106 lines
4.0 KiB
Python
106 lines
4.0 KiB
Python
|
|
import os, sys
|
|||
|
|
from datetime import datetime
|
|||
|
|
from dateutil import relativedelta
|
|||
|
|
|
|||
|
|
import django
|
|||
|
|
sys.path.append('../../../')
|
|||
|
|
os.environ['DJANGO_SETTINGS_MODULE'] = 'ds_events.settings'
|
|||
|
|
django.setup()
|
|||
|
|
|
|||
|
|
from time import sleep
|
|||
|
|
from pprint import pprint as ppr
|
|||
|
|
import pytz
|
|||
|
|
|
|||
|
|
from events.models import Organization, Scraper, Event as DSEvent
|
|||
|
|
import events.digitools as digitools
|
|||
|
|
|
|||
|
|
|
|||
|
|
try:
|
|||
|
|
venue, created = Organization.objects.get_or_create(
|
|||
|
|
name="Parkway Theater",
|
|||
|
|
city="Minneapolis",
|
|||
|
|
website="https://theparkwaytheater.com",
|
|||
|
|
is_venue = True
|
|||
|
|
)
|
|||
|
|
except Exception as e:
|
|||
|
|
venue = Organization.objects.get(name="Parkway Theater")
|
|||
|
|
|
|||
|
|
scraper,item_count_start = digitools.getScraper(venue)
|
|||
|
|
|
|||
|
|
tz = pytz.timezone("US/Central")
|
|||
|
|
|
|||
|
|
DATETIME_FORMAT = '%b %d, %Y %I:%M %p'
|
|||
|
|
|
|||
|
|
def get_events(ps, event_type):
|
|||
|
|
contents = ps.xpath('.//*/div[@class="summary-content sqs-gallery-meta-container"]')
|
|||
|
|
img_etc = ps.xpath('.//*/div[@class="summary-thumbnail-outer-container"]/a/div/img/@src')
|
|||
|
|
ps.xpath('.//*/span[@class="event-time-12hr"]/text()')
|
|||
|
|
for c,i in zip(contents,img_etc):
|
|||
|
|
try:
|
|||
|
|
event = {}
|
|||
|
|
event['calendar'] = scraper.calendar
|
|||
|
|
event['title'] = c.xpath('.//*/a[@class="summary-title-link"]/text()')[0]
|
|||
|
|
event['link'] = "https://theparkwaytheater.com" + c.xpath('.//*/a[@class="summary-title-link"]/@href')[0]
|
|||
|
|
event['date'] = c.xpath('.//div/div/time/text()')[0] + " " + c.xpath('.//*/span[@class="event-time-12hr"]/text()')[0].split("–")[0].strip()
|
|||
|
|
event['dateStamp'] = datetime.strptime(event['date'], DATETIME_FORMAT)
|
|||
|
|
event['desc'] = c.xpath('.//*/p/text()')[0]
|
|||
|
|
event['img_link'] = i
|
|||
|
|
event['details'] = {
|
|||
|
|
'description': event['desc'],
|
|||
|
|
'img_link': event['img_link'],
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
try:
|
|||
|
|
new_event = DSEvent.objects.update_or_create(
|
|||
|
|
calendar = scraper.calendar,
|
|||
|
|
scraper = scraper,
|
|||
|
|
event_type = event_type,
|
|||
|
|
show_title = event['title'],
|
|||
|
|
show_link = event['link'],
|
|||
|
|
show_date = datetime.strptime(event['date'], DATETIME_FORMAT),
|
|||
|
|
show_day = datetime.strptime(event['date'], DATETIME_FORMAT),
|
|||
|
|
more_details = event["details"],
|
|||
|
|
venue = venue
|
|||
|
|
)
|
|||
|
|
scraper.items+=1
|
|||
|
|
except Exception as e:
|
|||
|
|
try:
|
|||
|
|
event['date'] = c.xpath('.//div/div/time/text()')[0].split("–")[0] + " " + c.xpath('.//*/span[@class="event-time-12hr"]/text()')[0].split("–")[0].strip()
|
|||
|
|
event['dateStamp'] = datetime.strptime(event['date'], DATETIME_FORMAT)
|
|||
|
|
new_event = DSEvent.objects.update_or_create(
|
|||
|
|
calendar = scraper.calendar,
|
|||
|
|
scraper = scraper,
|
|||
|
|
event_type = event_type,
|
|||
|
|
show_title = event['title'],
|
|||
|
|
show_link = event['link'],
|
|||
|
|
show_date = datetime.strptime(event['date'], DATETIME_FORMAT),
|
|||
|
|
show_day = datetime.strptime(event['date'], DATETIME_FORMAT),
|
|||
|
|
more_details = event["details"],
|
|||
|
|
venue = venue
|
|||
|
|
)
|
|||
|
|
scraper.items+=1
|
|||
|
|
except Exception as e:
|
|||
|
|
print(e)
|
|||
|
|
print("\n\n+++\n\n")
|
|||
|
|
except Exception as e:
|
|||
|
|
continue
|
|||
|
|
|
|||
|
|
if len(sys.argv) >= 2:
|
|||
|
|
arg1 = sys.argv[1]
|
|||
|
|
br = digitools.getBrowser(arg1)
|
|||
|
|
else:
|
|||
|
|
print("No run_env")
|
|||
|
|
quit()
|
|||
|
|
|
|||
|
|
calendar_url = 'https://theparkwaytheater.com/live-events'
|
|||
|
|
ps = digitools.getSource(br, calendar_url)
|
|||
|
|
get_events(ps, "Mu")
|
|||
|
|
|
|||
|
|
calendar_url = "https://theparkwaytheater.com/movies"
|
|||
|
|
ps = digitools.getSource(br, calendar_url)
|
|||
|
|
get_events(ps, "Th")
|
|||
|
|
|
|||
|
|
# ppr(events)
|
|||
|
|
br.close()
|
|||
|
|
|
|||
|
|
digitools.updateScraper(scraper, item_count_start)
|