psiturk_scripting_example.py

from sqlalchemy import create_engine, MetaData, Table, text
import json
import pandas as pd
from boto.mturk.connection import MTurkConnection
import random
import sys
from psiturk.models import Participant
from psiturk.db import db_session, init_db
from psiturk.amt_services import MTurkServices
from psiturk.psiturk_config import PsiturkConfig

sandbox = False
if sandbox:
    host = 'mechanicalturk.sandbox.amazonaws.com'
else:
    host = 'mechanicalturk.amazonaws.com'
mtc = MTurkConnection(host=host)

db_url = "<redacted>" # or get it from psiturk config file...
table_name = 'turkdemo'
data_column_name = 'datastring'
# boilerplace sqlalchemy setup
engine = create_engine(db_url)
metadata = MetaData()
metadata.bind = engine
table = Table(table_name, metadata, autoload=True)

# instantiate psiturk stuff
config = PsiturkConfig()
config.load_config()
amt_services = MTurkServices(
    config.get('AWS Access', 'aws_access_key_id'), \
    config.get('AWS Access', 'aws_secret_access_key'),
    sandbox
    )

init_db()

#bonus reason
reason = "Thank you!"

# get all workers with status 'submitted'
workers = amt_services.get_workers(assignment_status='Submitted')

uniqueIds = [ worker['workerId'] + ':' + worker['assignmentId'] for worker in workers ]

# make a query and loop through
s = table.select()

#status codes of subjects who have pending mturk submissions
statuses = [3,4]
s = s.where(table.c.status.in_(statuses))
s = s.where(table.c.uniqueid.in_(uniqueIds))

# if you have workers you wish to exclude, add them here
exclude = []
if exclude:
    s = s.where(~table.c.uniqueid.in_(exclude))

hits_to_consider = []
if hits_to_consider:
    s = s.where(table.c.hitid.in_(hits_to_consider))

rows = s.execute()

data = []
for row in rows:
    data.append(row[data_column_name])

# Now we have all participant datastrings in a list.
# Let's make it a bit easier to work with:

# parse each participant's datastring as json object
# and take the 'data' sub-object

passed_assignment_ids = []
for part_json in data:
    part = json.loads(part_json)
    worker_id = part['workerId']
    assignment_id = part['assignmentId']
    accuracy_string = part['questiondata']['accuracy']
    accuracy = float(accuracy_string.rstrip('%')) / 100
    #accuracy = random.random() #hahahahahah that was almost really bad
    if accuracy < 0.5:
        # rejected!
        print "Rejecting {} because accuracy was {:.0%}".format(assignment_id, accuracy)
        feedback = "Sorry, your accuracy was below the threshold, suggesting that you \
        were not paying attention during the task."
        #mtc.reject_assignment(assignment_id, feedback=feedback) 
    else:
        print "Passing {} with accuracy of {:.0%}".format(assignment_id, accuracy)

        success = amt_services.approve_worker(assignment_id)

        if success:
            print 'approved', assignment_id
            try:
                part_db = Participant.query.\
                       filter(Participant.assignmentid == assignment_id).\
                       filter(Participant.endhit != None).\
                       one()
                amount = part_db.bonus
                status = part_db.status
                if status in [5,7]:
                    raise "this participant shouldn't have reached this point, they've already been credited or bonused...", assignment_id
                part_db.status = 5
                if amount <= 0:
                    print "bonus amount <=$0, no bonus given to", assignment_id
                elif status == 7:
                    print "bonus already awarded to ", assignment_id
                else:
                    success = amt_services.bonus_worker(assignment_id,
                                                             amount, reason)
                    if success:
                        print "gave bonus of $" + str(amount) + " to " + \
                        assignment_id
                        part_db.status = 7
                    else:
                        print "*** failed to bonus", assignment_id
                db_session.add(part_db)
                db_session.commit()
            except Exception as e:
                print e
                print "*** failed to bonus", assignment_id
        else:
            print '*** failed to approve', assignment_id