def load_from_democlub(csv_files, frozen_seats):
# Get list of existing refined issues in remote datastore, so can track what to delete
log("Getting list of refined issues")
refined_issues = RefinedIssue.all().filter("deleted =", False).fetch(100)
to_be_marked_deleted = {}
while refined_issues:
for refined_issue in refined_issues:
key_name = refined_issue.key().name()
log(" Marking before have refined issue key " + key_name)
to_be_marked_deleted[key_name] = refined_issue
refined_issues = RefinedIssue.all().filter("deleted =", False).filter('__key__ >', refined_issues[-1].key()).fetch(100)
# Load in CSV file and create/update all the issues
refined_issues_by_key = {}
for csv_file in csv_files:
log("Reading CSV file " + csv_file)
reader = csv.reader(open(csv_file, "rb"))
for row in reader:
if len(row) == 7:
row.append(None)
(democlub_id, question, reference_url, seat_name, created, updated, seat_slug, short_name) = row
key_name = democlub_id
# DemocracyClub has this constituency without its accent, YourNextMP has it with it.
seat_name = seat_name.replace("Ynys Mon", "Ynys Môn")
seat = find_democracyclub_seat_in_yournextmp(seat_name.decode('utf-8'), seat_slug.decode('utf-8'))
if seat.key().name() in frozen_seats:
log(" Frozen seat " + seat_name.decode('utf-8') + ", not storing issue")
else:
refined_issue = RefinedIssue(
democlub_id = int(democlub_id),
question = question.decode('utf-8'),
reference_url = reference_url.decode('utf-8'),
short_name = short_name and short_name.decode('utf-8') or None,
national = (seat.name == 'National'),
seat = seat,
created = convdate(created),
updated = convdate(updated),
key_name = key_name
)
log(" Storing local issue for " + seat_name + ": " + question)
refined_issues_by_key[key_name] = refined_issue
# record we still have this issue
if key_name in to_be_marked_deleted:
del to_be_marked_deleted[key_name]
log("Putting all refined issues")
put_in_batches(refined_issues_by_key.values())
# See which refined issues are left, i.e. are deleted
for key_name, refined_issue in to_be_marked_deleted.iteritems():
log(" Marking deleted issue for " + refined_issue.seat.name + ":" + refined_issue.question)
refined_issue.deleted = True
log("Putting marked deleted refined issues")
put_in_batches(to_be_marked_deleted.values())
def lookup_issues_by_id():
log("Getting all issues")
fs = RefinedIssue.all().fetch(100)
issues_by_id = {}
c = 0
while fs:
log(" getting batch from " + str(c))
for f in fs:
c = c + 1
issues_by_id[str(f.key())] = f
fs = RefinedIssue.all().filter('__key__ >', fs[-1].key()).fetch(100)
return issues_by_id
def load_from_democlub(csv_files, frozen_seats):
# Get list of existing refined issues in remote datastore, so can track what to delete
log("Getting list of refined issues")
refined_issues = RefinedIssue.all().filter("deleted =", False).fetch(100)
to_be_marked_deleted = {}
while refined_issues:
for refined_issue in refined_issues:
key_name = refined_issue.key().name()
log(" Marking before have refined issue key " + key_name)
to_be_marked_deleted[key_name] = refined_issue
refined_issues = RefinedIssue.all().filter("deleted =", False).filter('__key__ >', refined_issues[-1].key()).fetch(100)
# Load in CSV file and create/update all the issues
refined_issues_by_key = {}
for csv_file in csv_files:
log("Reading CSV file " + csv_file)
reader = csv.reader(open(csv_file, "rb"))
for row in reader:
(democlub_id, question, reference_url, seat_name, created, updated) = row
key_name = democlub_id
seat = find_seat(seat_name.decode('utf-8'))
if seat.key().name() in frozen_seats:
log(" Frozen seat " + seat_name + ", not storing issue: " + question)
else:
refined_issue = RefinedIssue(
democlub_id = int(democlub_id),
question = question.decode('utf-8'),
reference_url = reference_url.decode('utf-8'),
seat = seat,
created = convdate(created),
updated = convdate(updated),
key_name = key_name
)
log(" Storing local issue for " + seat_name + ": " + question)
refined_issues_by_key[key_name] = refined_issue
# record we still have this issue
if key_name in to_be_marked_deleted:
del to_be_marked_deleted[key_name]
log("Putting all refined issues")
put_in_batches(refined_issues_by_key.values())
# See which refined issues are left, i.e. are deleted
for key_name, refined_issue in to_be_marked_deleted.iteritems():
log(" Marking deleted issue for " + refined_issue.seat.name + ":" + refined_issue.question)
refined_issue.deleted = True
log("Putting marked deleted refined issues")
put_in_batches(to_be_marked_deleted.values())
def task_average_response_by_party(request, party_key_name, refined_issue_key_name):
party = Party.get_by_key_name(party_key_name)
refined_issue = RefinedIssue.get_by_key_name(refined_issue_key_name)
arbp = db.Query(AverageResponseByParty).filter('party =', party).filter('refined_issue =', refined_issue).get()
if not arbp:
arbp = AverageResponseByParty(party = party, refined_issue = refined_issue,
average_agreement = None, processing_running_total = 0,
processing_running_count = 0, processing_next_key = None)
chunk = db.Query(Candidacy).filter('deleted = ', False).filter('survey_filled_in =', True)
# carry on calculation where we left off
if arbp.processing_last_candidacy == None:
assert arbp.processing_running_total == 0
assert arbp.processing_running_count == 0
else:
chunk = chunk.filter('__key__ >', arbp.processing_last_candidacy.key())
# do 100 candidacies at a time, as too slow otherwise
chunk.fetch(10) # XXX
candidacy = None
for candidacy in chunk:
survey_response = db.Query(SurveyResponse).filter('candidacy =', candidacy).filter('refined_issue =', refined_issue).get()
if survey_response:
arbp.processing_running_total += survey_response.agreement
arbp.processing_running_count += 1
arbp.processing_last_candidacy = candidacy
# if we've finished, work out average
if candidacy == None:
arbp.average_agreement = float(arbp.processing_running_total) / float(arbp.processing_running_count)
arbp.put()
# calculate next chunk
if candidacy == None:
return HttpResponse("Calculation complete for " + party.name + " question: " + refined_issue.question)
else:
taskqueue.Queue('average-calc').add(taskqueue.Task(url='/task/average_response_by_party/' + party_key_name + "/" + refined_issue_key_name))
return HttpResponse("Done " + str(arbp.processing_running_count) + ", queued next chunk for " + party.name + " question: " + question)
def load_from_democlub(csv_files, frozen_seats):
# Get list of existing refined issues in remote datastore, so can track what to delete
log("Getting list of refined issues")
refined_issues = RefinedIssue.all().filter("deleted =", False).fetch(100)
to_be_marked_deleted = {}
while refined_issues:
for refined_issue in refined_issues:
key_name = refined_issue.key().name()
log(" Marking before have refined issue key " + key_name)
to_be_marked_deleted[key_name] = refined_issue
refined_issues = RefinedIssue.all().filter("deleted =", False).filter(
'__key__ >', refined_issues[-1].key()).fetch(100)
# Load in CSV file and create/update all the issues
refined_issues_by_key = {}
for csv_file in csv_files:
log("Reading CSV file " + csv_file)
reader = csv.reader(open(csv_file, "rb"))
for row in reader:
if len(row) == 7:
row.append(None)
(democlub_id, question, reference_url, seat_name, created, updated,
seat_slug, short_name) = row
key_name = democlub_id
# DemocracyClub has this constituency without its accent, YourNextMP has it with it.
seat_name = seat_name.replace("Ynys Mon", "Ynys Môn")
seat = find_democracyclub_seat_in_yournextmp(
seat_name.decode('utf-8'), seat_slug.decode('utf-8'))
if seat.key().name() in frozen_seats:
log(" Frozen seat " + seat_name.decode('utf-8') +
", not storing issue")
else:
refined_issue = RefinedIssue(
democlub_id=int(democlub_id),
question=question.decode('utf-8'),
reference_url=reference_url.decode('utf-8'),
short_name=short_name and short_name.decode('utf-8')
or None,
national=(seat.name == 'National'),
seat=seat,
created=convdate(created),
updated=convdate(updated),
key_name=key_name)
log(" Storing local issue for " + seat_name + ": " + question)
refined_issues_by_key[key_name] = refined_issue
# record we still have this issue
if key_name in to_be_marked_deleted:
del to_be_marked_deleted[key_name]
log("Putting all refined issues")
put_in_batches(refined_issues_by_key.values())
# See which refined issues are left, i.e. are deleted
for key_name, refined_issue in to_be_marked_deleted.iteritems():
log(" Marking deleted issue for " + refined_issue.seat.name + ":" +
refined_issue.question)
refined_issue.deleted = True
log("Putting marked deleted refined issues")
put_in_batches(to_be_marked_deleted.values())
