rig
rig copied to clipboard
bug: Inconsistencies between id and call_id for tool calling
- [x] I have looked for existing issues (including closed) about this
Bug Report
I am confused by the call_id vs id usage, all providers (including openai in the completion API) use the id for the tool_call_id but the response API uses the call_id but not the id. So I am wondering why it was added in the first place and if there is a way to make that a bit more consistent.
Also there are a lot of expect in with call_id (example). This is not great, they should be refactored.
Dug a bit into it:
arguments
string
A JSON string of the arguments to pass to the function.
call_id
string
The unique ID of the function tool call generated by the model.
name
string
The name of the function to run.
type
string
The type of the function tool call. Always function_call.
id
string
The unique ID of the function tool call.
status
string
The status of the item. One of in_progress, completed, or incomplete. Populated when items are returned via API.