def initFromFile(self, labels_filename):
filename = dir_tools.getDatasetDirectory(self.project, self.dataset)
filename += 'labels/' + labels_filename
if not dir_tools.checkFileExists(filename):
raise ValueError('The labels file %s does not exist.' % filename)
## Check whether the file contains families
families = False
with open(filename, 'r') as f:
header = f.readline()
fields = header.split(',')
if len(fields) == 3:
families = True
query = 'LOAD DATA LOCAL INFILE \'' + filename + '\' '
query += 'INTO TABLE ' + 'Labels' + ' '
query += 'FIELDS TERMINATED BY \',\' '
query += 'IGNORE 1 LINES '
if families:
query += '(instance_id, label, family) '
else:
query += '(instance_id, label) '
query += 'SET experiment_label_id = ' + str(
self.experiment_label_id) + ', '
if not families:
query += 'family = "other",'
query += 'iteration = 0, '
query += 'method = "init", '
query += 'annotation = "0"'
query += ';'
self.cursor.execute(query)
self.db.commit()
self.checkLabelsValidity()
def runNextIteration(experiment_id, iteration_number):
res = str(celeryRunNextIteration.s().apply_async())
if user_exp:
experiment = updateCurrentExperiment(experiment_id)
filename = experiment.getOutputDirectory()
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [datetime.datetime.now(), 'nextIteration', iteration_number]
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return res
def loadTrueLabels(self):
labels_file = dir_tools.getDatasetDirectory(self.project, self.dataset)
labels_file += 'labels/true_labels.csv'
# Loads the true labels in the table TrueLabels if the file exists
# Otherwise the table TrueLabels is not created
if not dir_tools.checkFileExists(labels_file):
print >> sys.stderr, 'No ground truth labels for this dataset'
return
exp = Experiment(self.project,
self.dataset,
self.db,
self.cursor,
experiment_name='true_labels')
exp.initLabels('true_labels.csv')
def currentAnnotations(experiment_id, iteration):
experiment = updateCurrentExperiment(experiment_id)
page = render_template('ActiveLearning/current_annotations.html',
project=experiment.project)
if user_exp:
filename = experiment.getOutputDirectory()
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [datetime.datetime.now(), 'displayAnnotatedInstances']
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return page
def removeLabel(experiment_id, inst_experiment_label_id, iteration_number,
instance_id):
labels_tools.removeLabel(session, inst_experiment_label_id, instance_id)
if user_exp:
experiment = updateCurrentExperiment(experiment_id)
filename = experiment.getOutputDirectory()
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [datetime.datetime.now(), 'removeLabel', instance_id]
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return ''
def runNextIteration(project, dataset, experiment_id, iteration_number):
res = str(celeryRunNextIteration.s().apply_async())
if user_exp:
experiment = ExperimentFactory.getFactory().fromJson(
project, dataset, experiment_id, db, cursor)
filename = dir_tools.getExperimentOutputDirectory(experiment)
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [datetime.datetime.now(), 'nextIteration', iteration_number]
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return res
def currentAnnotations(project, dataset, experiment_id, iteration):
page = render_template('ActiveLearning/current_annotations.html',
project=project)
if user_exp:
experiment = ExperimentFactory.getFactory().fromJson(
project, dataset, experiment_id, db, cursor)
filename = dir_tools.getExperimentOutputDirectory(experiment)
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [datetime.datetime.now(), 'displayAnnotatedInstances']
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return page
def changeFamilyLabel(experiment_id, label, family):
experiment = updateCurrentExperiment(experiment_id)
experiment_label_id = experiment.labels_id
labels_tools.changeFamilyLabel(session, label, family, experiment_label_id)
if user_exp:
filename = experiment.getOutputDirectory()
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [
datetime.datetime.now(), 'changeFamilyLabel', family, label
]
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return ''
def mergeFamilies(experiment_id, label, families, new_family_name):
experiment = updateCurrentExperiment(experiment_id)
experiment_label_id = experiment.labels_id
families = families.split(',')
labels_tools.mergeFamilies(session, label, families, new_family_name,
experiment_label_id)
if user_exp:
filename = experiment.getOutputDirectory()
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [datetime.datetime.now(), 'mergeFamilies', new_family_name]
to_print += map(str, families)
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return ''
def changeFamilyLabel(project, dataset, experiment_id, experiment_label_id,
label, family):
mysql_tools.useDatabase(cursor, project, dataset)
labels_tools.changeFamilyLabel(cursor, label, family, experiment_label_id)
if user_exp:
experiment = ExperimentFactory.getFactory().fromJson(
project, dataset, experiment_id, db, cursor)
filename = dir_tools.getExperimentOutputDirectory(experiment)
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [
datetime.datetime.now(), 'changeFamilyLabel', family, label
]
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return ''
def mergeFamilies(project, dataset, experiment_id, experiment_label_id, label,
families, new_family_name):
families = families.split(',')
mysql_tools.useDatabase(cursor, project, dataset)
labels_tools.mergeFamilies(cursor, label, families, new_family_name,
experiment_label_id)
if user_exp:
experiment = ExperimentFactory.getFactory().fromJson(
project, dataset, experiment_id, db, cursor)
filename = dir_tools.getExperimentOutputDirectory(experiment)
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [datetime.datetime.now(), 'mergeFamilies', new_family_name]
to_print += map(str, families)
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return ''
def addLabel(experiment_id, inst_experiment_label_id, iteration_number,
instance_id, label, family, method, annotation):
annotation = annotation == 'true'
labels_tools.addLabel(session, inst_experiment_label_id, instance_id,
label, family, iteration_number, method, annotation)
if user_exp:
experiment = updateCurrentExperiment(experiment_id)
filename = experiment.getOutputDirectory()
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [
datetime.datetime.now(), 'addLabel', iteration_number, instance_id,
label, family, method, annotation
]
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return ''
def removeLabel(project, dataset, experiment_id, inst_dataset,
inst_experiment_label_id, iteration_number, instance_id):
mysql_tools.useDatabase(cursor, project, inst_dataset)
labels_tools.removeLabel(cursor, inst_experiment_label_id, instance_id)
db.commit()
if user_exp:
experiment = ExperimentFactory.getFactory().fromJson(
project, dataset, experiment_id, db, cursor)
filename = dir_tools.getExperimentOutputDirectory(experiment)
filename += 'user_actions.log'
file_exists = dir_tools.checkFileExists(filename)
mode = 'a' if file_exists else 'w'
to_print = [
datetime.datetime.now(), 'removeLabel', project, dataset,
instance_id
]
to_print = map(str, to_print)
to_print = ','.join(to_print)
with open(filename, mode) as f:
print >> f, to_print
return ''
def loadTrueLabels(self):
labels_file = dir_tools.getDatasetDirectory(self.project, self.dataset)
labels_file += 'labels/true_labels.csv'
# Loads the true labels in the table TrueLabels if the file exists
# Otherwise the table TrueLabels is not created
if not dir_tools.checkFileExists(labels_file):
print >> sys.stderr, 'No ground truth labels for this dataset'
return
## Check whether the file contains families
families = False
with open(labels_file, 'r') as f:
header = f.readline()
fields = header.split(',')
if len(fields) == 3:
families = True
db, cursor = db_tools.getRawConnection()
if db_tools.isMysql():
query = 'CREATE TEMPORARY TABLE true_labels_import('
query += 'user_instance_id integer PRIMARY KEY, '
query += 'label varchar(200), '
query += 'family varchar(200) DEFAULT \'other\', '
query += 'dataset_id integer DEFAULT ' + str(
self.dataset_id) + ', '
query += 'id integer DEFAULT NULL'
query += ');'
cursor.execute(query)
query = 'LOAD DATA LOCAL INFILE \'' + labels_file + '\' '
query += 'INTO TABLE ' + 'true_labels_import' + ' '
query += 'FIELDS TERMINATED BY \',\' '
query += 'IGNORE 1 LINES '
if families:
query += '(user_instance_id, label, family) '
else:
query += '(user_instance_id, label) '
query += ';'
cursor.execute(query)
query = 'UPDATE true_labels_import t '
query += 'JOIN instances i '
query += 'ON i.user_instance_id = t.user_instance_id '
query += 'AND i.dataset_id = t.dataset_id '
query += 'SET t.id = i.id;'
cursor.execute(query)
query = 'INSERT INTO true_labels(instance_id, dataset_id, label, family) '
query += 'SELECT t.id, t.dataset_id, t.label, t.family '
query += 'FROM true_labels_import AS t;'
cursor.execute(query)
elif db_tools.isPostgresql():
query = 'CREATE TEMPORARY TABLE true_labels_import('
query += 'user_instance_id integer PRIMARY KEY, '
query += 'label true_labels_enum, '
query += 'family varchar(200) DEFAULT \'other\', '
query += 'dataset_id integer DEFAULT ' + str(
self.dataset_id) + ', '
query += 'id integer DEFAULT NULL'
query += ');'
cursor.execute(query)
with open(labels_file, 'r') as f:
if families:
query = 'COPY true_labels_import(user_instance_id,label,family) '
else:
query = 'COPY true_labels_import(user_instance_id,label) '
query += 'FROM STDIN '
query += 'WITH CSV HEADER DELIMITER AS \',\' ;'
cursor.copy_expert(sql=query, file=f)
query = 'UPDATE true_labels_import AS t '
query += 'SET id = i.id '
query += 'FROM instances AS i '
query += 'WHERE i.user_instance_id = t.user_instance_id '
query += 'AND i.dataset_id = t.dataset_id;'
cursor.execute(query)
query = 'INSERT INTO true_labels(instance_id, dataset_id, label, family) '
query += 'SELECT t.id, t.dataset_id, t.label, t.family '
query += 'FROM true_labels_import AS t;'
cursor.execute(query)
db_tools.closeRawConnection(db, cursor)
def initFromFile(self, labels_filename):
if labels_filename is None:
labels_type = 'none'
elif labels_filename == 'true_labels.csv':
labels_type = 'true_labels'
else:
labels_type = 'partial_labels'
exp_labels = db_tables.ExperimentsLabelsAlchemy(
experiment_id=self.experiment_id, labels_type=labels_type)
self.session.add(exp_labels)
self.session.commit()
self.labels_id = exp_labels.labels_id
self.labels_type = labels_type
if labels_type == 'partial_labels':
filename = dir_tools.getDatasetDirectory(self.project,
self.dataset)
filename += 'labels/' + labels_filename
if not dir_tools.checkFileExists(filename):
raise ValueError('The labels file %s does not exist.' %
filename)
## Check whether the file contains families
families = False
with open(filename, 'r') as f:
header = f.readline()
fields = header.split(',')
if len(fields) == 3:
families = True
db, cursor = db_tools.getRawConnection()
if db_tools.isMysql():
query = 'CREATE TEMPORARY TABLE labels_import('
query += 'instance_id integer, '
query += 'labels_id integer DEFAULT ' + str(
self.labels_id) + ', '
query += 'user_instance_id integer, '
query += 'label varchar(200), '
query += 'family varchar(200) DEFAULT \'other\', '
query += 'iteration integer DEFAULT 0, '
query += 'method varchar(200) DEFAULT \'init\', '
query += 'annotation boolean DEFAULT True'
query += ');'
cursor.execute(query)
query = 'LOAD DATA LOCAL INFILE \'' + filename + '\' '
query += 'INTO TABLE ' + 'labels_import' + ' '
query += 'FIELDS TERMINATED BY \',\' '
query += 'IGNORE 1 LINES '
if families:
query += '(user_instance_id, label, family) '
else:
query += '(user_instance_id, label) '
query += ';'
cursor.execute(query)
query = 'UPDATE labels_import l '
query += 'JOIN instances i '
query += 'ON i.user_instance_id = l.user_instance_id '
query += 'AND i.dataset_id = ' + str(self.dataset_id) + ' '
query += 'SET l.instance_id = i.id;'
cursor.execute(query)
query = 'INSERT INTO labels(instance_id,labels_id,label,family,iteration,method,annotation) '
query += 'SELECT instance_id,labels_id,label,family,iteration,method,annotation '
query += 'FROM labels_import;'
cursor.execute(query)
elif db_tools.isPostgresql():
query = 'CREATE TEMPORARY TABLE labels_import('
query += 'instance_id integer, '
query += 'labels_id integer DEFAULT ' + str(
self.labels_id) + ', '
query += 'user_instance_id integer, '
query += 'label labels_enum, '
query += 'family varchar(200) DEFAULT \'other\', '
query += 'iteration integer DEFAULT 0, '
query += 'method varchar(200) DEFAULT \'init\', '
query += 'annotation boolean DEFAULT True'
query += ');'
cursor.execute(query)
with open(filename, 'r') as f:
if families:
query = 'COPY labels_import(user_instance_id,label,family) '
else:
query = 'COPY labels_import(user_instance_id,label) '
query += 'FROM STDIN '
query += 'WITH CSV HEADER DELIMITER AS \',\' ;'
cursor.copy_expert(sql=query, file=f)
query = 'UPDATE labels_import AS l '
query += 'SET instance_id = i.id '
query += 'FROM instances AS i '
query += 'WHERE i.user_instance_id = l.user_instance_id '
query += 'AND i.dataset_id = ' + str(self.dataset_id) + ';'
cursor.execute(query)
query = 'INSERT INTO labels(instance_id,labels_id,label,family,iteration,method,annotation) '
query += 'SELECT instance_id,labels_id,label,family,iteration,method,annotation '
query += 'FROM labels_import;'
cursor.execute(query)
db_tools.closeRawConnection(db, cursor)
self.session.commit()