AquaDX/tools/data_convert.py

218 lines
6.8 KiB
Python

import argparse
import os
import shutil
from pathlib import Path
import orjson
import xmltodict
from hypy_utils import write
from hypy_utils.logging_utils import setup_logger
from hypy_utils.tqdm_utils import pmap
from wand.image import Image
log = setup_logger()
def convert_path(file: Path):
# Get path relative to source
rel = file.relative_to(src)
# If path is one-level under StreamingAssets, ignore it (e.g. StreamingAssets/A000/Data.xml)
if len(rel.parts) <= 2:
return
# Generate target file path
# Ignore the first segment of the relative path, and append to the destination
# Also collapse the single-item directory into the filename
# e.g. {src}/A000/music/music000001/Music.xml -> {dst}/music/000001.json
target = dst / '/'.join(rel.parts[1:-2])
file_id = ''.join(filter(str.isdigit, rel.parts[-2]))
file_id = file_id.zfill(6)
target = target / f'{file_id}.json'
return target
def convert_one(file: Path):
target = convert_path(file)
if target is None:
return
# Read xml
try:
xml = xmltodict.parse(file.read_text())
except Exception as e:
log.info(f'Error parsing {file}: {e}')
return
# There should only be one root element, expand it
assert len(xml) == 1, f'Expected 1 root element, got {len(xml)}'
xml = xml[list(xml.keys())[0]]
# Remove @xmlns:xsi and @xmlns:xsd
if '@xmlns:xsi' in xml:
del xml['@xmlns:xsi']
if '@xmlns:xsd' in xml:
del xml['@xmlns:xsd']
if target.exists():
log.info(f'Overwriting {target}')
# Write json
write(target, orjson.dumps(xml))
def convert_dds(file: Path):
target = convert_path(file)
if target is None:
return
# Convert dds to jpg
try:
with Image(filename=str(file)) as img:
img.format = 'jpeg'
img.save(filename=str(target.with_suffix('.png')))
except Exception as e:
log.info(f'Error converting {file}: {e}')
return
def get(d: dict, *keys: str):
"""
Get the first key that exists in the dictionary
:param d: Dictionary
:param keys: Recursive key in the format of keya.keyb.keyc...
"""
for k in keys:
ks = k.split('.')
cd = d
while len(ks) > 0:
cd = cd.get(ks.pop(0))
if cd is None:
break
if cd is not None:
return cd
return None
def convert_music_mai2(d: dict) -> (str, dict):
return d['name']['id'], {
'name': d['name']['str'],
'ver': d.get('version') or d.get('releaseTagName')['str'],
'composer': d['artistName']['str'],
'genre': d['genreName']['str'] or d['genreNames'],
'bpm': int(d['bpm']),
'lock': f"{d['lockType']} {d['subLockType']}",
'notes': [{
'lv': int(n['level']) + (int(n['levelDecimal']) / 10.0),
'designer': n['notesDesigner']['str'],
'lv_id': n['musicLevelID'],
'notes': int(n['maxNotes']),
} for n in d['notesData']['Notes'] if n['isEnable'] != 'false']
}
def convert_music_chu3(d: dict) -> (str, dict):
return d['name']['id'], {
'name': d['name']['str'],
'ver': d['releaseTagName']['str'],
'composer': d['artistName']['str'],
'genre': get(d, 'genreName.list.StringID.str'),
'lock': d['firstLock'],
'notes': [{
'lv': int(n['level']) + (int(n['levelDecimal']) / 100.0),
'designer': n.get('notesDesigner'),
'lv_id': n['type']['id'],
} for n in d['fumens']['MusicFumenData'] if n['enable'] != 'false']
}
def convert_music_ongeki(d: dict) -> (str, dict):
return d['Name']['id'], {
'name': d['Name']['str'],
'ver': d['VersionID']['id'],
'composer': d['ArtistName']['str'],
'genre': d['Genre']['str'],
'lock': f"{d['CostToUnlock']} {d['IsLockedAtTheBeginning']}",
'notes': [{
'lv': int(n['FumenConstIntegerPart']) + (int(n['FumenConstFractionalPart']) / 100.0),
'lv_id': i,
} for i, n in enumerate(d['FumenData']['FumenData']) if n['FumenFile']['path'] is not None],
'lunatic': d['IsLunatic']
}
if __name__ == '__main__':
agupa = argparse.ArgumentParser()
# Source can be one of the following:
# - maimai/Package/Sinmai_Data/StreamingAssets
# - chusan/App/data
# - ongeki/package/mu3_Data/StreamingAssets/GameData
agupa.add_argument('source', type=str, help='Package/Sinmai_Data/StreamingAssets directory')
agupa.add_argument('destination', type=str, help='Directory to extract to')
agupa.add_argument('-g', '--game', type=str, help='Game to convert', default='mai2', choices=['mai2', 'chu3', 'ongeki'])
args = agupa.parse_args()
src = Path(args.source)
dst = Path(args.destination)
# Special post-convert command to relocate stuff
if args.source == 'post-convert':
ori = dst
dst = dst.parent
# In assetbundle/dir, move each XXX_{id}_XXX.png to assetbundle/dir/{id}.png
for d in os.listdir(dst / 'assetbundle'):
d = dst / 'assetbundle' / d
if not d.is_dir():
continue
log.info(f'Relocating {d}')
for file in d.rglob('*.png'):
id = ''.join(filter(str.isdigit, file.stem))
shutil.move(file, d / f'{id}.png')
exit(0)
# Assert that A000 exists in the source directory
assert (src / 'A000').exists(), f'{src}/A000 does not exist'
# Assert that target directory does not exist
if dst.exists():
if input(f'{dst} already exists, delete? (y/n): ') == 'y':
log.info(f'Deleting {dst}')
shutil.rmtree(dst)
# Find all xml files in the source directory
files = list(src.rglob('*.xml'))
log.info(f'Found {len(files)} xml files')
# Multithreaded map
pmap(convert_one, files, desc='Converting', unit='file', chunksize=50)
log.info('> Finished converting')
# Find all .dds files in the source A000 directory
dds_files = list(src.rglob('*.dds'))
log.info(f'Found {len(dds_files)} dds files')
# Convert and copy dds files (CPU-intensive)
pmap(convert_dds, dds_files, desc='Converting DDS', unit='file', chunksize=50, max_workers=os.cpu_count() - 2)
log.info('> Finished converting DDS')
# Convert all music
log.info('Combining music')
music_files = list(dst.rglob('music/*.json'))
log.info(f'> Found {len(music_files)} music files')
jsons = [orjson.loads(f.read_text()) for f in music_files]
converter = {'mai2': convert_music_mai2, 'chu3': convert_music_chu3, 'ongeki': convert_music_ongeki}[args.game]
combined = {k: v for k, v in [converter(d) for d in jsons]}
# Write combined music
write(dst / '00/all-music.json', orjson.dumps(combined))