tane/apps/app_seeds/test/data/species_repository_test.dart
vjrj ea5ab0342d perf(species): seed the bundled catalog only once per version
Skip the ~1.5 MB catalog parse and the table scan on every launch — the
main startup cost. `seedBundledIfNeeded` re-seeds only when the recorded
version differs from `speciesCatalogVersion`, writing the version after
the seed commits so an interrupted seed retries next launch.

Adds `parseSpeciesCatalogVersion` and a test keeping the constant in sync
with the asset's `version` field.
2026-07-10 15:46:04 +02:00

286 lines
8.3 KiB
Dart

import 'package:commons_core/commons_core.dart';
import 'package:drift/drift.dart' show Value;
import 'package:flutter_test/flutter_test.dart';
import 'package:tane/data/species_catalog.dart';
import 'package:tane/data/species_repository.dart';
import 'package:tane/db/database.dart';
import '../support/test_support.dart';
const _seeds = [
SpeciesSeed(
scientificName: 'Solanum lycopersicum',
family: 'Solanaceae',
commonNames: {
'en': ['Tomato'],
'es': ['Tomate'],
},
),
SpeciesSeed(
scientificName: 'Phaseolus vulgaris',
family: 'Fabaceae',
commonNames: {
'en': ['Common bean'],
'es': ['Judía', 'Alubia'],
},
),
];
void main() {
late AppDatabase db;
late SpeciesRepository repo;
setUp(() {
db = newTestDatabase();
repo = SpeciesRepository(db, idGen: IdGen());
});
tearDown(() => db.close());
test(
'parseSpeciesCatalog reads scientific name, family and common names',
() {
const json = '''
{"version":3,"species":[
{"scientific_name":"Zea mays","family":"Poaceae","wikidata_qid":"Q11575","gbif_key":2703077,"common":{"en":["Maize"],"es":["Maíz"]}}
]}''';
final seeds = parseSpeciesCatalog(json);
expect(seeds.single.scientificName, 'Zea mays');
expect(seeds.single.family, 'Poaceae');
expect(seeds.single.wikidataQid, 'Q11575');
expect(seeds.single.gbifKey, 2703077);
expect(seeds.single.commonNames['es'], ['Maíz']);
},
);
test('parseSpeciesCatalog reads gbif and wikidata identifiers', () {
const json = '''
{"version":3,"species":[
{"scientific_name":"Zea mays","gbif_key":5290052,"wikidata_qid":"Q11575"},
{"scientific_name":"Unknown sp."}
]}''';
final seeds = parseSpeciesCatalog(json);
expect(seeds.first.gbifKey, 5290052);
expect(seeds.first.wikidataQid, 'Q11575');
// Missing identifiers stay null, they are optional.
expect(seeds.last.gbifKey, isNull);
expect(seeds.last.wikidataQid, isNull);
});
test('seedBundled persists gbif and wikidata identifiers', () async {
await repo.seedBundled(const [
SpeciesSeed(
scientificName: 'Zea mays',
gbifKey: 5290052,
wikidataQid: 'Q11575',
),
]);
final row = await db.select(db.species).getSingle();
expect(row.gbifKey, 5290052);
expect(row.wikidataQid, 'Q11575');
});
test(
'seedBundled backfills identifiers onto rows seeded before them',
() async {
// A pre-existing bundled row that predates the identifier fields.
await db
.into(db.species)
.insert(
SpeciesCompanion.insert(
id: 'legacy',
createdAt: 0,
updatedAt: '0',
lastAuthor: 'bundle',
scientificName: 'Zea mays',
isBundled: const Value(true),
),
);
await repo.seedBundled(const [
SpeciesSeed(
scientificName: 'Zea mays',
gbifKey: 5290052,
wikidataQid: 'Q11575',
),
]);
final rows = await db.select(db.species).get();
expect(rows.single.id, 'legacy'); // same row, not a duplicate
expect(rows.single.gbifKey, 5290052);
expect(rows.single.wikidataQid, 'Q11575');
},
);
test('seedBundled inserts species and is idempotent', () async {
await repo.seedBundled(_seeds);
await repo.seedBundled(_seeds); // second run must not duplicate
final species = await db.select(db.species).get();
expect(species.length, 2);
expect(species.every((s) => s.isBundled), isTrue);
});
test('search matches by scientific name', () async {
await repo.seedBundled(_seeds);
final results = await repo.search('phaseolus');
expect(results.single.scientificName, 'Phaseolus vulgaris');
});
test('search matches by common name and returns a localized label', () async {
await repo.seedBundled(_seeds);
final es = await repo.search('alubia', languageCode: 'es');
expect(es.single.scientificName, 'Phaseolus vulgaris');
expect(es.single.commonName, anyOf('Judía', 'Alubia'));
final en = await repo.search('tom', languageCode: 'en');
expect(en.single.commonName, 'Tomato');
expect(en.single.displayLabel, 'Tomato (Solanum lycopersicum)');
});
test('empty query returns nothing', () async {
await repo.seedBundled(_seeds);
expect(await repo.search(' '), isEmpty);
});
test('seedBundled stores wikidata id and gbif key', () async {
await repo.seedBundled(const [
SpeciesSeed(
scientificName: 'Solanum lycopersicum',
family: 'Solanaceae',
wikidataQid: 'Q13223',
gbifKey: 2930137,
commonNames: {
'en': ['Tomato'],
},
),
]);
final row = await db.select(db.species).getSingle();
expect(row.wikidataQid, 'Q13223');
expect(row.gbifKey, 2930137);
});
test('seedBundled backfills reference fields added later', () async {
// First seed lacks the enrichment fields (an older catalog).
await repo.seedBundled(const [
SpeciesSeed(
scientificName: 'Allium cepa',
commonNames: {
'en': ['Onion'],
},
),
]);
// A later catalog carries them; identity is preserved, fields are filled.
await repo.seedBundled(const [
SpeciesSeed(
scientificName: 'Allium cepa',
wikidataQid: 'Q23485',
gbifKey: 2857697,
viabilityYears: 1,
commonNames: {
'en': ['Onion'],
},
),
]);
final rows = await db.select(db.species).get();
expect(rows.length, 1);
expect(rows.single.wikidataQid, 'Q23485');
expect(rows.single.gbifKey, 2857697);
expect(rows.single.viabilityYears, 1);
});
test('search matches a non-Latin common name', () async {
await repo.seedBundled(const [
SpeciesSeed(
scientificName: 'Solanum lycopersicum',
commonNames: {
'ar': ['طماطم'],
'zh': ['番茄'],
},
),
]);
final ar = await repo.search('طماطم', languageCode: 'ar');
expect(ar.single.scientificName, 'Solanum lycopersicum');
expect(ar.single.commonName, 'طماطم');
final zh = await repo.search('番茄', languageCode: 'zh');
expect(zh.single.commonName, '番茄');
});
test('search respects the limit', () async {
final many = [
for (var i = 0; i < 20; i++)
SpeciesSeed(
scientificName: 'Genus species$i',
commonNames: const {
'en': ['Plant'],
},
),
];
await repo.seedBundled(many);
final results = await repo.search('Genus', limit: 5);
expect(results.length, 5);
});
test('classifyLabel infers the species named in a free-text label', () async {
await repo.seedBundled(_seeds);
final match = await repo.classifyLabel('Tomate cherry', languageCode: 'es');
expect(match, isNotNull);
expect(match!.scientificName, 'Solanum lycopersicum');
expect(match.family, 'Solanaceae');
expect(match.commonName, 'Tomate'); // best name for the locale
// No species named → no suggestion.
expect(await repo.classifyLabel('Girasol gigante'), isNull);
});
group('seedBundledIfNeeded', () {
test('seeds on first run and skips the reload once recorded', () async {
String? stored;
var loads = 0;
Future<List<SpeciesSeed>> load() async {
loads++;
return _seeds;
}
await repo.seedBundledIfNeeded(
version: 3,
readVersion: () async => stored,
writeVersion: (v) async => stored = v,
loadSeeds: load,
);
expect(loads, 1);
expect(stored, '3');
expect(await repo.search('Tomato'), isNotEmpty);
// Same version already recorded → the ~1.5 MB parse is never loaded again.
await repo.seedBundledIfNeeded(
version: 3,
readVersion: () async => stored,
writeVersion: (v) async => stored = v,
loadSeeds: load,
);
expect(loads, 1);
});
test('reseeds when the catalog version changes', () async {
var stored = '3';
var loads = 0;
await repo.seedBundledIfNeeded(
version: 4,
readVersion: () async => stored,
writeVersion: (v) async => stored = v,
loadSeeds: () async {
loads++;
return _seeds;
},
);
expect(loads, 1);
expect(stored, '4');
});
});
}