Trey T
3ee1563cb0
The book reader's word lookup used DictionaryService, a verb-conjugation index plus ~200 hand-typed words: ordinary nouns like "taza" returned nothing, and homographs always lost (tapping "como" in "como siempre" gave the verb "comer" because the verb index is checked first). Add a glossary phase to the books pipeline (build_glossary.py): every distinct Spanish word is translated once, in its sentence context, by the same Claude-Code-subagent LLM step the pipeline already uses for chapter translation. English front matter is excluded by an ES==EN paragraph-ratio heuristic. The glossary is bundled into book_<slug>.json and is now part of the pipeline for every book. In the app, Book carries the decoded glossary and BookReaderView resolves each tap automatically through cache -> glossary -> DictionaryService -> on-device LLM, citing which source answered so a curated glossary hit reads differently from a best-effort AI guess. book_olly-vol2.json regenerated with a 3,658-word glossary. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
103 lines
6.5 KiB
Markdown
103 lines
6.5 KiB
Markdown
# Books pipeline
|
||
|
||
Turns any EPUB into a chapter-structured JSON file the app bundles and reads.
|
||
|
||
## TL;DR
|
||
|
||
```bash
|
||
cd Conjuga/Scripts/books
|
||
./run.sh /path/to/book.epub --slug my-book-slug
|
||
```
|
||
|
||
This runs Phase 1 (extract) and Phase 2 (manifest jobs), then stops and tells you how many translation jobs are pending. Run those via Claude Code subagents (Phase 2.5 below), then re-run `./run.sh` to bundle the final file.
|
||
|
||
## Phases
|
||
|
||
| Phase | Script | What it does | Output |
|
||
|---|---|---|---|
|
||
| 1 | `extract_epub.py` | Unzip the EPUB, walk `content.opf` spine + `toc.ncx` navMap, group HTML files into chapters, strip HTML→text. | `build/<slug>/chapters.json` |
|
||
| 2 | `translate_chapters.py` | Split each chapter into ~30-paragraph translation batches. Each batch becomes a job with its own input/output file. **Resumable**: jobs whose output file already exists are skipped. | `build/<slug>/jobs/<jobid>.input.json` + `_pending.txt` |
|
||
| 2b | `build_glossary.py` | Tokenize every Spanish paragraph the same way the app does, collect the distinct words with example sentences, split into ~150-word glossary batches. **Resumable** the same way. | `build/<slug>/glossary/<jobid>.input.json` + `_pending.txt` |
|
||
| 2.5 | Claude Code subagents | Drain **both** manifests: translate the chapter jobs *and* the glossary jobs, writing each job's `<jobid>.output.json`. See "Running translations" below. | `build/<slug>/{jobs,glossary}/<jobid>.output.json` |
|
||
| 3 | `bundle_book.py` | Merge `chapters.json` + every translation `*.output.json` + every glossary `*.output.json` into the final bundled JSON the app reads. | `Conjuga/Conjuga/book_<slug>.json` |
|
||
|
||
`run.sh` chains 1 → 2 → 2b → 3. If Phase 2 or 2b produces pending jobs, Phase 3 still runs but bundles with placeholders so you can preview app structure before the LLM passes complete. Re-running `run.sh` after subagents fill in the outputs gives you the real bundled file.
|
||
|
||
The glossary is the book reader's primary word-lookup source: every distinct word translated once, in context, so taps are instant, cover the whole book, and don't mis-resolve homographs (e.g. "como" as the conjunction vs. the verb *comer*). This phase is a permanent part of the pipeline — every book imported this way gets a glossary.
|
||
|
||
## Adding a new book
|
||
|
||
1. **Drop the EPUB** anywhere on disk.
|
||
2. **Run Phase 1+2**:
|
||
```bash
|
||
cd Conjuga/Scripts/books
|
||
./run.sh /path/to/book.epub --slug my-book
|
||
```
|
||
Sanity-check the chapter list it prints. If chapter grouping looks wrong (e.g. an EPUB without a usable `toc.ncx`), `extract_epub.py` will need a fallback heuristic — see "Open assumptions" below.
|
||
|
||
3. **Run translations** (Phase 2.5). The default approach is to spawn Claude Code subagents from inside a Claude Code session pointed at this repo:
|
||
|
||
There are **two** manifests to drain — translation and glossary:
|
||
- `build/<slug>/jobs/_pending.txt` with prompt `build/<slug>/jobs/_prompt_template.md`
|
||
- `build/<slug>/glossary/_pending.txt` with prompt `build/<slug>/glossary/_prompt_template.md`
|
||
|
||
For each pending job ID, hand a subagent the matching prompt with `<JOB_INPUT_PATH>` / `<JOB_OUTPUT_PATH>` filled in. The subagent reads the input, produces the translation/glossary, and writes the output. Resumable — interrupted runs just leave the missing job IDs in `_pending.txt`.
|
||
|
||
Cluster jobs into agent batches of ~5–10 jobs each to keep per-agent context manageable. ~5 parallel agents is a good throughput target.
|
||
|
||
4. **Bundle**:
|
||
```bash
|
||
./run.sh /path/to/book.epub --slug my-book # re-running pulls in the new outputs
|
||
# or directly:
|
||
python3 bundle_book.py my-book --require-all
|
||
```
|
||
`--require-all` will fail loudly if any job is still missing.
|
||
|
||
5. **Bump `bookDataVersion`** in `DataLoader.swift` so the in-app store re-seeds the new book on next launch (or any time you re-run with new translations).
|
||
|
||
6. **Verify the file is bundled** in `Conjuga.xcodeproj`. The script writes `book_<slug>.json` into `Conjuga/Conjuga/Resources/`; if that folder is part of a recursive group reference, Xcode picks it up automatically. Otherwise, add it manually or via the `xcodeproj` ruby gem.
|
||
|
||
## File layout
|
||
|
||
```
|
||
Conjuga/Scripts/books/
|
||
├── extract_epub.py # Phase 1
|
||
├── translate_chapters.py # Phase 2
|
||
├── build_glossary.py # Phase 2b
|
||
├── bundle_book.py # Phase 3
|
||
├── run.sh # Orchestrator
|
||
└── build/ # gitignored
|
||
└── <slug>/
|
||
├── chapters.json
|
||
├── jobs/ # translation jobs
|
||
│ ├── _pending.txt
|
||
│ ├── _prompt_template.md
|
||
│ ├── ch01_b00.input.json
|
||
│ ├── ch01_b00.output.json
|
||
│ └── ...
|
||
└── glossary/ # glossary jobs (Phase 2b)
|
||
├── _pending.txt
|
||
├── _prompt_template.md
|
||
├── gloss_b00.input.json
|
||
├── gloss_b00.output.json
|
||
└── ...
|
||
```
|
||
|
||
The final output (`book_<slug>.json`) lives at `Conjuga/Conjuga/book_<slug>.json` so the iOS app bundle includes it. (Existing `textbook_data.json` / `conjuga_data.json` use the same layout — files in the app target root rather than a Resources subgroup.)
|
||
|
||
## Open assumptions
|
||
|
||
- **TOC drives chapter boundaries.** If an EPUB ships without a usable `toc.ncx`, or the navMap is too granular (e.g. one navPoint per page), `extract_epub.py` will need a fallback that groups by `<h1>` headings in spine order.
|
||
- **Spanish bold tags = inline emphasis.** The Olly Richards books bold vocab hints inside paragraphs. We strip the bold and let the in-app dictionary lookup handle definitions instead. If a future book uses bold for something else (titles, etc.), revisit.
|
||
- **Translation is per-paragraph 1:1.** Subagents must preserve paragraph count and order. `bundle_book.py` will warn + pad/truncate if a job's output array length doesn't match its input — but that's a sign the subagent misbehaved.
|
||
|
||
## Out of scope (intentional)
|
||
|
||
- OCR of vocab image tables (use `Scripts/textbook/` if your book is image-heavy).
|
||
- Exercise extraction (textbook pipeline).
|
||
- Per-occurrence word sense disambiguation. The glossary has one entry per
|
||
distinct word, translated in context; a word genuinely used in two senses in
|
||
the same book gets its dominant sense. The runtime `DictionaryService` + the
|
||
on-device LLM remain as fallbacks for anything the glossary misses.
|
||
- Cover image extraction (covers are derived from a color hash in the app for now).
|