def delete_entry(str_id):
# For return path
id = ObjectId(str_id)
sample = db.get_sample_by_id(id)
if sample is None:
set_error('Item to delete not found.')
return redirect('/')
dataset_id = sample['dataset_id']
readonly = db.is_readonly_dataset_id(dataset_id)
sample_index, sample_count, prev_sample_id, next_sample_id = db.get_sample_index(dataset_id, id)
if readonly:
set_error('Dataset is protected.')
elif db.delete_sample(id):
set_notice('Item deleted.')
else:
set_error('Could not delete item.')
# Redirect: To next sample in set if it exists. Otherwise to the dataset page.
if next_sample_id is not None and next_sample_id != id:
# Return to same kind of page (info/annotation/diff_annotation)
is_annotation = request.args.get('annotate')
is_differential = request.args.get('differential')
if is_annotation:
suffix = '?differential=1' if is_differential else ''
return redirect('/annotate/' + str(next_sample_id) + suffix)
else:
return redirect('/info/' + str(next_sample_id))
else:
return redirect('/dataset/' + str(dataset_id))
def upload_archive(dataset_id, file, _filename, is_new_dataset):
zipf = ZipFile(file.stream, 'r')
print 'ZIPF:'
for nfo in zipf.infolist():
if nfo.file_size > config.max_image_file_size:
return error_redirect(
'One of the images (%s) exceeds the max file size (%d).' %
(nfo.filename, config.max_image_file_size))
n_added = 0
for nfo in zipf.infolist():
# Check
image_filename = secure_filename(nfo.filename)
basename, ext = os.path.splitext(image_filename)
if ext not in config.image_extensions:
print 'Not an image: %s (%s)' % (ext, image_filename)
continue
# Extract
full_fn = make_unique_server_image_filename(image_filename)
zfile = zipf.open(nfo)
with open(full_fn, 'wb') as fid:
contents = zfile.read()
fid.write(contents)
print 'Written to ', full_fn
# Add
if add_uploaded_image(dataset_id, full_fn, image_filename) is not None:
n_added += 1
else:
# Cleanup invalid
if os.path.isfile(full_fn):
os.remove(full_fn)
set_notice('%d images added.' % n_added)
ds_url_suffix = '?new=true' if is_new_dataset else ''
return redirect('/dataset/%s' % str(dataset_id) + ds_url_suffix)
def enqueue_all_validation_sets():
n = 0
m = 0
for model in db.get_models(details=False, status=db.model_status_trained):
nm = enqueue_validation_sets_for_model(model['_id'])
if nm:
n += nm
m += 1
set_notice('Enqueued %d sets from %d models.' % (n, m))
return redirect('/admin')
def delete_dataset(dataset_id_str):
dataset_id = ObjectId(dataset_id_str)
dataset_info = db.get_dataset_by_id(dataset_id)
if dataset_info is None:
return render_template("404.html")
if db.is_readonly_dataset(dataset_info):
set_error('Dataset is protected.')
return redirect('/dataset/' + str(dataset_info['_id']))
db.delete_dataset(dataset_id)
set_notice('Dataset "%s" deleted.' % dataset_info['name'])
return redirect('/')
def enqueue_all_images():
n = 0
m = 0
for model in db.get_models(details=False, status=db.model_status_trained):
if not model['primary']:
nm = enqueue_images_for_model(model['_id'])
if nm:
n += nm
m += 1
set_notice('Enqueued %d images from %d models.' % (n, m))
return redirect('/admin')
def dataset_set_threshold(dataset_id_str, new_threshold_str):
dataset_id = ObjectId(dataset_id_str)
dataset_info = db.get_dataset_by_id(dataset_id)
new_threshold = float(new_threshold_str)
if dataset_info.get('threshold_prob') == new_threshold:
set_notice('Threshold not updated: Values are identical.')
else:
db.set_dataset_threshold_prob(dataset_id, new_threshold)
count = db.remove_machine_annotations_for_dataset(dataset_id)
set_notice('Threshold updated. %d annotations removed.' % count)
return redirect('/dataset/' + dataset_id_str)
def admin_retrain():
data = request.form
model_name = data['train_model_name']
# Race condition here; but it's just for the admin page anyway
if re.match('^[\w-]+$', model_name) is None:
set_error(
'Invalid model name. Must be non-empty, only alphanumeric characters, dashes and underscores.'
)
return redirect('/admin')
if db.get_model_by_name(model_name):
set_error('Model exists. Please pick a different name.')
return redirect('/admin')
# Validate tag
tag_name = data['train_label']
dsets = list(db.get_datasets_by_tag(tag_name))
if not len(dsets):
set_error('Train tag does not match any datasets.')
return redirect('/admin')
is_primary = (data.get('train_primary') == 'on')
dataset_only = (data.get('dataset_only') == 'on')
train_sample_limit_s = data['train_sample_limit']
if len(train_sample_limit_s):
try:
train_sample_limit = int(train_sample_limit_s)
if train_sample_limit <= 0:
raise ValueError()
except ValueError as e:
set_error('Invalid sample limit. Either leave empty for no limit, '
'or supply a valid number greater than zero.')
return redirect('/admin')
else:
train_sample_limit = None
# Add scheduled model training to DB. Will be picked up by worker.
rec = db.add_model(model_name=model_name,
margin=96,
sample_limit=train_sample_limit,
train_tag=tag_name,
scheduled_primary=is_primary,
status=db.model_status_scheduled,
dataset_only=dataset_only)
set_notice('Model training scheduled.')
return redirect('/model/' + str(rec['_id']))
def dataset_rerun(dataset_id_str):
dataset_id = ObjectId(dataset_id_str)
count = db.remove_machine_annotations_for_dataset(dataset_id)
set_notice('%d annotations removed.' % count)
return redirect('/dataset/' + dataset_id_str)
def delete_expired_datasets():
ds = find_old_datasets()
delete_datasets(ds)
set_notice('%d datasets deleted' % len(ds))
return redirect('/admin')
def enqueue_validation_sets(model_id_s):
model_id = ObjectId(model_id_s)
n = enqueue_validation_sets_for_model(model_id)
set_notice('Enqueued %d sets.' % n)
return redirect('/model/' + model_id_s)
def set_primary_model(model_id_s):
model_id = ObjectId(model_id_s)
db.set_primary_model(model_id)
set_notice('Updated primary model.')
return redirect('/model/' + model_id_s)
def enqueue_images(model_id_s):
model_id = ObjectId(model_id_s)
n = enqueue_images_for_model(model_id)
set_notice('Enqueued %d images.' % n)
return redirect('/model/' + model_id_s)
