def retrieve_job(job_id: str) -> None:
mturk = connect_mturk()
results = mturk.list_assignments_for_hit(HITId=job_id)
if results['NumResults'] > 0:
for assignment in results['Assignments']:
xml_doc = xmltodict.parse(assignment['Answer'])
logging.info("Worker's answer was:")
if isinstance(xml_doc['QuestionFormAnswers']['Answer'], list):
# Multiple fields in HIT layout
for answer_field in xml_doc['QuestionFormAnswers']['Answer']:
logging.info("For input field: " +
answer_field['QuestionIdentifier'])
logging.info("Submitted answer: " +
answer_field['FreeText'])
else:
# One field found in HIT layout
logging.info("For input field: " +
xml_doc['QuestionFormAnswers']['Answer']
['QuestionIdentifier'])
logging.info(
"Submitted answer: " +
xml_doc['QuestionFormAnswers']['Answer']['FreeText'])
else:
logging.info("No results ready yet")
def submit(allow_duplicate: bool = False,
name: Optional[List[str]] = None,
all_tasks: bool = False,
from_csv: Optional[str] = None):
if name == tuple() and not all_tasks:
raise ValueError("No task to submit")
client = aws.connect_mturk()
config = get_config()
tasks = config['tasks']
if name != tuple():
tasks = [task for task in tasks if task['name'] in name]
for task in tasks:
logging.info(f"Submitting task {task['name']}")
if from_csv is None:
generator_name = retrieve_generator(task['name'])
generator = generator_name(client)
else:
generator = csv_generator(client, from_csv)
for sample in generator:
if not allow_duplicate \
and not is_job_in_records(task['name'], sample):
create_job(client, task, sample)
def delete(all_hits: bool = False, hit_id: Optional[List[str]] = None):
client = aws.connect_mturk()
if all_hits:
aws.delete_all_hits(client)
elif hit_id is not None:
for hit in hit_id:
aws.delete_hit(client, hit)
else:
aws.delete_recorded_hits(client)
def progress(all_hits: bool = False,
all_recorded: bool = False,
hit_id: Optional[List[str]] = None):
client = aws.connect_mturk()
if all_hits:
aws.progress_all_hits(client)
elif all_recorded:
aws.progress_recorded_hits(client)
elif hit_id != tuple():
aws.progress_hits(client, hit_id)
else:
raise ValueError("No job to delete")
def bonus(from_csv: Optional[str] = None,
amount: float = 0.,
message: str = "",
output: str = ""):
client = aws.connect_mturk()
assignments = []
workers = []
done = []
if output != "":
with open(output, 'r', newline='') as fid:
reader = csv.reader(fid, delimiter=',')
for row in reader:
done.append(row[0])
if from_csv is not None:
with open(from_csv, 'r', newline='') as fid:
header = next(fid).split(',')
reader = csv.reader(fid, delimiter=',')
for row in reader:
row_dict = {k: v for k, v in zip(header, row)}
assignment_id = row_dict['assignment_id']
if assignment_id in done:
continue
if assignment_id not in assignments:
assignments.append(assignment_id)
workers.append(row_dict['worker_id'])
else:
raise ValueError("No job to reward")
for assignment, worker in zip(assignments, workers):
aws.send_bonus(client, assignment, worker, amount, message)
if output != "":
with open(output, 'a', newline='') as fid:
writer = csv.writer(fid, delimiter=',')
for assignment, worker in zip(assignments, workers):
writer.writerow(assignment, worker, amount, message)
response = client.create_hit_type(
Title=task['title'],
Description=task['description'],
Keywords=task['keywords'],
Reward=str(task['reward']),
AssignmentDurationInSeconds=task['assignment_duration'],
AutoApprovalDelayInSeconds=task['auto_approval_delay'])
return response
if __name__ == "__main__":
config = get_config()
tasks = config['tasks']
jobs = []
client = connect_mturk()
for task in tasks:
logging.info(f"Submitting task {task['name']}")
generator = retrieve_generator(task['name'])(client)
for sample in generator:
job = create_job(client, task, sample)
jobs.append(job)
logging.info("All tasks were successfully submitted.")
logging.info(f"Information is recorded in {config['job_filename']}.")