• Hi Guest!

    This is a notice regarding recent upgrades to the Hub. Over the last month, we have added several new features to improve your experience.
    You can check out the details in our official announcement!
Resource icon
Had a few errors with encoding in files, and trying to read directories that were named .var
My fixes are below - sorry about formatting. Try to format it online?

#!/bin/python3

#usage:
# python3 extract_appearance.py "Saves/scene/*.json" "AddonPackages/*.var"

import json
import sys
import os
import zipfile
import re
import glob
import traceback
import logging

SCRIPT_VERSION = '2.1'

# logging.getLogger().setLevel(logging.WARNING)
logging.getLogger().setLevel(logging.INFO)

try:
from tqdm import tqdm as tqdmfunc
except ImportError:
logging.warning("tqdm not installed, not showing progress bar")

def tqdmfunc(iterator):
yield from iterator


def update_dict(d, fn, key=None):
if isinstance(d, dict):
for k, v in d.items():
d[k] = update_dict(v, fn, k)
# return not needed to update dicts by reference, but used
# here to assist in assigning subtrees
return d
elif isinstance(d, list):
return [update_dict(i, fn) for i in d]
else:
return fn(d, key)


def update_dict_selfname(d, self_name):
def update_fn(v, key=None):
if key == 'id' and isinstance(v, str):
return v.replace("SELF:/", self_name + ":/")
else:
return v

update_dict(d, update_fn)


def getPersonAtoms(scenejson):
for atom in scenejson['atoms']:
if atom['type'] == "Person":
yield atom


class BaseExtractor(object):
"""docstring for BaseExtractor"""
out_dir_verify = "./customLooks"
#out_dir_verify = "./Custom/Atom/Person/"

def __init__(self, noclobber=False):
super(BaseExtractor, self).__init__()
self.noclobber = noclobber
self.dupes = set()
self.out_dir = self.getOutDir()

os.makedirs(self.out_dir, exist_ok=True)

def getOutDir(self):
if not os.path.isdir(self.out_dir_verify):
logging.error("Script is not being run from the VaM root directory!")
logging.warning("Can't cleanly merge into Person/Appearance presets; making ExtractedAppearance folder instead.")
return self.out_dir_fallback
else:
return self.out_dir_wanted

def extractFromVar(self, var):
# Check if the path is a file and has a .var extension
if not os.path.isfile(var) or not var.endswith('.var'):
print(f"{var} is not a valid .var file or is a directory. Skipping.")
return

__, filename = os.path.split(var)
self_name, __ = os.path.splitext(filename)

author = self_name.split('.')[0]

try:
with zipfile.ZipFile(var) as varzip:
infodict = {
zi.filename: zi
for zi in varzip.infolist()
}
zip_scenes = {
k: v for k, v in infodict.items()
if k.startswith("Saves/scene") and k.endswith(".json")
}
for scenejsonpath in zip_scenes.keys():
logging.debug("%s:%s", var, scenejsonpath)

__, sjp_filename = os.path.split(scenejsonpath)
sjp_plain, sjp_ext = os.path.splitext(sjp_filename)

with varzip.open(scenejsonpath, 'r') as fp:
scenejson = json.load(fp)

if False:
update_dict_selfname(scenejson, self_name)
else:
# This method is potentially inaccurate but MUCH faster:
scenejson = json.loads(json.dumps(scenejson).replace("SELF:/", self_name + ":/"))

def readthumb():
return varzip.open(scenejsonpath.replace('.json', '.jpg'), 'r')

def outnameFn(atom):
return f"Preset_{self_name}.{sjp_plain}{atom['id'].replace('Person', '')}".replace('#', '').replace('/', '-')

self.extractFromSceneJson(scenejson, outnameFn, readthumb)

except json.JSONDecodeError:
print(var)
traceback.print_exc()
os.rename(var, var + '.invalid')
except zipfile.BadZipFile:
print(var)
traceback.print_exc()
os.rename(var, var + '.invalid')
except Exception as e:
print(f"An error occurred with file {var}: {e}. Skipping to next file.")
traceback.print_exc()

def extractFromSceneJsonPath(self, scenejsonpath):
__, sjp_filename = os.path.split(scenejsonpath)
sjp_plain, sjp_ext = os.path.splitext(sjp_filename)

print('loading: ' + scenejsonpath)
def readthumb():
return open(scenejsonpath.replace('.json', '.jpg'), 'rb')

def outnameFn(atom):
return f"Preset_!local.{sjp_plain}{atom['id'].replace('Person', '')}".replace('#', '').replace('/', '-')

try:
with open(scenejsonpath, 'r', encoding='utf-8') as fp:
self.extractFromSceneJson(json.load(fp), outnameFn, readthumb)

except json.JSONDecodeError:
print(f"Error decoding JSON in file {scenejsonpath}. Skipping to next file.")
except Exception as e:
print(f"An error occurred with file {scenejsonpath}: {e}. Skipping to next file.")


def presetOutname(self, outname, ext):
return os.path.join(self.out_dir, outname + ext)

def extractFromSceneJson(self, scenejson, outnameFn, readthumb):
# print([atom['id'] for atom in scenejson['atoms'] if atom['type'] == 'Person'])

for atom in getPersonAtoms(scenejson):
outname = outnameFn(atom)
if self.noclobber and os.path.isfile(self.presetOutname(outname, '.vap')):
logging.info(f"Skipping {outname!r}")
continue

# print(atom['id'])
preset = {
"setUnlistedParamsToDefault": "true",
'storables': self.filterPersonStorables(atom)
}

self.savePreset(preset, outname, readthumb)

def savePreset(self, preset, outname, readthumb):
preset_hash = json.dumps(preset, sort_keys=True)
if preset_hash not in self.dupes:
outpath = self.presetOutname(outname, '.vap')
with open(outpath, 'w', encoding='utf-8') as fp:
logging.info("-> %s", outpath)
fp.write(json.dumps(preset, indent=3, ensure_ascii=False))
try:
img_outpath = self.presetOutname(outname, '.jpg')
with open(img_outpath, 'wb') as fp:
with readthumb() as fp2:
fp.write(fp2.read())
except (KeyError, FileNotFoundError):
# No image in archive
pass
self.dupes.add(preset_hash)
else:
# logging.warn("Skip duplicate %s", outname)
pass

def filterPersonStorables(self, atom):
raise NotImplementedError()

class AppearanceExtractor(BaseExtractor):
out_dir_wanted = "./Custom/Atom/Person/Appearance/extracted/"
out_dir_fallback = "./ExtractedAppearance/"

def filterPersonStorables(self, atom):
storables = atom['storables']
storables_new = []
for storable in storables:
storable.pop("position", None)
storable.pop("rotation", None)

# Remove all keys containing these substrings
for key in [*storable.keys()]:
if any(ss in key.lower() for ss in ['position', 'rotation']):
storable.pop(key)

# Remove all storables whose ids contain these substrings
if any(ss in storable['id'].lower() for ss in ["control", "trigger", "plugin", "preset", "animation"]):
# print("pop", storable['id'])
storables.remove(storable)
continue

# Remove transient morphs
if storable['id'] == "geometry" and 'morphs' in storable:
for morph in [*storable['morphs']]:
if any(re.match(ss, morph.get('uid', '')) for ss in [
r'Breast Impact',
r'^OpenXXL$',
r'^Eyelids (Top|Bottom) (Down|Up) (Left|Right)$', r'Brow .*(Up|Down)', r'^(Left|Right) Fingers',
r'^Mouth Open', r'Tongue In-Out', r'^Smile',
'Shock', 'Surprise', 'Fear', 'Pain', 'Concentrate', 'Eyes Closed', r'^Flirting',
]):
# print(morph)
storable['morphs'].remove(morph)

# Remove storables with just an id and no properties
if len(storable.keys()) > 1:
storables_new.append(storable)

return storables_new


class OutfitExtractor(BaseExtractor):
out_dir_wanted = "./Custom/Atom/Person/Clothing/extracted/"
out_dir_fallback = "./ExtractedOutfits/"

def filterPersonStorables(self, atom):
storables = atom['storables']

clothing_master = next(filter(lambda v: v['id'] == 'geometry' and 'clothing' in v.keys(), storables))
clothing = clothing_master['clothing']
try:
clothing_ids = [c.get('internalId', c['id']) for c in clothing]
except KeyError:
print(clothing)
raise

storables_new = [
{
"id": "geometry",
"clothing": clothing
}
]
for storable in storables:
if any(storable['id'].startswith(id_) for id_ in clothing_ids):
storables_new.append(storable)
# else:
# print(storable['id'], clothing_ids)

# print(storables_new)
return storables_new

def add_bool_arg(parser, name, default=False, **kwargs):
group = parser.add_mutually_exclusive_group(required=False)
group.add_argument('--' + name, dest=name, action='store_true', **kwargs)
group.add_argument('--no-' + name, dest=name, action='store_false')
parser.set_defaults(**{name: default})


def parse_args():
import argparse
parser = argparse.ArgumentParser(
description="Exports presets",
formatter_class=argparse.ArgumentDefaultsHelpFormatter
)
parser.add_argument('input_globs', help="Input files or fileglobs", nargs='+')
add_bool_arg(parser, 'appearance', default=True, help="Extract appearance (body) presets")
add_bool_arg(parser, 'outfit', default=False, help="Extract outfit (clothing) presets")

parser.add_argument('--no-clobber', action='store_true', help="If set, don't overwrite files.", default=False)
return parser.parse_args()


def main():
print('Appearance Preset Extractor', SCRIPT_VERSION, 'by cartitolaire')

if not sys.argv[1:]:
print('')
print('This tool automatically exports presets from existing scenes and collects them in an "extracted" folder under presets. The presets will use the thumbnail for the source scene.')
print('')
print("Example usages:")
print(' ', sys.argv[0], 'AddonPackages/Anon.SceneName.1.var')
print(' ', sys.argv[0], '"AddonPackages/Anon*.var"')
print(' ', sys.argv[0], '"Saves/scene/*.json" "AddonPackages/*.var"')
print(' ', sys.argv[0], '--outfit --no-appearance AddonPackages/Anon.SceneName.1.var')
return

args = parse_args()

extractors = []
if args.appearance:
extractors.append(AppearanceExtractor)
if args.outfit:
extractors.append(OutfitExtractor)

try:
for etype in extractors:
extractor = etype(noclobber=args.no_clobber)
iterable = sum((glob.glob(a, recursive=True) for a in args.input_globs), [])
for filepath in tqdmfunc(iterable):
__, filename = os.path.split(filepath)
self_name, ext = os.path.splitext(filename)

try:
if ext == ".var":
extractor.extractFromVar(filepath)
elif ext == ".json":
extractor.extractFromSceneJsonPath(filepath)
else:
raise NotImplementedError(ext)
except KeyError:
pass

except KeyboardInterrupt:
return


if __name__ == "__main__":
main()
Upvote 0
After extracting about 1000 presets from the scenes, it seems that this script handles only around 40% of them. In 40% of cases it does not transfer character morphs from json to vap file. It does not matter if there are 1 or 3 characters in the scene. I wish it would be corrected.
EDIT: I found out that if I delete 4 lines - 195 to 199, all morphs are transfered successfully to the VAP file. strange. Could the author look what happens?
cartitolaire
cartitolaire
Can you send an example of a problematic scene? Those lines are meant to filter out properties that are irrelevant to the pose, which helps with filesizes and deduplication, but it could somehow be catching something significant.
Upvote 1
Question: If you run it twice will it skip ones already created?
Upvote 1
Extra points for properly shaming look creators who don't package appearance presets in their .vars
Upvote 1
I have no basic python knowledge, but I did it using your explanation and chatgpt

It works well and has been very successful

thank you so much bro!
Upvote 0
It did a good job, just kept crashing with any version of ascorad's Heads Will Roll scene. Suggest to look into why it did that.
cartitolaire
cartitolaire
Crashing? Hanging, or crashing with a stacktrace?
Upvote 0
Thank you so much, i hate all those "scenes" that only shows the look and nothing more.
cartitolaire
cartitolaire
I hate that pattern so much.
Upvote 1
Amazing
Upvote 0
Ok, call me impressed. This script plowed through 1300 vars in a few minutes. It's sooo much faster than "Var Looks Scanner plugin". Thank you for this incredible timesaver. You might consider compiling this with cx_freeze and releasing it as an executable. Kudos!
Upvote 0
Very handy utility
Upvote 0
I was getting some errors, so I had to put "encoding='utf8'" in some places in the code, but now it is working fine. Although I still work on legacy 'appearances' and I will have to manually convert all files to that format, your script is going to save a lot of my time.

I still get "json.decoder.JSONDecodeError: Expecting property name enclosed in double quotes" errors on some scenes, but I think those are just corrupted JSON files and it has nothing to do with your code.
cartitolaire
cartitolaire
Interesting. I didn’t encounter that with any of my scenes but I’ll add the explicit encoding, that’s probably best practice anyway.
Upvote 0
You're a god !
Upvote 0
saves a lot of time!
Upvote 0
Back
Top Bottom