sporada/nopaque
1
0
Fork 0
mirror of https://gitlab.ub.uni-bielefeld.de/sfb1288inf/nopaque.git synced 2024-12-25 10:54:18 +00:00
Code Issues Projects Releases Wiki Activity

Implement force download parameter in model insert_defaults methods

Browse Source
This commit is contained in:
Patrick Jentsch 2023-11-28 12:10:55 +01:00
parent bdcc80a66f
commit 9c22370eea
1 changed files with 71 additions and 67 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
app/models.py
View File

@ -953,7 +953,7 @@ class TesseractOCRPipelineModel(FileMixin, HashidMixin, db.Model):
return self.user.hashid return self.user.hashid
@staticmethod @staticmethod
def insert_defaults(): def insert_defaults(force_download=False):
nopaque_user = User.query.filter_by(username='nopaque').first() nopaque_user = User.query.filter_by(username='nopaque').first()
defaults_file = os.path.join( defaults_file = os.path.join(
os.path.dirname(os.path.abspath(__file__)), os.path.dirname(os.path.abspath(__file__)),
@ -966,6 +966,7 @@ class TesseractOCRPipelineModel(FileMixin, HashidMixin, db.Model):
if model is not None: if model is not None:
model.compatible_service_versions = m['compatible_service_versions'] model.compatible_service_versions = m['compatible_service_versions']
model.description = m['description'] model.description = m['description']
model.filename = f'{model.id}.traineddata'
model.publisher = m['publisher'] model.publisher = m['publisher']
model.publisher_url = m['publisher_url'] model.publisher_url = m['publisher_url']
model.publishing_url = m['publishing_url'] model.publishing_url = m['publishing_url']
@ -973,7 +974,7 @@ class TesseractOCRPipelineModel(FileMixin, HashidMixin, db.Model):
model.is_public = True model.is_public = True
model.title = m['title'] model.title = m['title']
model.version = m['version'] model.version = m['version']
continue else:
model = TesseractOCRPipelineModel( model = TesseractOCRPipelineModel(
compatible_service_versions=m['compatible_service_versions'], compatible_service_versions=m['compatible_service_versions'],
description=m['description'], description=m['description'],
@ -990,6 +991,7 @@ class TesseractOCRPipelineModel(FileMixin, HashidMixin, db.Model):
db.session.flush(objects=[model]) db.session.flush(objects=[model])
db.session.refresh(model) db.session.refresh(model)
model.filename = f'{model.id}.traineddata' model.filename = f'{model.id}.traineddata'
if not os.path.exists(model.path) or force_download:
r = requests.get(m['url'], stream=True) r = requests.get(m['url'], stream=True)
pbar = tqdm( pbar = tqdm(
desc=f'{model.title} ({model.filename})', desc=f'{model.title} ({model.filename})',
@ -1080,7 +1082,7 @@ class SpaCyNLPPipelineModel(FileMixin, HashidMixin, db.Model):
return self.user.hashid return self.user.hashid
@staticmethod @staticmethod
def insert_defaults(): def insert_defaults(force_download=False):
nopaque_user = User.query.filter_by(username='nopaque').first() nopaque_user = User.query.filter_by(username='nopaque').first()
defaults_file = os.path.join( defaults_file = os.path.join(
os.path.dirname(os.path.abspath(__file__)), os.path.dirname(os.path.abspath(__file__)),
@ -1093,6 +1095,7 @@ class SpaCyNLPPipelineModel(FileMixin, HashidMixin, db.Model):
if model is not None: if model is not None:
model.compatible_service_versions = m['compatible_service_versions'] model.compatible_service_versions = m['compatible_service_versions']
model.description = m['description'] model.description = m['description']
model.filename = m['url'].split('/')[-1]
model.publisher = m['publisher'] model.publisher = m['publisher']
model.publisher_url = m['publisher_url'] model.publisher_url = m['publisher_url']
model.publishing_url = m['publishing_url'] model.publishing_url = m['publishing_url']
@ -1101,10 +1104,11 @@ class SpaCyNLPPipelineModel(FileMixin, HashidMixin, db.Model):
model.title = m['title'] model.title = m['title']
model.version = m['version'] model.version = m['version']
model.pipeline_name = m['pipeline_name'] model.pipeline_name = m['pipeline_name']
continue else:
model = SpaCyNLPPipelineModel( model = SpaCyNLPPipelineModel(
compatible_service_versions=m['compatible_service_versions'], compatible_service_versions=m['compatible_service_versions'],
description=m['description'], description=m['description'],
filename=m['url'].split('/')[-1],
publisher=m['publisher'], publisher=m['publisher'],
publisher_url=m['publisher_url'], publisher_url=m['publisher_url'],
publishing_url=m['publishing_url'], publishing_url=m['publishing_url'],
@ -1118,7 +1122,7 @@ class SpaCyNLPPipelineModel(FileMixin, HashidMixin, db.Model):
db.session.add(model) db.session.add(model)
db.session.flush(objects=[model]) db.session.flush(objects=[model])
db.session.refresh(model) db.session.refresh(model)
model.filename = m['url'].split('/')[-1] if not os.path.exists(model.path) or force_download:
r = requests.get(m['url'], stream=True) r = requests.get(m['url'], stream=True)
pbar = tqdm( pbar = tqdm(
desc=f'{model.title} ({model.filename})', desc=f'{model.title} ({model.filename})',