Modernize packaging and add unified inference module#39
Merged
Conversation
- Replace setup.py + requirements files with pyproject.toml (PEP 517/518) - Fix eval() usage with ast.literal_eval() for security - Fix bare retry() decorator calls in gpt.py and claude.py - Fix relative imports in inference/main.py - Remove fire dependency from evaluate.py, replace with argparse - Remove unused imports (sys, Union) - Add deprecation warnings to legacy model modules - Add new BaseModel ABC and OpenAIModel for unified inference - Add openai-compatible model option to inference CLI - Add initial test suite for evaluation utilities - Run ruff check/format and ty check across codebase
The newer OpenAI API requires max_completion_tokens for chat completions. The old max_tokens parameter is rejected by recent models.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
setup.py+eval_requirements.txt/infer_requirements.txtwith a modernpyproject.toml(PEP 517/518, setuptools backend)eval()withast.literal_eval()inparse_raw_output()retry()decorator calls ingpt.pyandclaude.py(were no-ops, now properly applied)inference/main.pyfiredependency fromevaluate.py, replace withargparsesys,Union)claude.py,gpt.py,llama.py,rwkv.py)BaseModelABC andOpenAIModelfor unified inference via any OpenAI-compatible endpointopenai-compatiblemodel option with--base-url,--api-mode, and--model-idCLI argumentstests/test_evaluation.py) coveringedit_distanceandparse_raw_outputruff check --fix,ruff format, andty checkacross the codebaseTest plan
pytest tests/ -v)ruff checkpasses with zero errorsruff formatproduces no changesty checkpasses with zero errorsmango-evalCLI still works with existing inference resultsopenai-compatiblemodel option works end-to-end against an OpenAI-compatible endpoint