Refactor/interfaces

[finitearth]

No description provided.

[Copilot]

Pull Request Overview

This PR refactors the interface for predictors, optimizers, and LLMs by introducing a unified ExperimentConfig, removing legacy *args/**kwargs parameters and updating factory functions to pass configuration. Key changes include updating constructor signatures across modules, refactoring configuration management in config.py, and revising factory functions in predictors, optimizers, and llms.

Reviewed Changes

Copilot reviewed 50 out of 50 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
promptolution/predictors/classificator.py	Updated init signatures to accept ExperimentConfig; adjusted super() calls.
promptolution/predictors/base_predictor.py	Added config parameter and updated abstract method signatures.
promptolution/predictors/init.py	Revised get_predictor signature and default value for predictor type.
promptolution/optimizers/*.py and base_optimizer.py	Refactored constructors to enable unified configuration and removed legacy parameters.
promptolution/llms/*.py	Updated LLM constructors and factory function get_llm to incorporate ExperimentConfig.
promptolution/config.py	Refactored configuration class to simplify and centralize config management.
promptolution/helpers.py	Adjusted experiment functions to use the updated ExperimentConfig.
Other files (exemplars, callbacks, init.py, notebooks)	Minor updates to support the refactored interfaces and configuration.

[Copilot]

Pull Request Overview

This PR refactors the LLM interfaces across the package by standardizing how configurations are passed, replacing generic **kwargs with explicit parameters, and updating factory functions and documentation. Key changes include:

• Using the unified ExperimentConfig in LLM classes and helper functions.
• Updating the parameter signatures and docstrings for VLLM, LocalLLM, APILLM, and BaseLLM.
• Refining auxiliary modules, configuration handling, and documentation (notebooks, docs, workflows) to reflect these changes.

Reviewed Changes

Copilot reviewed 59 out of 59 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
promptolution/llms/vllm.py	Refactored constructor signature to accept llm_kwargs and ExperimentConfig; removed del method.
promptolution/llms/local_llm.py	Updated constructor to include config parameter and call the new BaseLLM initializer.
promptolution/llms/base_llm.py	Updated initialization to rely on ExperimentConfig instead of variadic parameters.
promptolution/llms/api_llm.py	Switched from model_id to llm in the constructor and updated _invoke_model call to include retry logic.
promptolution/llms/init.py	Modified get_llm factory function to work with ExperimentConfig and harmonize LLM instantiation.
promptolution/helpers.py	Adjusted experiment and evaluation workflows to use the new ExperimentConfig and revised helper APIs.
Other files (config, callbacks, docs, workflows)	Miscellaneous updates to documentation, configuration, and CI/CD workflows reflecting broader changes.

Comments suppressed due to low confidence (1)

promptolution/llms/api_llm.py:108

• [nitpick] The APILLM attribute 'llm' is used as the model identifier in the API call, which may be ambiguous compared to other LLM classes that use 'model_id'. Consider renaming it to 'model_id' for clarity and consistency.

            _invoke_model(prompt, system_prompt, self.max_tokens, self.llm, self.client, self.semaphore)