def create(request):
if request.method == 'POST':
dataset = Dataset()
dataset.owner = request.user
dataset.name = request.POST['name']
dataset.number_of_labels = request.POST['number_of_labels']
dataset.description = request.POST['description']
if not dataset.privacy_validation(request.POST['privacy']):
return render(request, '400.html', status=400)
dataset_file = request.FILES['dataset']
reader = csv.reader(dataset_file, encoding='utf-8')
header_list = reader.next()
label_name = request.POST.get('label_name', 'CLASS')
append_label_column = request.POST.get('append_label_column', False)
if not append_label_column:
label_index = header_list.index(label_name)
header_list.pop(label_index)
header_list.append(label_name)
dataset.header = csvlist_to_string(header_list).strip()
dataset.save()
samples_count = 0
for row_list in reader:
samples_count += 1
if not append_label_column:
label_string = row_list.pop(label_index)
row = csvlist_to_string(row_list).strip()
sample = Sample(dataset=dataset,
data=row,
original_index=samples_count)
sample.save()
if not append_label_column and label_string:
label = Label(owner=request.user,
sample=sample,
label=label_string)
label.save()
sample.times_labeled = 1
sample.save()
dataset.number_of_samples = samples_count
dataset.save()
return HttpResponseRedirect(
reverse('datasets_show', args=(dataset.id, )))
def create(request):
if request.method == 'POST':
dataset = Dataset()
dataset.owner = request.user
dataset.name = request.POST['name']
dataset.number_of_labels = request.POST['number_of_labels']
dataset.description = request.POST['description']
if not dataset.privacy_validation(request.POST['privacy']):
return render(request, '400.html', status=400)
dataset_file = request.FILES['dataset']
reader = csv.reader(dataset_file, encoding='utf-8')
header_list = reader.next()
label_name = request.POST.get('label_name', 'CLASS')
append_label_column = request.POST.get('append_label_column', False)
if not append_label_column:
label_index = header_list.index(label_name)
header_list.pop(label_index)
header_list.append(label_name)
dataset.header = csvlist_to_string(header_list).strip()
dataset.save()
samples_count = 0
for row_list in reader:
samples_count += 1
if not append_label_column:
label_string = row_list.pop(label_index)
row = csvlist_to_string(row_list).strip()
sample = Sample(dataset=dataset, data=row,
original_index=samples_count)
sample.save()
if not append_label_column and label_string:
label = Label(owner=request.user, sample=sample,
label=label_string)
label.save()
sample.times_labeled = 1
sample.save()
dataset.number_of_samples = samples_count
dataset.save()
return HttpResponseRedirect(
reverse('datasets_show', args=(dataset.id,)))
def _update_datasets(app):
"""
Updates from conp-datasets
"""
from app import db
from app.models import ArkId
from app.models import Dataset as DBDataset
from app.models import DatasetAncestry as DBDatasetAncestry
from sqlalchemy import exc
from datalad import api
from datalad.api import Dataset as DataladDataset
import fnmatch
import json
from pathlib import Path
import git
datasetsdir = Path(app.config['DATA_PATH']) / 'conp-dataset'
datasetsdir.mkdir(parents=True, exist_ok=True)
# Initialize the git repository object
try:
repo = git.Repo(datasetsdir)
except git.exc.InvalidGitRepositoryError:
repo = git.Repo.clone_from('https://github.com/CONP-PCNO/conp-dataset',
datasetsdir,
branch='master')
# Update to latest commit
origin = repo.remotes.origin
origin.pull('master')
repo.submodule_update(recursive=False, keep_going=True)
d = DataladDataset(path=datasetsdir)
if not d.is_installed():
api.clone(source='https://github.com/CONP-PCNO/conp-dataset',
path=datasetsdir)
d = DataladDataset(path=datasetsdir)
try:
d.install(path='', recursive=True)
except Exception as e:
print("\033[91m")
print("[ERROR ] An exception occurred in datalad update.")
print(e.args)
print("\033[0m")
return
print('[INFO ] conp-dataset update complete')
print('[INFO ] Updating subdatasets')
for ds in d.subdatasets():
print('[INFO ] Updating ' + ds['gitmodule_url'])
subdataset = DataladDataset(path=ds['path'])
if not subdataset.is_installed():
try:
api.clone(source=ds['gitmodule_url'], path=ds['path'])
subdataset = DataladDataset(path=ds['path'])
subdataset.install(path='')
except Exception as e:
print("\033[91m")
print(
"[ERROR ] An exception occurred in datalad install for " +
str(ds) + ".")
print(e.args)
print("\033[0m")
continue
# The following relates to the DATS.json files
# of the projects directory in the conp-dataset repo.
# Skip directories that aren't projects.
patterns = [app.config['DATA_PATH'] + '/conp-dataset/projects/*']
if not any(
fnmatch.fnmatch(ds['path'], pattern) for pattern in patterns):
continue
dirs = os.listdir(ds['path'])
descriptor = ''
for file in dirs:
if fnmatch.fnmatch(file.lower(), 'dats.json'):
descriptor = file
if descriptor == '':
print("\033[91m")
print('[ERROR ] DATS.json file can`t be found in ' + ds['path'] +
".")
print("\033[0m")
continue
try:
with open(os.path.join(ds['path'], descriptor), 'r') as f:
dats = json.load(f)
except Exception as e:
print("\033[91m")
print("[ERROR ] Descriptor file can't be read.")
print(e.args)
print("\033[0m")
continue
# use dats.json data to fill the datasets table
# avoid duplication / REPLACE instead of insert
dataset = DBDataset.query.filter_by(
dataset_id=ds['gitmodule_name']).first()
# pull the timestamp of the first commit in the git log for the dataset create date
createDate = datetime.utcnow()
try:
createTimeStamp = os.popen(
"git -C {} log --pretty=format:%ct --reverse | head -1".format(
ds['path'])).read()
createDate = datetime.fromtimestamp(int(createTimeStamp))
except Exception:
print("[ERROR ] Create Date couldnt be read.")
firstMergeDate = datetime.utcnow()
try:
firstMergeTimeStamp = os.popen(
"git -C {} log --pretty=format:%ct --reverse {} | head -1".
format(app.config['DATA_PATH'] + "/conp-dataset",
ds['path'])).read()
firstMergeDate = datetime.fromtimestamp(int(firstMergeTimeStamp))
except Exception:
print(
"[ERROR ] First merge date of the submodule dataset could not be read."
)
# last commit in the git log for the dataset update date
updateDate = datetime.utcnow()
try:
createTimeStamp = os.popen(
"git -C {} log --pretty=format:%ct | head -1".format(
ds['path'])).read()
updateDate = datetime.fromtimestamp(int(createTimeStamp))
except Exception:
print("[ERROR ] Update Date couldnt be read.")
# get the remote URL
remoteUrl = None
try:
remoteUrl = os.popen(
"git -C {} config --get remote.origin.url".format(
ds['path'])).read()
except Exception:
print("[ERROR ] Remote URL couldnt be read.")
if dataset is None:
dataset = DBDataset()
dataset.dataset_id = ds['gitmodule_name']
dataset.date_created = createDate
dataset.date_added_to_portal = firstMergeDate
if dataset.date_created != createDate:
dataset.date_created = createDate
# check for dataset ancestry
extraprops = dats.get('extraProperties', [])
for prop in extraprops:
if prop.get('category') == 'parent_dataset_id':
for x in prop.get('values', []):
if x.get('value', None) is None:
continue
datasetAncestry = DBDatasetAncestry()
datasetAncestry.id = str(uuid.uuid4())
datasetAncestry.parent_dataset_id = 'projects/' + \
x.get('value', None)
datasetAncestry.child_dataset_id = dataset.dataset_id
try:
db.session.merge(datasetAncestry)
db.session.commit()
except exc.IntegrityError:
# we already have a record of this ancestry
db.session.rollback()
if not dataset.date_added_to_portal:
dataset.date_added_to_portal = firstMergeDate
dataset.date_updated = updateDate
dataset.fspath = ds['path']
dataset.remoteUrl = remoteUrl
dataset.description = dats.get('description',
'No description in DATS.json')
dataset.name = dats.get('title', os.path.basename(dataset.dataset_id))
db.session.merge(dataset)
db.session.commit()
# if the dataset does not have an ARK identifier yet, generate it
dataset_with_ark_id_list = [
row[0] for row in db.session.query(ArkId.dataset_id).all()
]
if dataset.dataset_id not in dataset_with_ark_id_list:
new_ark_id = ark_id_minter(app, 'dataset')
save_ark_id_in_database(app, 'dataset', new_ark_id,
dataset.dataset_id)
print('[INFO ] ' + ds['gitmodule_name'] + ' updated.')
def _update_datasets(app):
"""
Updates from conp-datasets
"""
from app import db, config
from app.models import Dataset as DBDataset
from datalad import api
from datalad.api import Dataset as DataladDataset
import fnmatch
import json
datasetspath = app.config['DATA_PATH']
d = DataladDataset(path=datasetspath + '/conp-dataset')
if not d.is_installed():
api.clone(source='https://github.com/CONP-PCNO/conp-dataset',
path=datasetspath + '/conp-dataset')
d = DataladDataset(path=datasetspath + '/conp-dataset')
d.install(path='', recursive=True)
try:
d.update(path='', merge=True, recursive=True)
except Exception as e:
print("\033[91m")
print("[ERROR ] An exception occurred in datalad update.")
print(e.args)
print("\033[0m")
return
print('[INFO ] conp-dataset update complete')
print('[INFO ] Updating subdatasets')
for ds in d.subdatasets():
print('[INFO ] Updating ' + ds['gitmodule_url'])
subdataset = DataladDataset(path=ds['path'])
if not subdataset.is_installed():
try:
api.clone(source=ds['gitmodule_url'], path=ds['path'])
subdataset = DataladDataset(path=ds['path'])
subdataset.install(path='')
except Exception as e:
print("\033[91m")
print(
"[ERROR ] An exception occurred in datalad install for " +
str(ds) + ".")
print(e.args)
print("\033[0m")
continue
dirs = os.listdir(ds['path'])
descriptor = ''
for file in dirs:
if fnmatch.fnmatch(file.lower(), 'dats.json'):
descriptor = file
if descriptor == '':
print("\033[91m")
print('[ERROR ] DATS.json file can`t be found in ' + ds['path'] +
".")
print("\033[0m")
continue
try:
with open(os.path.join(ds['path'], descriptor), 'r') as f:
dats = json.load(f)
except Exception as e:
print("\033[91m")
print("[ERROR ] Descriptor file can't be read.")
print(e.args)
print("\033[0m")
continue
# use dats.json data to fill the datasets table
# avoid duplication / REPLACE instead of insert
dataset = DBDataset.query.filter_by(
dataset_id=ds['gitmodule_name']).first()
if dataset is None:
dataset = DBDataset()
dataset.dataset_id = ds['gitmodule_name']
dataset.date_created = datetime.utcnow()
dataset.date_updated = datetime.utcnow()
dataset.fspath = ds['path']
dataset.description = dats.get('description',
'No description in DATS.json')
dataset.name = dats.get('title', os.path.basename(dataset.dataset_id))
db.session.merge(dataset)
db.session.commit()
print('[INFO ] ' + ds['gitmodule_name'] + ' updated.')
def _update_datasets(app):
"""
Updates from conp-datasets
"""
from app import db, config
from app.models import Dataset as DBDataset
from datalad import api
from datalad.api import Dataset as DataladDataset
import fnmatch
import json
from pathlib import Path
import git
datasetsdir = Path(app.config['DATA_PATH']) / 'conp-dataset'
datasetsdir.mkdir(parents=True, exist_ok=True)
# Initialize the git repository object
try:
repo = git.Repo(datasetsdir)
except git.exc.InvalidGitRepositoryError as e:
repo = git.Repo.clone_from(
'https://github.com/CONP-PCNO/conp-dataset',
datasetsdir,
branch='master'
)
# Update to latest commit
origin = repo.remotes.origin
origin.pull('master')
repo.submodule_update(recursive=False, keep_going=True)
d = DataladDataset(path=datasetsdir)
if not d.is_installed():
api.clone(
source='https://github.com/CONP-PCNO/conp-dataset',
path=datasetsdir
)
d = DataladDataset(path=datasetsdir)
try:
d.install(path='', recursive=True)
except Exception as e:
print("\033[91m")
print("[ERROR ] An exception occurred in datalad update.")
print(e.args)
print("\033[0m")
return
print('[INFO ] conp-dataset update complete')
print('[INFO ] Updating subdatasets')
for ds in d.subdatasets():
print('[INFO ] Updating ' + ds['gitmodule_url'])
subdataset = DataladDataset(path=ds['path'])
if not subdataset.is_installed():
try:
api.clone(
source=ds['gitmodule_url'],
path=ds['path']
)
subdataset = DataladDataset(path=ds['path'])
subdataset.install(path='')
except Exception as e:
print("\033[91m")
print("[ERROR ] An exception occurred in datalad install for " + str(ds) + ".")
print(e.args)
print("\033[0m")
continue
dirs = os.listdir(ds['path'])
descriptor = ''
for file in dirs:
if fnmatch.fnmatch(file.lower(), 'dats.json'):
descriptor = file
if descriptor == '':
print("\033[91m")
print('[ERROR ] DATS.json file can`t be found in ' + ds['path'] + ".")
print("\033[0m")
continue
try:
with open(os.path.join(ds['path'], descriptor), 'r') as f:
dats = json.load(f)
except Exception as e:
print("\033[91m")
print("[ERROR ] Descriptor file can't be read.")
print(e.args)
print("\033[0m")
continue
# use dats.json data to fill the datasets table
# avoid duplication / REPLACE instead of insert
dataset = DBDataset.query.filter_by(dataset_id=ds['gitmodule_name']).first()
if dataset is None:
dataset = DBDataset()
dataset.dataset_id = ds['gitmodule_name']
dataset.date_created = datetime.utcnow()
dataset.date_updated = datetime.utcnow()
dataset.fspath = ds['path']
dataset.description = dats.get('description', 'No description in DATS.json')
dataset.name = dats.get(
'title',
os.path.basename(dataset.dataset_id)
)
db.session.merge(dataset)
db.session.commit()
print('[INFO ] ' + ds['gitmodule_name'] + ' updated.')
