def populate_uid(name, force=False, api=None):
"""
For a TwitterUser, populate its uid based on its stored screen name,
if uid==0 (default value, indicating it hasn't been set yet).
if force==True, do it even if uid isn't 0
Only do this for active users.
see https://dev.twitter.com/docs/api/1.1/get/users/lookup
for explanation of get_user call
see https://dev.twitter.com/docs/working-with-timelines
for explanation of max_id, since_id usage
see also:
https://dev.twitter.com/docs/error-codes-responses
https://dev.twitter.com/docs/rate-limiting
"""
if api is None:
api = authenticated_api(username=settings.TWITTER_DEFAULT_USERNAME)
qs_tweeps = TwitterUser.objects.filter(is_active=True, name=name)
for tweep in qs_tweeps:
if tweep.uid == 0 or force is True:
try:
user_status = api.get_user(screen_name=name)
tweep.uid = user_status['id']
tweep.save()
print 'updated user \'%s\' uid to %d' % (name, tweep.uid)
except tweepy.error.TweepError as e:
print 'Failed to find user \'%s\'. Error: %s' % (name, e)
finally:
time.sleep(set_wait_time(api.last_response))
def handle(self, *args, **options):
api = authenticated_api(username=settings.TWITTER_DEFAULT_USERNAME)
qs_tweeps = TwitterUser.objects.filter(is_active=True)
if options.get('user', None):
qs_tweeps = qs_tweeps.filter(name=options.get('user'))
for tweep in qs_tweeps:
print 'user: %s' % tweep.name
# check user status, update twitter user name if it has changed
if tweep.uid == 0:
print 'uid has not been set yet - skipping.'
continue
try:
user_status = api.get_user(id=tweep.uid)
if user_status['screen_name'] != tweep.name:
print ' -- updating screen name to %s' % \
user_status['screen_name']
former_names = tweep.former_names
if not tweep.former_names:
former_names = '{}'
oldnames = json.loads(former_names)
oldnames[datetime.datetime.now().strftime('%c')] = \
tweep.name
tweep.former_names = json.dumps(oldnames)
tweep.name = user_status['screen_name']
#TODO: Is this save unnecessary, since it gets saved below?
tweep.save()
except tweepy.error.TweepError as e:
print 'Error: %s' % e
#go to the next tweep in the for loop
continue
finally:
time.sleep(set_wait_time(api.last_response))
def handle(self, *args, **options):
api = authenticated_api(username=settings.TWITTER_DEFAULT_USERNAME)
qs_tweeps = TwitterUser.objects.filter(is_active=True)
if options.get('user', None):
qs_tweeps = qs_tweeps.filter(name=options.get('user'))
for tweep in qs_tweeps:
print 'user: %s' % tweep.name
# check user status, update twitter user name if it has changed
if tweep.uid == 0:
print 'uid has not been set yet - skipping.'
continue
try:
user_status = api.get_user(id=tweep.uid)
if user_status['screen_name'] != tweep.name:
print ' -- updating screen name to %s' % \
user_status['screen_name']
former_names = tweep.former_names
if not tweep.former_names:
former_names = '{}'
oldnames = json.loads(former_names)
oldnames[datetime.datetime.utcnow().strftime(
'%Y-%m-%dT%H:%M:%SZ')] = tweep.name
tweep.former_names = json.dumps(oldnames)
tweep.name = user_status['screen_name']
#TODO: Is this save unnecessary, since it gets saved below?
tweep.save()
except tweepy.error.TweepError as e:
print 'Error: %s' % e
#go to the next tweep in the for loop
continue
finally:
time.sleep(set_wait_time(api.last_response))
def handle(self, *args, **options):
api = authenticated_api(username=settings.TWITTER_DEFAULT_USERNAME)
qs_tweeps = TwitterUser.objects.filter(is_active=True)
if options.get('user', None):
qs_tweeps = qs_tweeps.filter(name=options.get('user'))
paginator = Paginator(qs_tweeps, 100)
page_count = paginator.num_pages
for page_counter in range(1, page_count + 1):
print "Page %s of %s" % (page_counter, page_count)
qs_page = paginator.page(page_counter)
tweep_map = {}
for tweep in qs_page:
# check user status, update twitter user name if it has changed
if tweep.uid == 0:
print 'user: %s' % tweep.name
print ' -- uid has not been set yet - skipping.'
continue
else:
tweep_map[tweep.uid] = tweep
if tweep_map:
try:
user_statuses = api.lookup_users(user_ids=tweep_map.keys())
for user_status in user_statuses:
tweep = tweep_map[user_status['id']]
print 'user: %s' % tweep.name
if user_status['screen_name'] != tweep.name:
print ' -- updating screen name to %s' % \
user_status['screen_name']
former_names = tweep.former_names
if not tweep.former_names:
former_names = '{}'
oldnames = json.loads(former_names)
oldnames[datetime.datetime.utcnow().strftime(
'%Y-%m-%dT%H:%M:%SZ')] = tweep.name
tweep.former_names = json.dumps(oldnames)
tweep.name = user_status['screen_name']
#TODO: Is this save unnecessary, since it gets saved below?
tweep.save()
except tweepy.error.TweepError as e:
print 'Error: %s' % e
#go to the next tweep in the for loop
continue
finally:
time.sleep(set_wait_time(api.last_response))
def handle(self, *args, **options):
api = authenticated_api(username=settings.TWITTER_DEFAULT_USERNAME)
job = TwitterUserTimelineJob()
job.save()
qs_tweeps = TwitterUser.objects.filter(is_active=True)
if options.get('user', None):
qs_tweeps = qs_tweeps.filter(name=options.get('user'))
else:
# NOTE: randomizing here might be healthier when considering
# possibility of multiple parallel jobs running and competing
# for api calls but this is an instinctual call, not data-driven
qs_tweeps = qs_tweeps.order_by('?')
for tweep in qs_tweeps:
print 'user: %s' % tweep.name
# can't do this unless we have a twitter user_id stored
if tweep.uid == 0:
skipmsg = 'uid has not been set yet - skipping this ' + \
'user. May need to run populate_uids if this ' + \
'is an old database.'
print skipmsg
error = TwitterUserTimelineError(job=job, user=tweep,
error=skipmsg)
error.save()
continue
# now move on to determining first tweet id to get
since_id = 1
# set since_id if they have any statuses recorded
if tweep.items.count() > 0:
max_dict = tweep.items.all().aggregate(Max('twitter_id'))
since_id = max_dict['twitter_id__max']
max_id = 0
# update their record (auto_now) as we're checking it now
tweep.save()
while True:
stop = False
try:
print 'since: %s' % (since_id)
if max_id:
print 'max: %s' % max_id
timeline = api.user_timeline(id=tweep.uid,
since_id=since_id,
max_id=max_id, count=200)
else:
timeline = api.user_timeline(id=tweep.uid,
since_id=since_id,
count=200)
except tweepy.error.TweepError as e:
print 'ERROR: %s' % e
error = TwitterUserTimelineError(job=job, user=tweep,
error=e)
error.save()
timeline = []
if len(timeline) == 0:
# Nothing new; stop for this user
stop = True
new_status_count = 0
for status in timeline:
# eg 'Mon Oct 15 20:15:12 +0000 2012'
dt_aware = dt_aware_from_created_at(status['created_at'])
try:
item, created = TwitterUserItem.objects.get_or_create(
twitter_user=tweep,
twitter_id=status['id'],
date_published=dt_aware,
item_text=status['text'],
item_json=json.dumps(status),
place=status['place'] or '',
source=status['source'])
if created:
max_id = item.twitter_id - 1
new_status_count += 1
else:
print 'skip: id %s' % item.id
except IntegrityError as ie:
print 'ERROR: %s' % ie
error = TwitterUserTimelineError(job=job, user=tweep,
error=ie)
error.save()
print 'saved: %s item(s)' % new_status_count
job.num_added += new_status_count
# max new statuses per call is 200, so check for less than
# a reasonable fraction of that to see if we should stop
if new_status_count < 150:
print 'stop: < 150 new statuses'
stop = True
if max_id < since_id:
# Got 'em all, stop for this user
print 'stop: max_id < since_id'
stop = True
# Check response codes for issues
response_status = api.last_response.status
if response_status >= 400:
print 'error:', api.last_response.getheader('status')
error = TwitterUserTimelineError(job=job, user=tweep,
error=e)
error.save()
stop = True
job.save()
# wait before next call no matter what
time.sleep(set_wait_time(api.last_response))
if stop:
break
def handle(self, *args, **options):
api = authenticated_api(username=settings.TWITTER_DEFAULT_USERNAME)
qs_tweeps = TwitterUser.objects.filter(is_active=True)
if options.get('user', None):
qs_tweeps = qs_tweeps.filter(name=options.get('user'))
qs_tweeps = qs_tweeps.order_by('date_last_checked')
for tweep in qs_tweeps:
print 'user: %s' % tweep.name
# can't do this unless we have a twitter user_id stored
if tweep.uid == 0:
print 'uid has not been set yet - skipping this user. ' + \
'May need to run populate_uids if this is an old ' + \
'database.'
continue
# now move on to determining first tweet id to get
since_id = 1
# set since_id if they have any statuses recorded
if tweep.items.count() > 0:
max_dict = tweep.items.all().aggregate(Max('twitter_id'))
since_id = max_dict['twitter_id__max']
max_id = 0
# update their record (auto_now) as we're checking it now
tweep.save()
while True:
stop = False
try:
print 'since: %s' % (since_id)
if max_id:
print 'max: %s' % max_id
timeline = api.user_timeline(id=tweep.uid,
since_id=since_id,
max_id=max_id,
count=200)
else:
timeline = api.user_timeline(id=tweep.uid,
since_id=since_id,
count=200)
except tweepy.error.TweepError as e:
print 'ERROR: %s' % e
timeline = []
if len(timeline) == 0:
# Nothing new; stop for this user
stop = True
new_status_count = 0
for status in timeline:
# eg 'Mon Oct 15 20:15:12 +0000 2012'
dt_aware = dt_aware_from_created_at(status['created_at'])
try:
item, created = TwitterUserItem.objects.get_or_create(
twitter_user=tweep,
twitter_id=status['id'],
date_published=dt_aware,
item_text=status['text'],
item_json=json.dumps(status),
place=status['place'] or '',
source=status['source'])
if created:
max_id = item.twitter_id - 1
new_status_count += 1
else:
print 'skip: id %s' % item.id
except IntegrityError as ie:
print 'ERROR: %s' % ie
print 'saved: %s item(s)' % new_status_count
# max new statuses per call is 200, so check for less than
# a reasonable fraction of that to see if we should stop
if new_status_count < 150:
print 'stop: < 150 new statuses'
stop = True
if max_id < since_id:
# Got 'em all, stop for this user
print 'stop: max_id < since_id'
stop = True
# Check response codes for issues
response_status = api.last_response.status
if response_status >= 400:
print 'error:', api.last_response.getheader('status')
stop = True
# wait before next call no matter what
time.sleep(set_wait_time(api.last_response))
if stop:
break
def handle(self, *args, **options):
api = authenticated_api(username=settings.TWITTER_DEFAULT_USERNAME)
job = TwitterUserTimelineJob()
job.save()
qs_tweeps = TwitterUser.objects.filter(is_active=True)
if options.get('user', None):
qs_tweeps = qs_tweeps.filter(name=options.get('user'))
else:
# NOTE: randomizing here might be healthier when considering
# possibility of multiple parallel jobs running and competing
# for api calls but this is an instinctual call, not data-driven
qs_tweeps = qs_tweeps.order_by('?')
for tweep in qs_tweeps:
print 'user: %s' % tweep.name
# can't do this unless we have a twitter user_id stored
if tweep.uid == 0:
skipmsg = 'uid has not been set yet - skipping this ' + \
'user. May need to run populate_uids if this ' + \
'is an old database.'
print skipmsg
error = TwitterUserTimelineError(job=job,
user=tweep,
error=skipmsg)
error.save()
continue
# now move on to determining first tweet id to get
since_id = 1
# set since_id if they have any statuses recorded
if tweep.items.count() > 0:
max_dict = tweep.items.all().aggregate(Max('twitter_id'))
since_id = max_dict['twitter_id__max']
max_id = 0
# update their record (auto_now) as we're checking it now
tweep.save()
while True:
# wait before next call no matter what;
# use getattr() because api might be None the first time or
# after errors
time.sleep(set_wait_time(getattr(api, 'last_response', None)))
job.save()
stop = False
try:
print 'since: %s' % (since_id)
if max_id:
print 'max: %s' % max_id
timeline = api.user_timeline(id=tweep.uid,
since_id=since_id,
max_id=max_id,
count=200)
else:
timeline = api.user_timeline(id=tweep.uid,
since_id=since_id,
count=200)
except tweepy.error.TweepError as e:
print 'ERROR: %s' % e
error = TwitterUserTimelineError(job=job,
user=tweep,
error=e)
error.save()
timeline = []
break
if len(timeline) == 0:
# Nothing new; stop for this user
stop = True
new_status_count = 0
for status in timeline:
# eg 'Mon Oct 15 20:15:12 +0000 2012'
dt_aware = dt_aware_from_created_at(status['created_at'])
try:
item, created = TwitterUserItem.objects.get_or_create(
twitter_user=tweep,
twitter_id=status['id'],
date_published=dt_aware,
item_text=status['text'],
item_json=json.dumps(status),
place=status['place'] or '',
source=status['source'])
if created:
max_id = item.twitter_id - 1
new_status_count += 1
else:
print 'skip: id %s' % item.id
except IntegrityError as ie:
print 'ERROR: %s' % ie
error = TwitterUserTimelineError(job=job,
user=tweep,
error=ie)
error.save()
print 'saved: %s item(s)' % new_status_count
job.num_added += new_status_count
# max new statuses per call is 200, so check for less than
# a reasonable fraction of that to see if we should stop
if new_status_count < 150:
print 'stop: < 150 new statuses'
stop = True
if max_id < since_id:
# Got 'em all, stop for this user
print 'stop: max_id < since_id'
stop = True
# Check response codes for issues
response_status = api.last_response.status_code
if response_status >= 400:
print 'error:', api.last_response.getheader('status')
error = TwitterUserTimelineError(job=job,
user=tweep,
error=e)
error.save()
stop = True
if stop:
break
