libra-ebook 0.1.0

AI-powered personal ebook library manager

Libra

AI-powered personal ebook library manager. Scan your collection, auto-tag with AI, search by meaning, clean up messy filenames.

Features

• AI tagging — auto-generate tags, clean titles, detect authors from messy filenames
• Semantic search — find files by concept, not just keywords ("那篇讲垃圾回收的论文")
• Three-layer search — FTS5 keyword + vector similarity + path matching
• Multi-version grouping — automatically groups editions and formats of the same work
• Rename — batch rename files based on canonical metadata, cross-platform safe
• fzf integration — interactive browse with preview, open, edit, copy path
• Controlled vocabulary — tags from a predefined taxonomy, extensible

Install

# From PyPI
pip install libra-ebook

# From source
git clone https://github.com/balusch/libra && cd libra
uv sync

Quick Start

# Set your OpenAI API key
export OPENAI_API_KEY="sk-..."

# Scan your library
libra scan ~/Documents/EBooks

# Search
libra search "distributed systems"
libra search "垃圾回收" --fzf

# Browse interactively
libra browse --tag ai

# See what you have
libra stats
libra tags
libra dupes

Commands

Command	Description
libra scan [PATH]	Scan directory, AI tag & embed (--limit N to cap)
libra search QUERY	Three-layer search (FTS + semantic + path)
libra browse	Interactive fzf browser
libra list	List files with filters
libra show ID|PATH	Show file details (by ID or path)
libra open ID	Open file with system app
libra edit ID	Edit tags/metadata
libra related ID	Find similar files
libra prefer ID	Mark preferred version
libra retag	Re-run AI tagging on selected files
libra summarize	Generate AI summaries from file content
libra rename [PATH]	Rename files from metadata (--id ID for single file)
libra tags	List tags with counts
libra dupes	Find duplicates + versions
libra stats	Library statistics
libra export	Export index (JSON/CSV)

Configuration

# ~/.config/libra/config.toml

[api]
openai_key = "sk-..."          # or use OPENAI_API_KEY env var

[library]
paths = ["~/Documents/EBooks"]

[search]
weight_fts = 0.4                 # keyword search weight
weight_semantic = 0.5            # semantic search weight
weight_path = 0.1                # path match weight

The database is stored at {library_path}/.libra/libra.db.

Shell Completion

# bash
libra --install-completion bash

# zsh
libra --install-completion zsh

# fish
libra --install-completion fish

Development

uv sync --all-groups   # install with dev deps
make check             # lint + typecheck + test
make format            # auto-format
make build             # build distribution

License

MIT