Skip to content

GEMINI.md

Latest commit

 

History

History
77 lines (52 loc) · 2.73 KB

GEMINI.md

File metadata and controls

77 lines (52 loc) · 2.73 KB
Raw

GEMINI.md

Project Overview

This project is a Rust-based command-line interface (CLI) called "Fluent CLI". It provides a unified interface for interacting with multiple Large Language Model (LLM) providers, including OpenAI, Anthropic, and Google.

The project is structured as a Rust workspace with multiple crates, each responsible for a specific functionality:

  • fluent-cli: The main CLI application.
  • fluent-core: Core utilities and configuration.
  • fluent-engines: LLM engine implementations.
  • fluent-agent: Agentic capabilities and tools.
  • fluent-storage: Storage and persistence layer.
  • fluent-sdk: SDK for external integrations.

Fluent CLI has advanced features such as:

  • Agentic Capabilities: An experimental agentic system with a ReAct loop, a tool system for file operations, shell commands, and code analysis.
  • Pipeline Execution: A YAML-based pipeline system for defining and executing multi-step workflows.
  • Model Context Protocol (MCP): Integration with MCP for tool integration.
  • Configuration Management: YAML-based configuration for multiple LLM engines.
  • Self-Reflection and Learning: An experimental system for self-reflection and learning.

Building and Running

Building the Project

To build the project, run the following command:

cargo build --release

Running the CLI

The main executable is fluent. You can use it with various subcommands:

  • Direct LLM Queries:

    fluent openai-gpt4 "Explain quantum computing"

  • Agent Commands:

    fluent agent

  • Pipeline Commands:

    fluent pipeline -f pipeline.yaml -i "process this data"

  • MCP (Model Context Protocol) Commands:

    fluent mcp server --stdio

Running Tests

To run the test suite, use the following command:

cargo test

Development Conventions

  • Modular Architecture: The project follows a modular architecture with functionalities separated into different crates.
  • Error Handling: The project has a comprehensive error handling system using the anyhow and thiserror crates.
  • Logging: The project uses the tracing and env_logger crates for logging.
  • Configuration: The application is configured using YAML files (config.yaml, fluent_config.toml).
  • Code Quality: The project has scripts for security audits (scripts/security_audit.sh) and code quality checks (scripts/code_quality_check.sh).
  • Dependencies: The project uses a wide range of dependencies, including tokio for asynchronous programming, clap for command-line argument parsing, serde for serialization/deserialization, and reqwest for making HTTP requests.