Book recommender: library integration
Design for integrating the book recommender with the garden library, making it the single source of truth for all book data.
Design for integrating the book-recommender with the garden’s library collection.
Goal
Replace the standalone library-audit.json with the garden’s existing library collection. The library page becomes the single interface for managing books and triggering recommendations.
Approach
1. Extend library frontmatter
Add audit fields to each book’s markdown frontmatter:
# Existing fields
title: 1Q84
author: Haruki Murakami
status: to-read
# New audit fields
type: fiction
purpose: personal
reason: "Not on my radar; I really liked previous Murakami books"
notes: null
rating: null
review: null2. Recommender reads from library
recommend.py reads markdown files from src/content/library/ instead of library-audit.json. Parses frontmatter using Python’s yaml library.
3. Write back recommendations
After scoring, the script updates top picks with a recommended field in their frontmatter:
recommended: true
recommended_mood: comfort-zone
recommended_score: 57.0This lets the library page highlight current picks visually.
4. Library page UI
Two upgrades to the existing garden library page:
- Inline audit editing: each book entry has an expandable form for type, purpose, reason, notes, rating
- A “get recommendations” trigger that runs the recommender and opens the dashboard
The display layer (covers, collection view, status filters) stays unchanged.