apilist.txt

# cosette Module Documentation

## cosette.core

- `def find_block(r)`
    Find the message in `r`.

- `def contents(r)`
    Helper to get the contents from response `r`.

- `def usage(inp, out)`
    Slightly more concise version of `CompletionUsage`.

- `@patch def __add__(self, b)`
    Add together each of `input_tokens` and `output_tokens`

- `def wrap_latex(text, md)`
    Replace OpenAI LaTeX codes with markdown-compatible ones

- `class Client`
    - `def __init__(self, model, cli)`
        Basic LLM messages client.


- `@patch @delegates(Completions.create) def __call__(self, msgs, sp, maxtok, stream, **kwargs)`
    Make a call to LLM.

- `def mk_toolres(r, ns, obj)`
    Create a `tool_result` message from response `r`.

- `@patch @delegates(Client.__call__) def structured(self, msgs, tools, obj, ns, **kwargs)`
    Return the value of all tool calls (generally used for structured outputs)

- `class Chat`
    - `def __init__(self, model, cli, sp, tools, tool_choice)`
        OpenAI chat client.

    - `@property def use`

## cosette.toolloop

- `@patch @delegates(Completions.create) def toolloop(self, pr, max_steps, trace_func, cont_func, **kwargs)`
    Add prompt `pr` to dialog and get a response from the model, automatically following up with `tool_use` messages