Ask for **httpx_kwargs, not httpx_client (#1)

Author: vrslev

README.md

@@ -144,7 +144,8 @@ client = any_llm_client.OpenAIClient(config, ...)
 
 #### Timeouts, proxy & other HTTP settings
 
-Pass custom [HTTPX](https://www.python-httpx.org) client:
+
+Pass custom [HTTPX](https://www.python-httpx.org) kwargs to `any_llm_client.get_client()`:
 
 ```python
 import httpx
@@ -154,14 +155,14 @@ import any_llm_client
 
 async with any_llm_client.get_client(
     ...,
-    httpx_client=httpx.AsyncClient(
-        mounts={"https://api.openai.com": httpx.AsyncHTTPTransport(proxy="http://localhost:8030")},
-        timeout=httpx.Timeout(None, connect=5.0),
-    ),
+    mounts={"https://api.openai.com": httpx.AsyncHTTPTransport(proxy="http://localhost:8030")},
+    timeout=httpx.Timeout(None, connect=5.0),
 ) as client:
     ...
 ```
 
+Default timeout is `httpx.Timeout(None, connect=5.0)` (5 seconds on connect, unlimited on read, write or pool).
+
 #### Retries
 
 By default, requests are retried 3 times on HTTP status errors. You can change the retry behaviour by supplying `request_retry` parameter:

any_llm_client/clients/openai.py

@@ -11,7 +11,7 @@
 import typing_extensions
 
 from any_llm_client.core import LLMClient, LLMConfig, LLMError, Message, MessageRole, OutOfTokensOrSymbolsError
-from any_llm_client.http import make_http_request, make_streaming_http_request
+from any_llm_client.http import get_http_client_from_kwargs, make_http_request, make_streaming_http_request
 from any_llm_client.retry import RequestRetryConfig
 
 
@@ -101,12 +101,13 @@ class OpenAIClient(LLMClient):
     def __init__(
         self,
         config: OpenAIConfig,
-        httpx_client: httpx.AsyncClient | None = None,
+        *,
         request_retry: RequestRetryConfig | None = None,
+        **httpx_kwargs: typing.Any,  # noqa: ANN401
     ) -> None:
         self.config = config
-        self.httpx_client = httpx_client or httpx.AsyncClient()
         self.request_retry = request_retry or RequestRetryConfig()
+        self.httpx_client = get_http_client_from_kwargs(httpx_kwargs)
 
     def _build_request(self, payload: dict[str, typing.Any]) -> httpx.Request:
         return self.httpx_client.build_request(

any_llm_client/clients/yandexgpt.py

@@ -10,7 +10,7 @@
 import typing_extensions
 
 from any_llm_client.core import LLMClient, LLMConfig, LLMError, Message, OutOfTokensOrSymbolsError
-from any_llm_client.http import make_http_request, make_streaming_http_request
+from any_llm_client.http import get_http_client_from_kwargs, make_http_request, make_streaming_http_request
 from any_llm_client.retry import RequestRetryConfig
 
 
@@ -70,12 +70,13 @@ class YandexGPTClient(LLMClient):
     def __init__(
         self,
         config: YandexGPTConfig,
-        httpx_client: httpx.AsyncClient | None = None,
+        *,
         request_retry: RequestRetryConfig | None = None,
+        **httpx_kwargs: typing.Any,  # noqa: ANN401
     ) -> None:
         self.config = config
-        self.httpx_client = httpx_client or httpx.AsyncClient()
         self.request_retry = request_retry or RequestRetryConfig()
+        self.httpx_client = get_http_client_from_kwargs(httpx_kwargs)
 
     def _build_request(self, payload: dict[str, typing.Any]) -> httpx.Request:
         headers: typing.Final = {"x-data-logging-enabled": "false"}

any_llm_client/http.py

@@ -8,6 +8,15 @@
 from any_llm_client.retry import RequestRetryConfig
 
 
+DEFAULT_HTTP_TIMEOUT: typing.Final = httpx.Timeout(None, connect=5.0)
+
+
+def get_http_client_from_kwargs(kwargs: dict[str, typing.Any]) -> httpx.AsyncClient:
+    kwargs_with_defaults: typing.Final = kwargs.copy()
+    kwargs_with_defaults.setdefault("timeout", DEFAULT_HTTP_TIMEOUT)
+    return httpx.AsyncClient(**kwargs_with_defaults)
+
+
 async def make_http_request(
     *,
     httpx_client: httpx.AsyncClient,

any_llm_client/main.py

@@ -1,8 +1,6 @@
 import functools
 import typing
 
-import httpx
-
 from any_llm_client.clients.mock import MockLLMClient, MockLLMConfig
 from any_llm_client.clients.openai import OpenAIClient, OpenAIConfig
 from any_llm_client.clients.yandexgpt import YandexGPTClient, YandexGPTConfig
@@ -18,43 +16,43 @@
     def get_client(
         config: AnyLLMConfig,
         *,
-        httpx_client: httpx.AsyncClient | None = None,
         request_retry: RequestRetryConfig | None = None,
+        **httpx_kwargs: typing.Any,  # noqa: ANN401
     ) -> LLMClient: ...  # pragma: no cover
 else:
 
     @functools.singledispatch
     def get_client(
         config: typing.Any,  # noqa: ANN401, ARG001
         *,
-        httpx_client: httpx.AsyncClient | None = None,  # noqa: ARG001
         request_retry: RequestRetryConfig | None = None,  # noqa: ARG001
+        **httpx_kwargs: typing.Any,  # noqa: ANN401, ARG001
     ) -> LLMClient:
         raise AssertionError("unknown LLM config type")
 
     @get_client.register
     def _(
         config: YandexGPTConfig,
         *,
-        httpx_client: httpx.AsyncClient | None = None,
         request_retry: RequestRetryConfig | None = None,
+        **httpx_kwargs: typing.Any,  # noqa: ANN401
     ) -> LLMClient:
-        return YandexGPTClient(config=config, httpx_client=httpx_client, request_retry=request_retry)
+        return YandexGPTClient(config=config, request_retry=request_retry, **httpx_kwargs)
 
     @get_client.register
     def _(
         config: OpenAIConfig,
         *,
-        httpx_client: httpx.AsyncClient | None = None,
         request_retry: RequestRetryConfig | None = None,
+        **httpx_kwargs: typing.Any,  # noqa: ANN401
     ) -> LLMClient:
-        return OpenAIClient(config=config, httpx_client=httpx_client, request_retry=request_retry)
+        return OpenAIClient(config=config, request_retry=request_retry, **httpx_kwargs)
 
     @get_client.register
     def _(
         config: MockLLMConfig,
         *,
-        httpx_client: httpx.AsyncClient | None = None,  # noqa: ARG001
         request_retry: RequestRetryConfig | None = None,  # noqa: ARG001
+        **httpx_kwargs: typing.Any,  # noqa: ANN401, ARG001
     ) -> LLMClient:
         return MockLLMClient(config=config)

tests/test_http.py

@@ -0,0 +1,27 @@
+import copy
+import typing
+
+import httpx
+
+from any_llm_client.http import DEFAULT_HTTP_TIMEOUT, get_http_client_from_kwargs
+
+
+class TestGetHttpClientFromKwargs:
+    def test_http_timeout_is_added(self) -> None:
+        original_kwargs: typing.Final = {"mounts": {"http://": None}}
+        passed_kwargs: typing.Final = copy.deepcopy(original_kwargs)
+
+        result: typing.Final = get_http_client_from_kwargs(passed_kwargs)
+
+        assert result.timeout == DEFAULT_HTTP_TIMEOUT
+        assert original_kwargs == passed_kwargs
+
+    def test_http_timeout_is_not_modified_if_set(self) -> None:
+        timeout: typing.Final = httpx.Timeout(7, connect=5, read=3)
+        original_kwargs: typing.Final = {"mounts": {"http://": None}, "timeout": timeout}
+        passed_kwargs: typing.Final = copy.deepcopy(original_kwargs)
+
+        result: typing.Final = get_http_client_from_kwargs(passed_kwargs)
+
+        assert result.timeout == timeout
+        assert original_kwargs == passed_kwargs

tests/test_mock_client.py

@@ -1,5 +1,4 @@
 import typing
-from unittest import mock
 
 from polyfactory.factories.pydantic_factory import ModelFactory
 
@@ -12,7 +11,7 @@ class MockLLMConfigFactory(ModelFactory[any_llm_client.MockLLMConfig]): ...
 
 async def test_mock_client_request_llm_message_returns_config_value() -> None:
     config: typing.Final = MockLLMConfigFactory.build()
-    response: typing.Final = await any_llm_client.get_client(config, httpx_client=mock.Mock()).request_llm_message(
+    response: typing.Final = await any_llm_client.get_client(config).request_llm_message(
         **LLMFuncRequestFactory.build()
     )
     assert response == config.response_message
@@ -21,8 +20,6 @@ async def test_mock_client_request_llm_message_returns_config_value() -> None:
 async def test_mock_client_request_llm_partial_responses_returns_config_value() -> None:
     config: typing.Final = MockLLMConfigFactory.build()
     response: typing.Final = await consume_llm_partial_responses(
-        any_llm_client.get_client(config, httpx_client=mock.Mock()).stream_llm_partial_messages(
-            **LLMFuncRequestFactory.build()
-        )
+        any_llm_client.get_client(config).stream_llm_partial_messages(**LLMFuncRequestFactory.build())
     )
     assert response == config.stream_messages

tests/test_openai_client.py

@@ -1,5 +1,4 @@
 import typing
-from unittest import mock
 
 import faker
 import httpx
@@ -35,8 +34,7 @@ async def test_ok(self, faker: faker.Faker) -> None:
         )
 
         result: typing.Final = await any_llm_client.get_client(
-            OpenAIConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
+            OpenAIConfigFactory.build(), transport=httpx.MockTransport(lambda _: response)
         ).request_llm_message(**LLMFuncRequestFactory.build())
 
         assert result == expected_result
@@ -47,8 +45,7 @@ async def test_fails_without_alternatives(self) -> None:
             json=ChatCompletionsNotStreamingResponse.model_construct(choices=[]).model_dump(mode="json"),
         )
         client: typing.Final = any_llm_client.get_client(
-            OpenAIConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
+            OpenAIConfigFactory.build(), transport=httpx.MockTransport(lambda _: response)
         )
 
         with pytest.raises(pydantic.ValidationError):
@@ -89,10 +86,7 @@ async def test_ok(self, faker: faker.Faker) -> None:
         response: typing.Final = httpx.Response(
             200, headers={"Content-Type": "text/event-stream"}, content=response_content
         )
-        client: typing.Final = any_llm_client.get_client(
-            config,
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
-        )
+        client: typing.Final = any_llm_client.get_client(config, transport=httpx.MockTransport(lambda _: response))
 
         result: typing.Final = await consume_llm_partial_responses(client.stream_llm_partial_messages(**func_request))
 
@@ -106,8 +100,7 @@ async def test_fails_without_alternatives(self) -> None:
             200, headers={"Content-Type": "text/event-stream"}, content=response_content
         )
         client: typing.Final = any_llm_client.get_client(
-            OpenAIConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
+            OpenAIConfigFactory.build(), transport=httpx.MockTransport(lambda _: response)
         )
 
         with pytest.raises(pydantic.ValidationError):
@@ -119,8 +112,7 @@ class TestOpenAILLMErrors:
     @pytest.mark.parametrize("status_code", [400, 500])
     async def test_fails_with_unknown_error(self, stream: bool, status_code: int) -> None:
         client: typing.Final = any_llm_client.get_client(
-            OpenAIConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: httpx.Response(status_code))),
+            OpenAIConfigFactory.build(), transport=httpx.MockTransport(lambda _: httpx.Response(status_code))
         )
 
         coroutine: typing.Final = (
@@ -144,8 +136,7 @@ async def test_fails_with_unknown_error(self, stream: bool, status_code: int) ->
     async def test_fails_with_out_of_tokens_error(self, stream: bool, content: bytes | None) -> None:
         response: typing.Final = httpx.Response(400, content=content)
         client: typing.Final = any_llm_client.get_client(
-            OpenAIConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
+            OpenAIConfigFactory.build(), transport=httpx.MockTransport(lambda _: response)
         )
 
         coroutine: typing.Final = (
@@ -244,14 +235,13 @@ def test_with_alternation(
         self, messages: list[any_llm_client.Message], expected_result: list[ChatCompletionsMessage]
     ) -> None:
         client: typing.Final = any_llm_client.OpenAIClient(
-            config=OpenAIConfigFactory.build(force_user_assistant_message_alternation=True), httpx_client=mock.Mock()
+            OpenAIConfigFactory.build(force_user_assistant_message_alternation=True)
         )
         assert client._prepare_messages(messages) == expected_result  # noqa: SLF001
 
     def test_without_alternation(self) -> None:
         client: typing.Final = any_llm_client.OpenAIClient(
-            config=OpenAIConfigFactory.build(force_user_assistant_message_alternation=False),
-            httpx_client=mock.Mock(),
+            OpenAIConfigFactory.build(force_user_assistant_message_alternation=False)
         )
         assert client._prepare_messages(  # noqa: SLF001
             [

tests/test_unknown_client.py

@@ -1,5 +1,3 @@
-from unittest import mock
-
 import faker
 import pytest
 
@@ -8,4 +6,4 @@
 
 def test_unknown_client_raises_assertion_error(faker: faker.Faker) -> None:
     with pytest.raises(AssertionError):
-        any_llm_client.get_client(faker.pyobject(), httpx_client=mock.Mock())  # type: ignore[arg-type]
+        any_llm_client.get_client(faker.pyobject())  # type: ignore[arg-type]

tests/test_yandexgpt_client.py

@@ -29,8 +29,7 @@ async def test_ok(self, faker: faker.Faker) -> None:
         )
 
         result: typing.Final = await any_llm_client.get_client(
-            YandexGPTConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
+            YandexGPTConfigFactory.build(), transport=httpx.MockTransport(lambda _: response)
         ).request_llm_message(**LLMFuncRequestFactory.build())
 
         assert result == expected_result
@@ -40,8 +39,7 @@ async def test_fails_without_alternatives(self) -> None:
             200, json=YandexGPTResponse(result=YandexGPTResult.model_construct(alternatives=[])).model_dump(mode="json")
         )
         client: typing.Final = any_llm_client.get_client(
-            YandexGPTConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
+            YandexGPTConfigFactory.build(), transport=httpx.MockTransport(lambda _: response)
         )
 
         with pytest.raises(pydantic.ValidationError):
@@ -70,7 +68,7 @@ async def test_ok(self, faker: faker.Faker) -> None:
 
         result: typing.Final = await consume_llm_partial_responses(
             any_llm_client.get_client(
-                config, httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response))
+                config, transport=httpx.MockTransport(lambda _: response)
             ).stream_llm_partial_messages(**func_request)
         )
 
@@ -83,8 +81,7 @@ async def test_fails_without_alternatives(self) -> None:
         response: typing.Final = httpx.Response(200, content=response_content)
 
         client: typing.Final = any_llm_client.get_client(
-            YandexGPTConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
+            YandexGPTConfigFactory.build(), transport=httpx.MockTransport(lambda _: response)
         )
 
         with pytest.raises(pydantic.ValidationError):
@@ -96,8 +93,7 @@ class TestYandexGPTLLMErrors:
     @pytest.mark.parametrize("status_code", [400, 500])
     async def test_fails_with_unknown_error(self, stream: bool, status_code: int) -> None:
         client: typing.Final = any_llm_client.get_client(
-            YandexGPTConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: httpx.Response(status_code))),
+            YandexGPTConfigFactory.build(), transport=httpx.MockTransport(lambda _: httpx.Response(status_code))
         )
 
         coroutine: typing.Final = (
@@ -121,8 +117,7 @@ async def test_fails_with_unknown_error(self, stream: bool, status_code: int) ->
     async def test_fails_with_out_of_tokens_error(self, stream: bool, response_content: bytes | None) -> None:
         response: typing.Final = httpx.Response(400, content=response_content)
         client: typing.Final = any_llm_client.get_client(
-            YandexGPTConfigFactory.build(),
-            httpx_client=httpx.AsyncClient(transport=httpx.MockTransport(lambda _: response)),
+            YandexGPTConfigFactory.build(), transport=httpx.MockTransport(lambda _: response)
         )
 
         coroutine: typing.Final = (