funkpower/catena
1
0
Fork 0
mirror of https://git.pumpin.online/funkpower/catena.git synced 2026-10-23 03:07:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

vocabmerge: a subject scheme, supplied rather than compiled in

Browse Source
This commit is contained in:
funkpower
2026-09-19 19:06:57 -04:00
parent 05b363ea8a
commit 484017fc18
9 changed files with 801 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/bibclean/README.md
+52 -1
View File
@@ -251,7 +251,7 @@ One heading per line — an `id`, a `name`, and the `count` of works under it:
| `--rules <FILE>` | Vocabulary specific to this collection |
| `--man` | Write this page as a manual page, in roff |
The rules file is JSON, and today holds one thing:
The rules file is JSON and holds two things.
```json
{"device_prefixes": ["alexskobo"]}
@@ -261,6 +261,57 @@ An e-reader stamps its own name onto every book it touches. The generic device
names are built in; one named after its owner is a fact about a single
collection, so it is supplied rather than compiled in.
### A subject scheme
The other thing is a closed list of subjects to file under. A catalogue that
lets anything be a subject ends up with forty thousand of them and no way to
browse; a scheme is the other approach.
```json
{
"scheme": {
"subjects": ["Philosophy", "Phenomenology", "Theology", "History", "Fiction"],
"aliases": {"Philosophie": "Philosophy", "Church History": "History"},
"drop": ["General", "Humanities", "Non-Fiction"],
"parents": {"Phenomenology": "Philosophy"},
"blocks_fiction": ["Biography"],
"fiction_genres": ["Fiction"]
}
}
```
Which subjects a collection uses is the most collection-specific thing there
is — a theology library and a science library agree on none of them — so it
arrives in the file rather than compiled in.
With one, every heading nothing else has already claimed gets an answer: a
spelling the scheme knows is a merge onto the spelling it uses, and a heading
the scheme does not have at all is a delete.
```
$ vocabmerge --kind tag --rules scheme.json tags.ndjson
{"kind":"tag","action":"merge","keep":"Philosophy","drop":"Philosophie",…,"method":"scheme_alias","confidence":0.95,"evidence":"the scheme files \"Philosophie\" under \"Philosophy\""}
{"kind":"tag","action":"delete","keep":null,"drop":"General",…,"method":"scheme_absent","confidence":0.8,"evidence":"\"General\" is not in the scheme, on 88 books"}
{"kind":"tag","action":"delete","keep":null,"drop":"Underwater Basketry",…,"method":"scheme_absent","confidence":0.8,"evidence":"\"Underwater Basketry\" is not in the scheme, on 1 books"}
{"kind":"tag","action":"merge","keep":"History","drop":"Church History",…,"method":"scheme_alias","confidence":0.95,"evidence":"the scheme files \"Church History\" under \"History\""}
```
The deletes are the lower confidence of the two on purpose: an alias is a
statement the scheme makes, while an absence only says the scheme does not
list it — which may equally mean the scheme is short a heading.
Four rules the scheme applies, for anything using it as a library:
- **Parents come along.** Phenomenology files under Philosophy and
Phenomenology, so a shelf browsed by the parent still finds the child.
- **Order is scheme order**, not the order the headings arrived in.
- **A study is not a novel.** A work already under a heading in
`blocks_fiction` — or whose title says "Companion to", "A Life of", "Diary",
"Very Short Introduction" — loses the `fiction_genres`. A historical novel
may still be both.
- **A book tagged only with headings the scheme rejects has no subject yet**,
which is what a catalogue built from filenames looks like.
### Exit codes
| Code | Meaning |