Elamur: **Mova** · [Parse](https://mova.elamur.ai/free/en/parse/) · [Dictionary](https://md.dictionary.elamur.ai/free/en/) · [Translate](https://md.translate.elamur.ai/free/en/) · [Research](https://md.research.elamur.ai/free/en/) · [MCP](https://mcp.elamur.ai/) · [Groupants](https://md.groupants.elamur.ai/free/en/)

[Mova](https://md.mova.elamur.ai/free/en/) · [this page in HTML](https://mova.elamur.ai/free/en/docs/tools/school-sents/)

[home](https://md.mova.elamur.ai/free/en/) / [docs](https://md.mova.elamur.ai/free/en/docs/) / [tools](https://md.mova.elamur.ai/free/en/docs/tools/) / school-sents — sentence pools for the annotator and the cycle

# school-sents — sentence pools for the annotator and the cycle

`school-sents` cuts texts into sentences and writes them as a pool (`id<TAB>sentence`, or for lessons a four-column TSV) for the SLM, the annotator and the [teacher cycle](https://md.mova.elamur.ai/free/en/docs/tools/cycle/).

## Build

```
cargo build --release -p school-sents
```

## Forms

| Command | Input → output |
| --- | --- |
| `school-sents <school-dir> <out.tsv>` | every `<course>/*.lesson`: the `explain:` and `fact:` values, cut into sentences; fragments under three words dropped. Columns: course, lesson file, index in the lesson, sentence |
| `school-sents corpus <corpus-dir> <pool.tsv>` | the open corpus (release 0.4 layout, `corpus/open`): books (`book:<file>:<n>`), the pupils' dictionary (`dict:<word>:<n>`) and the train-split questions of the maths scripts (`math:<id>:<n>`); the contrast sets are tests and are not read |
| `school-sents books <books-dir> <pool.tsv> [--skip <file.txt>]… [--exclude <test.conllu>]…` | books only; `--exclude` drops every sentence that is a sentence of a held-out test set |
| `school-sents usfm <dir> <pool.tsv> <prefix> [BOOK,…\|-] [--hebrew-verses]` | scripture in USFM (`*.usfm` or `*.usfm.zst`), one sentence per id `<prefix>:<BOOK>:<n>`; `<pool.tsv>.verses` maps ids to verses; `--hebrew-verses` maps each sentence to the Hebrew verse numbering |
| `school-sents notes <dir> <pool.tsv> <prefix>` | Markdown notes (`*.md` except README.md) with markup taken off; `<prefix>:<file stem>:<n>` |
| `school-sents addressed <pool.tsv> <prefix>=<in.tsv>…` | texts by address (`address<TAB>text`); `<prefix>:<n>`; `<pool.tsv>.verses` maps ids to addresses |
| `school-sents scans <dir> <pool.tsv> <prefix>` | English prose of scanned books (`*.txt`): running heads, page numbers and line-end hyphens repaired; OCR noise (under 70% known English forms) left out; `<pool.tsv>.counts` per book |

Book texts lose the Project Gutenberg header and footer; a «sentence» over 200 words (a table of contents, a run-together list) is left out. Sentences with Hebrew or Aramaic script are counted in `notes` and `scans`.

## Limits

The cut is rule-based (`.`, `!`, `?` before a space and a capital or an opening quote, with extra title and numbering rules for prose). Lessons keep the older cut on purpose: the parse database's sentence ids count them.

- [← previous roundtrip — where each layer of Mova loses information](https://md.mova.elamur.ai/free/en/docs/tools/roundtrip/)
- [next → Sea Spider — mova-seaspider](https://md.mova.elamur.ai/free/en/docs/tools/seaspider/)
- [↑ back to tools](https://md.mova.elamur.ai/free/en/docs/tools/)