atliliw/langchainRustAgent

A Rust-based retrieval-augmented generation (RAG) conversational system supporting multiple retrieval modes and dialogue compression strategies. Intended for developers and teams building RAG chat apps or experimenting with retrieval + LLM workflows. Key features: - Three retrieval modes: vector (Qdrant), BM25 (MongoDB), and RRF hybrid; endpoint to compare all three. - Dialogue compression: hierarchical and summary compression with protection for important info. - Multi-format document ingestion: TXT, PDF, Markdown, JSON, CSV; file upload via single-page web UI. - Backend: Rust + Axum; vector DB: Qdrant; BM25: MongoDB; conversation store: SQLite. - Uses OpenAI embeddings (text-embedding-ada-002) and GPT-3.5/4 for LLMs. - APIs for chat (/api/chat), upload, search variants, stats; config via config.toml; run with cargo. MIT license.

このRepoのdescription / GitHub topics / READMEのAI要約に出現したAI関連キーワードをカテゴリ別に表示。各バッジは該当カテゴリの詳細ランキングへリンクします。