beeeeg update

scripts/build-pokemon-catalog.py

@@ -0,0 +1,96 @@
+"""Build the species-level catalog from a pinned PokéAPI CSV snapshot.
+
+Run with --source-dir to reuse downloaded CSVs. Default downloads into a temporary
+folder. Encounter absence is not evidence of trade/transfer/event exclusivity.
+"""
+import argparse
+import concurrent.futures
+import csv
+import json
+from pathlib import Path
+import tempfile
+import urllib.request
+
+REVISION = '575291cdb197a7e3a320297be276c9de4ef8401a'
+TABLES = ['pokemon', 'pokemon_species', 'pokemon_species_names', 'pokemon_types',
+          'pokemon_stats', 'type_names', 'versions', 'version_names', 'version_groups',
+          'pokemon_dex_numbers', 'pokedex_version_groups', 'pokemon_game_indices', 'encounters']
+
+
+def build(directory):
+    def rows(name):
+        with (directory / f'{name}.csv').open() as f:
+            return list(csv.DictReader(f))
+
+    species = rows('pokemon_species')
+    pokemon = rows('pokemon')
+    species_for = {int(p['id']): int(p['species_id']) for p in pokemon}
+    defaults = {int(p['species_id']): int(p['id']) for p in pokemon if p['is_default'] == '1'}
+    names = {int(p['pokemon_species_id']): p['name'] for p in rows('pokemon_species_names') if p['local_language_id'] == '9'}
+    type_names = {int(t['type_id']): t['name'] for t in rows('type_names') if t['local_language_id'] == '9'}
+    types = {}
+    for p in rows('pokemon_types'):
+        types.setdefault(int(p['pokemon_id']), {})[int(p['slot'])] = type_names[int(p['type_id'])]
+    stats = {}
+    for p in rows('pokemon_stats'):
+        stats.setdefault(int(p['pokemon_id']), {})[int(p['stat_id'])] = int(p['base_stat'])
+    stat_fields = ['hp', 'attack', 'defense', 'specialAttack', 'specialDefense', 'speed']
+    mons = []
+    for s in species:
+        number = int(s['id'])
+        default = defaults[number]
+        mons.append(dict(number=number, name=names[number], generation=int(s['generation_id']),
+                         type1=types[default][1], type2=types[default].get(2),
+                         **{name: stats[default][i + 1] for i, name in enumerate(stat_fields)}))
+
+    groups = {int(g['id']): g for g in rows('version_groups')}
+    versions = rows('versions')
+    game_names = {int(v['version_id']): v['name'] for v in rows('version_names') if v['local_language_id'] == '9'}
+    games = [dict(sourceId=int(v['id']), name=game_names.get(int(v['id']), v['identifier'].replace('-', ' ').title()),
+                  generation=int(groups[int(v['version_group_id'])]['generation_id'])) for v in versions]
+    for game, version in zip(games, versions):
+        if version['identifier'].endswith('-japan'):
+            game['name'] += ' (Japan)'
+    versions_for_group = {}
+    for v in versions:
+        versions_for_group.setdefault(int(v['version_group_id']), []).append(int(v['id']))
+    groups_for_dex = {}
+    for d in rows('pokedex_version_groups'):
+        groups_for_dex.setdefault(int(d['pokedex_id']), []).append(int(d['version_group_id']))
+    entries = {}
+    # Membership and internal game indices establish presence, not how it is obtained.
+    for p in rows('pokemon_dex_numbers'):
+        for group in groups_for_dex.get(int(p['pokedex_id']), []):
+            for version in versions_for_group.get(group, []):
+                entries[(int(p['species_id']), version)] = 'unknown'
+    for p in rows('pokemon_game_indices'):
+        if int(p['game_index']) > 0:
+            entries[(species_for[int(p['pokemon_id'])], int(p['version_id']))] = 'unknown'
+    for p in rows('encounters'):
+        entries[(species_for[int(p['pokemon_id'])], int(p['version_id']))] = 'catchable'
+    return dict(source={'repository': 'https://github.com/PokeAPI/pokeapi', 'revision': REVISION,
+                        'scope': 'All species and game versions in this snapshot; default-form types/stats. Unknown means presence without confirmed encounter data. Missing pairs have no source record, not proven absence.'},
+                pokemon=mons, games=games, entries=[[p, g, availability] for (p, g), availability in sorted(entries.items())])
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--source-dir', type=Path)
+    args = parser.parse_args()
+    with tempfile.TemporaryDirectory() as temporary:
+        directory = args.source_dir or Path(temporary)
+        if not args.source_dir:
+            def download(table):
+                url = f'https://raw.githubusercontent.com/PokeAPI/pokeapi/{REVISION}/data/v2/csv/{table}.csv'
+                with urllib.request.urlopen(url, timeout=60) as response:
+                    (directory / f'{table}.csv').write_bytes(response.read())
+            with concurrent.futures.ThreadPoolExecutor(max_workers=4) as executor:
+                list(executor.map(download, TABLES))
+        catalog = build(directory)
+    target = Path(__file__).resolve().parent.parent / 'data/pokemon-catalog.json'
+    target.write_text(json.dumps(catalog, ensure_ascii=False, separators=(',', ':')) + '\n')
+    print(f"Built {len(catalog['pokemon'])} species, {len(catalog['games'])} versions, {len(catalog['entries'])} entries")
+
+
+if __name__ == '__main__':
+    main()