Conversation
This change introduces support for OpenRouter as a new LLM provider and refactors the model provider architecture to be more extensible. Key changes include: - Added an `OpenRouterProvider` to `llm_provider.py`. - Refactored `call_llm.py` to use a factory pattern for instantiating providers, removing the previous `if/elif/else` structure. - Modified `pyproject.toml` to use optional dependencies (`extras`) for each provider (`google`, `ollama`, `huggingface`, `openrouter`), allowing users to install only the libraries they need. - Updated `conf.py` to prefix `ModelName` enum members with their provider (e.g., `google_gemini_2_5_flash`) for more robust provider detection. - Updated `README.md` files for both the core and CLI packages to reflect the new installation instructions and model naming conventions. - Fixed all tests to align with the new architecture.
Summary of ChangesHello @leo-gan, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request significantly enhances the flexibility and extensibility of the anonymization tool by integrating OpenRouter as a new LLM provider. It introduces a more modular architecture for interacting with various LLMs, making it easier to add future providers and manage dependencies. The changes also streamline the installation process and update documentation to reflect the new provider and improved structure. Highlights
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces support for the OpenRouter provider and significantly refactors the LLM provider handling. The new architecture, which uses a provider abstraction layer and a factory pattern, is a major improvement that enhances modularity and makes the system more extensible. The dependency management is also cleaner with the use of optional dependencies for each provider. I've identified a couple of areas for improvement, mainly concerning robustness and maintainability in the new provider implementation.
| completion = client.chat.completions.create( | ||
| model=model_name, messages=[{"role": "user", "content": prompt}] | ||
| ) | ||
| return completion.choices[0].message.content or "" |
There was a problem hiding this comment.
There's a potential IndexError here if the completion.choices list returned by the API is empty. While this is unlikely for a successful call, it's safer to guard against it to prevent the application from crashing. This change ensures the code gracefully handles an empty choices list.
| return completion.choices[0].message.content or "" | |
| return completion.choices[0].message.content or "" if completion.choices else "" |
| logging.error( | ||
| f"Attempt {attempt + 1} failed with JSON decode error: {e}, " | ||
| f"response: {response_text[:200]}..." | ||
| f"response: {raw_text[:200] if 'raw_text' in locals() else 'N/A'}..." | ||
| ) |
There was a problem hiding this comment.
The if 'raw_text' in locals() check is redundant. The json.JSONDecodeError exception can only be raised if json.loads() is called, which means raw_text has already been successfully assigned in the try block. You can simplify this logging statement for better readability.
| logging.error( | |
| f"Attempt {attempt + 1} failed with JSON decode error: {e}, " | |
| f"response: {response_text[:200]}..." | |
| f"response: {raw_text[:200] if 'raw_text' in locals() else 'N/A'}..." | |
| ) | |
| logging.error( | |
| f"Attempt {attempt + 1} failed with JSON decode error: {e}, " | |
| f"response: {raw_text[:200]}..." | |
| ) |
| def get_provider(provider_name: str) -> LLMProvider: | ||
| """Factory function to get a provider instance.""" | ||
| if provider_name == "google": | ||
| return GoogleProvider() | ||
| if provider_name == "ollama": | ||
| return OllamaProvider() | ||
| if provider_name == "huggingface": | ||
| return HuggingFaceProvider() | ||
| if provider_name == "openrouter": | ||
| return OpenRouterProvider() | ||
| raise ValueError(f"Unknown provider: {provider_name}") |
There was a problem hiding this comment.
The get_provider factory function can be made more scalable and arguably more readable by using a dictionary to map provider names to their corresponding classes. This avoids a growing chain of if statements as more providers are added and is a more common pattern for factories.
| def get_provider(provider_name: str) -> LLMProvider: | |
| """Factory function to get a provider instance.""" | |
| if provider_name == "google": | |
| return GoogleProvider() | |
| if provider_name == "ollama": | |
| return OllamaProvider() | |
| if provider_name == "huggingface": | |
| return HuggingFaceProvider() | |
| if provider_name == "openrouter": | |
| return OpenRouterProvider() | |
| raise ValueError(f"Unknown provider: {provider_name}") | |
| def get_provider(provider_name: str) -> LLMProvider: | |
| """Factory function to get a provider instance.""" | |
| provider_map = { | |
| "google": GoogleProvider, | |
| "ollama": OllamaProvider, | |
| "huggingface": HuggingFaceProvider, | |
| "openrouter": OpenRouterProvider, | |
| } | |
| provider_class = provider_map.get(provider_name) | |
| if not provider_class: | |
| raise ValueError(f"Unknown provider: {provider_name}") | |
| return provider_class() |
1 participant
No description provided.