Skip to main content

Module vlm

robonix_pilot

Module vlm 

Source

Structs§

AccumulatedToolCall 🔒
FnCall
FnDef 🔒
Message
One message in an OpenAI Chat Completions conversation. Spec: https://platform.openai.com/docs/api-reference/chat/create#chat/create-messages
ToolCall
ToolDef
VlmClient
Direct HTTP client for an OpenAI-compatible chat-completions endpoint. Cheap to clone — async_openai::Client wraps a reqwest::Client (an Arc<...> internally). No mutex needed when sharing across tasks.
VlmUsage

Enums§

VlmStreamItem
Item yielded by the chat completion stream. planner.rs matches on this enum to drive token streaming, tool dispatch, and finish handling.

Constants§

MAX_OPEN_RETRIES 🔒

Functions§

build_openai_messages 🔒
build_openai_tools 🔒
open_retry_delay 🔒
parse_usage 🔒
Read the standard Chat Completions usage shape without requiring every OpenAI-compatible provider to deserialize optional detail fields equally.
process_stream_line 🔒
rejects_optional_request_fields 🔒
Compatibility fallback is only safe when a client-error response names an optional field that Pilot added. Unrelated 4xx responses must retain their original diagnosis instead of being retried with a misleading warning.