Improve token counting; account for functions

[Pwuts]

Background

Part of #4799

When using the OpenAI Function Call API, the specified functions count towards token consumption. However, they are currently not accounted for in token calculation and compensation logic.

Changes

autogpt.llm.providers.openai:

• Make get_openai_command_spec take a CommandRegistry and Config as arguments instead of an Agent
• Add function count_openai_functions_tokens
• Add function format_function_specs_as_typescript_ns

autogpt.llm.base:

• Add "function" as a valid role in a message
• Add ResponseMessageDict
• In LLMResponse, disable unused attributes prompt_tokens_used, completion_tokens_used

autogpt.llm.utils:

• Account for functions when calculating max_tokens
• Use better response object type definitions
• token_counter > count_message_tokens: allow single message arguments

Documentation

x

Test Plan

CI

PR Quality Checklist

• My pull request is atomic and focuses on a single change.
• I have thoroughly tested my changes with multiple different prompts.
• I have considered potential risks and mitigations for my changes.
• I have documented my changes clearly and comprehensively.
• I have not snuck in any "extra" small tweaks changes.
• I have run the following commands against my code to ensure it passes our linters:

  black .
  isort .
  mypy
  autoflake --remove-all-unused-imports --recursive --ignore-init-module-imports --ignore-pass-after-docstring autogpt tests --in-place

[netlify]

✅ Deploy Preview for auto-gpt-docs canceled.

Name	Link
🔨 Latest commit	17ed047
🔍 Latest deploy log	https://app.netlify.com/sites/auto-gpt-docs/deploys/64aaf9f8e7faff00081d4fd5

[Auto-GPT-Bot]

You changed AutoGPT's behaviour. The cassettes have been updated and will be merged to the submodule when this Pull Request gets merged.

[codecov]

Codecov Report

Patch coverage: 64.28% and no project coverage change.

Comparison is base (f6da992) 50.55% compared to head (60f1a87) 50.56%.

❗ Current head 60f1a87 differs from pull request most recent head 7c140b0. Consider uploading reports for the commit 7c140b0 to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #4919   +/-   ##
=======================================
  Coverage   50.55%   50.56%           
=======================================
  Files         116      115    -1     
  Lines        4860     4818   -42     
  Branches      657      638   -19     
=======================================
- Hits         2457     2436   -21     
+ Misses       2219     2205   -14     
+ Partials      184      177    -7     

Impacted Files	Coverage Δ
autogpt/llm/chat.py	61.44% <33.33%> (-4.79%)	⬇️
autogpt/llm/utils/token_counter.py	77.50% <55.55%> (-6.88%)	⬇️
autogpt/llm/providers/openai.py	88.79% <60.00%> (-2.04%)	⬇️
autogpt/llm/utils/__init__.py	48.00% <60.00%> (-0.58%)	⬇️
autogpt/agent/__init__.py	100.00% <100.00%> (ø)
autogpt/llm/base.py	93.33% <100.00%> (+2.42%)	⬆️

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[Auto-GPT-Bot]

You changed AutoGPT's behaviour. The cassettes have been updated and will be merged to the submodule when this Pull Request gets merged.

[collijk]

Very nice work on figuring out the function token counting. All comments here optional or for discussion.

[Auto-GPT-Bot]

You changed AutoGPT's behaviour. The cassettes have been updated and will be merged to the submodule when this Pull Request gets merged.

[Auto-GPT-Bot]

You changed AutoGPT's behaviour. The cassettes have been updated and will be merged to the submodule when this Pull Request gets merged.