters respectively
        in the OpenAI client.

        :param api_key:
            The OpenAI API key.
            You can set it with an environment variable `OPENAI_API_KEY`, or pass with this parameter
            during initialization.
        :param model:
            The name of the model to use for calculating embeddings.
            The default model is `text-embedding-ada-002`.
        :param dimensions:
            The number of dimensions of the resulting embeddings. Only `text-embedding-3` and
            later models support this parameter.
        :param api_base_url:
            Overrides default base URL for all HTTP requests.
        :param organization:
            Your organization ID. See OpenAI's
            [production best practices](https://platform.openai.com/docs/guides/production-best-practices/setting-up-your-organization)
            for more information.
        :param prefix:
            A string to add at the beginning of each text to embed.
        :param suffix:
            A string to add at the end of each text to embed.
        :param timeout:
            Timeout for OpenAI client calls. If not set, it defaults to either the
            `OPENAI_TIMEOUT` environment variable, or 30 seconds.
        :param max_retries:
            Maximum number of retries to contact OpenAI after an internal error.
            If not set, it defaults to either 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).
        """
        self.model = model
        self.dimensions = dimensions
        self.api_base_url = api_base_url
        self.organization = organization
        self.prefix = prefix
        self.suffix = suffix
        self.api_key = api_key
        self.http_client_kwargs = http_client_kwargs

        if timeout is None:
            timeout = float(os.environ.get("OPENAI_TIMEOUT", "30.0"))
        if max_retries is None:
            max_retries = int(os.environ.get("OPENAI_MAX_RETRIES", "5"))

        client_kwargs: Dict[str, Any] = {
            "api_key": api_key.resolve_value(),
            "organization": organization,
            "base_url": api_base_url,
            "timeout": timeout,
            "max_retries": max_retries,
        }

        self.client = OpenAI(http_client=init_http_client(self.http_client_kwargs, async_client=False), **client_kwargs)
        self.async_client = AsyncOpenAI(
            http_client=init_http_client(self.http_client_kwargs, async_client=True), **client_kwargs
        )

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

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

        :returns:
            Dictionary with serialized data.
        """
        return default_to_dict(
            self,
            model=self.model,
            api_base_url=self.api_base_url,
            organization=self.organization,
            prefix=self.prefix,
            suffix=self.suffix,
            dimensions=self.dimensions,
            api_key=self.api_key.to_dict(),
            http_client_kwargs=self.http_client_kwargs,
        )

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

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

    def _prepare_input(self, text: str) -> Dict[str, Any]:
        if not isinstance(text, str):
            raise TypeError(
                "OpenAITextEmbedder expects a string as an input."
                "In case you want to embed a list of Documents, please use the OpenAIDocumentEmbedder."
            )

        text_to_embed = self.prefix + text + self.suffix

        kwargs: Dict[str, Any] = {"model": self.model, "input": text_to_embed}
        if self.dimensions is not None:
            kwargs["dimensions"] = self.dimensions
        return kwargs

    def _prepare_output(self, result: CreateEmbeddingResponse) -> Dict[str, Any]:
        return {"embedding": result.data[0].embedding, "meta": {"model": result.model, "usage": dict(result.usage)}}

    @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.
        """
        create_kwargs = self._prepare_input(text=text)
        response = self.client.embeddings.create(**create_kwargs)
        return self._prepare_output(result=response)

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

        This is the asynchronous version of the `run` method. It has the same parameters and return values
        but can be used with `await` in async code.

        :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.
        """
        create_kwargs = self._prepare_input(text=text)
        response = await self.async_client.embeddings.create(**create_kwargs)
        return self._prepare_output(result=response)

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 import AsyncOpenAI, OpenAI	1✔
9	from openai.types import CreateEmbeddingResponse	1✔
10
11	from haystack import component, default_from_dict, default_to_dict	1✔
12	from haystack.utils import Secret, deserialize_secrets_inplace	1✔
13	from haystack.utils.http_client import init_http_client	1✔
14
15
16	@component	1✔
17	class OpenAITextEmbedder:	1✔
18	"""
19	Embeds strings using OpenAI models.
20
21	You can use it to embed user query and send it to an embedding Retriever.
22
23	### Usage example
24
25	```python
26	from haystack.components.embedders import OpenAITextEmbedder
27
28	text_to_embed = "I love pizza!"
29
30	text_embedder = OpenAITextEmbedder()
31
32	print(text_embedder.run(text_to_embed))
33
34	# {'embedding': [0.017020374536514282, -0.023255806416273117, ...],
35	# 'meta': {'model': 'text-embedding-ada-002-v2',
36	# 'usage': {'prompt_tokens': 4, 'total_tokens': 4}}}
37	```
38	"""
39
40	def __init__( # pylint: disable=too-many-positional-arguments	1✔
41	self,
42	api_key: Secret = Secret.from_env_var("OPENAI_API_KEY"),
43	model: str = "text-embedding-ada-002",
44	dimensions: Optional[int] = None,
45	api_base_url: Optional[str] = None,
46	organization: Optional[str] = None,
47	prefix: str = "",
48	suffix: str = "",
49	timeout: Optional[float] = None,
50	max_retries: Optional[int] = None,
51	http_client_kwargs: Optional[Dict[str, Any]] = None,
52	):
53	"""
54	Creates an OpenAITextEmbedder component.
55
56	Before initializing the component, you can set the 'OPENAI_TIMEOUT' and 'OPENAI_MAX_RETRIES'
57	environment variables to override the `timeout` and `max_retries` parameters respectively
58	in the OpenAI client.
59
60	:param api_key:
61	The OpenAI API key.
62	You can set it with an environment variable `OPENAI_API_KEY`, or pass with this parameter
63	during initialization.
64	:param model:
65	The name of the model to use for calculating embeddings.
66	The default model is `text-embedding-ada-002`.
67	:param dimensions:
68	The number of dimensions of the resulting embeddings. Only `text-embedding-3` and
69	later models support this parameter.
70	:param api_base_url:
71	Overrides default base URL for all HTTP requests.
72	:param organization:
73	Your organization ID. See OpenAI's
74	[production best practices](https://platform.openai.com/docs/guides/production-best-practices/setting-up-your-organization)
75	for more information.
76	:param prefix:
77	A string to add at the beginning of each text to embed.
78	:param suffix:
79	A string to add at the end of each text to embed.
80	:param timeout:
81	Timeout for OpenAI client calls. If not set, it defaults to either the
82	`OPENAI_TIMEOUT` environment variable, or 30 seconds.
83	:param max_retries:
84	Maximum number of retries to contact OpenAI after an internal error.
85	If not set, it defaults to either the `OPENAI_MAX_RETRIES` environment variable, or set to 5.
86	:param http_client_kwargs:
87	A dictionary of keyword arguments to configure a custom `httpx.Client`or `httpx.AsyncClient`.
88	For more information, see the [HTTPX documentation](https://www.python-httpx.org/api/#client).
89	"""
90	self.model = model	1✔
91	self.dimensions = dimensions	1✔
92	self.api_base_url = api_base_url	1✔
93	self.organization = organization	1✔
94	self.prefix = prefix	1✔
95	self.suffix = suffix	1✔
96	self.api_key = api_key	1✔
97	self.http_client_kwargs = http_client_kwargs	1✔
98
99	if timeout is None:	1✔
100	timeout = float(os.environ.get("OPENAI_TIMEOUT", "30.0"))	1✔
101	if max_retries is None:	1✔
102	max_retries = int(os.environ.get("OPENAI_MAX_RETRIES", "5"))	1✔
103
104	client_kwargs: Dict[str, Any] = {	1✔
105	"api_key": api_key.resolve_value(),
106	"organization": organization,
107	"base_url": api_base_url,
108	"timeout": timeout,
109	"max_retries": max_retries,
110	}
111
112	self.client = OpenAI(http_client=init_http_client(self.http_client_kwargs, async_client=False), **client_kwargs)	1✔
113	self.async_client = AsyncOpenAI(	1✔
114	http_client=init_http_client(self.http_client_kwargs, async_client=True), **client_kwargs
115	)
116
117	def _get_telemetry_data(self) -> Dict[str, Any]:	1✔
118	"""
119	Data that is sent to Posthog for usage analytics.
120	"""
121	return {"model": self.model}	×
122
123	def to_dict(self) -> Dict[str, Any]:	1✔
124	"""
125	Serializes the component to a dictionary.
126
127	:returns:
128	Dictionary with serialized data.
129	"""
130	return default_to_dict(	1✔
131	self,
132	model=self.model,
133	api_base_url=self.api_base_url,
134	organization=self.organization,
135	prefix=self.prefix,
136	suffix=self.suffix,
137	dimensions=self.dimensions,
138	api_key=self.api_key.to_dict(),
139	http_client_kwargs=self.http_client_kwargs,
140	)
141
142	@classmethod	1✔
143	def from_dict(cls, data: Dict[str, Any]) -> "OpenAITextEmbedder":	1✔
144	"""
145	Deserializes the component from a dictionary.
146
147	:param data:
148	Dictionary to deserialize from.
149	:returns:
150	Deserialized component.
151	"""
152	deserialize_secrets_inplace(data["init_parameters"], keys=["api_key"])	1✔
153	return default_from_dict(cls, data)	1✔
154
155	def _prepare_input(self, text: str) -> Dict[str, Any]:	1✔
156	if not isinstance(text, str):	1✔
157	raise TypeError(	1✔
158	"OpenAITextEmbedder expects a string as an input."
159	"In case you want to embed a list of Documents, please use the OpenAIDocumentEmbedder."
160	)
161
162	text_to_embed = self.prefix + text + self.suffix	1✔
163
164	kwargs: Dict[str, Any] = {"model": self.model, "input": text_to_embed}	1✔
165	if self.dimensions is not None:	1✔
166	kwargs["dimensions"] = self.dimensions	1✔
167	return kwargs	1✔
168
169	def _prepare_output(self, result: CreateEmbeddingResponse) -> Dict[str, Any]:	1✔
170	return {"embedding": result.data[0].embedding, "meta": {"model": result.model, "usage": dict(result.usage)}}	1✔
171
172	@component.output_types(embedding=List[float], meta=Dict[str, Any])	1✔
173	def run(self, text: str):	1✔
174	"""
175	Embeds a single string.
176
177	:param text:
178	Text to embed.
179
180	:returns:
181	A dictionary with the following keys:
182	- `embedding`: The embedding of the input text.
183	- `meta`: Information about the usage of the model.
184	"""
185	create_kwargs = self._prepare_input(text=text)	1✔
186	response = self.client.embeddings.create(**create_kwargs)	×
187	return self._prepare_output(result=response)	×
188
189	@component.output_types(embedding=List[float], meta=Dict[str, Any])	1✔
190	async def run_async(self, text: str):	1✔
191	"""
192	Asynchronously embed a single string.
193
194	This is the asynchronous version of the `run` method. It has the same parameters and return values
195	but can be used with `await` in async code.
196
197	:param text:
198	Text to embed.
199
200	:returns:
201	A dictionary with the following keys:
202	- `embedding`: The embedding of the input text.
203	- `meta`: Information about the usage of the model.
204	"""
205	create_kwargs = self._prepare_input(text=text)	×
206	response = await self.async_client.embeddings.create(**create_kwargs)	×
207	return self._prepare_output(result=response)	×

deepset-ai / haystack / 14591523099

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