No description
  • HTML 66.4%
  • CSS 13.2%
  • Python 12.3%
  • Go 8%
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2026-09-13 08:41:31 -07:00
.claude Check in the book-deep-dive agent and document Claude Code config 2026-09-08 16:39:54 -07:00
fiction Add a generated Files section to book indexes 2026-09-08 16:30:26 -07:00
non-fiction Organize and index new books 2026-09-08 17:19:59 -07:00
scripts Organize and index new books 2026-09-08 17:19:59 -07:00
tools fix(booksearch): mark MCP tools read-only 2026-09-13 08:41:31 -07:00
.gitattributes Add PDF support to Git LFS tracking 2026-03-23 06:34:36 -07:00
.gitignore Add unified task-based repository workflows 2026-09-08 16:27:40 -07:00
.mcp.json Refactor book indexing and search system to utilize snapshot-based database connections, enhancing data consistency during updates. Replace old node handling with atomic book replacement, and improve embedding and summarization processes for better performance. 2026-03-22 16:59:56 -07:00
CONTRIBUTING.md Check in the book-deep-dive agent and document Claude Code config 2026-09-08 16:39:54 -07:00
cspell.json Add book indexing and search functionality with model integration 2026-03-22 15:09:36 -07:00
pyproject.toml Organize and index new books 2026-09-08 17:19:59 -07:00
README.md Check in the book-deep-dive agent and document Claude Code config 2026-09-08 16:39:54 -07:00
Taskfile.yml Add unified task-based repository workflows 2026-09-08 16:27:40 -07:00
uv.lock Organize and index new books 2026-09-08 17:19:59 -07:00

Books

A browsable book collection with source files, Markdown conversions, metadata, and a local semantic-search service. The collection is organized by subject; the Go tooling under tools/bookindex builds a RAPTOR-style vector index and exposes it through MCP.

Browse the collection

Repository layout

Path Purpose
fiction/ Fiction titles and their category index
non-fiction/ Technical and non-technical subject trees
new_books/ Staging area for books that have not been classified
scripts/ Conversion, metadata, organization, and index-maintenance utilities
tools/bookindex/ Go indexer, distributed worker, and MCP search server
.claude/ Checked-in Claude Code agents and slash commands

Each searchable book directory contains:

  • an EPUB or MOBI source file;
  • one converted Markdown content file; and
  • an index.md with YAML frontmatter for title, year, keywords, and summary.

PDF-only folders can be browsed in the repository, but the current semantic indexer discovers only directories containing EPUB or MOBI files.

Common workflows

To add or reorganize books, start with the contribution guide. The scripts can move, rename, or delete files, so review their documented effects before running the batch workflow.

The root Taskfile is the main entry point. Run task --list to see repository tasks, task setup to create the uv environment, and task validate for the standard validation suite.

To build or serve semantic search, see the Bookindex guide. Its shortest local validation loop is:

cd tools/bookindex
go test ./...
go vet ./...
task build

Category index.md files and the section list above are generated by scripts/build_index.py. The script deliberately leaves the rest of this README intact.

Agent-assisted workflows

The repository ships its own Claude Code agents and slash commands: /organize for batch ingestion, /reconvert for repairing books with missing Markdown, and /deep-dive for cross-source research through the booksearch MCP server. See the Claude Code configuration guide for each agent's role and the side effects to expect.

Content and large files

Book archives are stored with Git LFS according to .gitattributes. Install Git LFS before cloning or adding EPUB, MOBI, PDF, ZIP, or BZ2 files. Only add material that you have the right to store and use; no repository-wide content license is declared here.