Add the attachments feature to LLMSession

[tattn]

• Add attachments feature to LLMSession
• Replace LLMClient with LLMSession in the example app
• Improve the stability of the model downloading

[gemini-code-assist]

Summary of Changes

Hello @tattn, 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 LocalLLMClient by integrating multi-modal attachment support directly into the LLMSession class. This refactoring centralizes message history and interaction logic within LLMSession, streamlining the API for sending prompts with attachments and improving the example application's architecture. Additionally, it includes improvements to file downloading robustness and verbosity control for Llama models.

Highlights

• Core LLMSession Enhancements: The LLMSession is now the central component for managing LLM interactions, including maintaining message history and handling multi-modal inputs (attachments). It has been made @Observable for seamless integration with SwiftUI, and its internal Generator class now uses Locked wrappers for thread-safe access to client and messages.
• Multi-modal Attachment Support: A new LLMAttachment struct has been introduced, allowing for generic attachment types (currently images) to be passed alongside text prompts to the LLM. The LLMInput.Message struct has also been updated to conform to Identifiable and include an id property for better SwiftUI integration.
• Example Application Refactoring: The LocalLLMClientExample app's UI components (BottomBar, ChatView) and ViewModel (ChatViewModel) have been updated to leverage the new LLMSession and LLMAttachment capabilities, simplifying message flow, enabling image input, and centralizing message history management within the LLMSession.
• Robust File Downloading: The underlying file downloading mechanism has been improved. The FileDownloader now verifies downloaded files by size in addition to existence, and the Downloader cleans up partially downloaded files upon error, enhancing reliability. The HuggingFaceAPI has been updated to retrieve file sizes for this purpose.
• Centralized Verbosity Control: Verbosity settings for Llama models are now managed through LlamaClient.Parameter.Options, providing a more consistent and centralized approach to debugging output across the Llama client and its multimodal context.

Using Gemini Code Assist

The 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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in issue comments and review comments.
Help	/gemini help	Displays a list of available commands.

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 .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist is currently in preview and 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 to provide feedback.

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

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

The pull request introduces the attachments feature to LLMSession, enhancing the example app and the underlying library. Several improvements are suggested, including safer unwrapping of optionals, allowing messages with only attachments, improving the efficiency of file download checks, and using more type-safe access to attachment content.