Struct ProxyContext 

pub struct ProxyContext {

    pub start_time: Instant,
    pub request_body: Vec<u8>,
    pub model: Option<String>,
    pub selected_backend: Option<BackendInfo>,
    pub provider_name: Option<String>,
    pub stream_state: Option<StreamState>,
    pub response_body: Vec<u8>,
    pub is_streaming: bool,
    pub rewritten_path: Option<String>,
    pub is_stream_response: bool,
    pub is_conversion_request: bool,
    pub stream_body_parsed_offset: usize,
    pub normalized_path: Option<String>,
    pub should_convert_stream_response: bool,
    pub upstream_status: Option<u16>,
    pub upstream_content_type: Option<String>,
    pub stream_chunks_parsed: usize,
    pub pending_conversation_messages: Option<Vec<ChatMessage>>,
    pub pending_instructions: Option<String>,
}

Proxy context attached to each request session.

Fields

start_time: Instant

Request start time for duration tracking.

request_body: Vec<u8>

Collected request body bytes.

model: Option<String>

Model name parsed from request.

selected_backend: Option<BackendInfo>

Selected backend for this request.

provider_name: Option<String>

Provider name.

stream_state: Option<StreamState>

Stream state for SSE conversion (also used for non-streaming conversion context).

response_body: Vec<u8>

Response body collected for conversion.

is_streaming: bool

Whether streaming is enabled.

rewritten_path: Option<String>

Rewritten upstream path.

is_stream_response: bool

Whether this is a streaming response (for conversion tracking).

is_conversion_request: bool

Whether this is a conversion request (Responses API -> Chat API).

stream_body_parsed_offset: usize

Offset in response_body that has been parsed (to avoid re-parsing events).

normalized_path: Option<String>

Request path after optional routing prefix stripping.

should_convert_stream_response: bool

Whether current upstream response should be converted as SSE stream.

upstream_status: Option<u16>

Upstream status code captured in response_filter for diagnostics.

upstream_content_type: Option<String>

Upstream content-type captured in response_filter for diagnostics.

stream_chunks_parsed: usize

Number of valid upstream chat stream chunks parsed.

pending_conversation_messages: Option<Vec<ChatMessage>>

Conversation messages before upstream response (for follow-up turn storage).

pending_instructions: Option<String>

Effective instructions used for this request after previous_response expansion.

Implementations

impl ProxyContext

pub fn new() -> Self

Create a new proxy context.

pub fn init_from_request_body(&mut self)

Parse model name and stream flag from request body. Initializes StreamState for ALL conversion requests (both streaming and non-streaming). StreamState holds ResponseRequestContext for protocol-aligned response generation.

pub fn set_response_request_context(&mut self, context: ResponseRequestContext)

Set the response request context from a parsed ResponseRequest. This should be called during request_body_filter processing.

Trait Implementations

impl Debug for ProxyContext

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for ProxyContext

fn default() -> Self

Returns the “default value” for a type.