Httpx Client

infra.llm.httpx_client

Summary

HTTP client ports for OpenAI-compatible chat completion endpoints.

---

Provide synchronous and asynchronous httpx wrappers that issue chat completion requests with bearer authentication.

Classes

AsyncClientPort

AsyncClientPort(
    api_key: str = "",
    endpoint: str = "",
    timeout: float = 60.0,
)

Asynchronous httpx client for chat completion APIs.

Attributes:

Name	Type	Description
api_key	str	Bearer token used for authentication.
endpoint	str	Base URL of the completions API.
timeout	float	Request timeout in seconds.
_http_client	AsyncClient	Underlying asynchronous HTTP client.

Initialize the asynchronous client.

Parameters:

Name	Type	Description	Default
api_key	str	Bearer token used for authentication.	''
endpoint	str	Base URL of the completions API.	''
timeout	float	Request timeout in seconds.	60.0

Functions

aclose async

aclose()

Close the underlying asynchronous HTTP client.

chat_completion async

chat_completion(
    model: str,
    messages: list,
    temperature: float = 0.0,
    max_tokens: int = 4096,
) -> dict

Send an asynchronous chat completion request.

Parameters:

Name	Type	Description	Default
model	str	Model identifier to invoke.	required
messages	list	Chat messages in API format.	required
temperature	float	Sampling temperature.	0.0
max_tokens	int	Maximum tokens to generate.	4096

Returns:

Name	Type	Description
dict	dict	Decoded JSON response body.

SyncClientPort

SyncClientPort(
    api_key: str = "",
    endpoint: str = "",
    timeout: float = 60.0,
)

Synchronous httpx client for chat completion APIs.

Attributes:

Name	Type	Description
api_key	str	Bearer token used for authentication.
endpoint	str	Base URL of the completions API.
timeout	float	Request timeout in seconds.
_client	Client	Underlying synchronous HTTP client.

Initialize the synchronous client.

Parameters:

Name	Type	Description	Default
api_key	str	Bearer token used for authentication.	''
endpoint	str	Base URL of the completions API.	''
timeout	float	Request timeout in seconds.	60.0

Functions

chat_completion

chat_completion(
    model: str,
    messages: list,
    temperature: float = 0.0,
    max_tokens: int = 4096,
) -> dict

Send a chat completion request.

Parameters:

Name	Type	Description	Default
model	str	Model identifier to invoke.	required
messages	list	Chat messages in API format.	required
temperature	float	Sampling temperature.	0.0
max_tokens	int	Maximum tokens to generate.	4096

Returns:

Name	Type	Description
dict	dict	Decoded JSON response body.

close

close()

Close the underlying HTTP client.