import os, sys from datetime import datetime from dateutil import relativedelta sys.path.append('/var/www/digisnaxx.ado/scrapers') import dtss dtss.getReady() from time import sleep from pprint import pprint as ppr import pytz from events.models import Organization, Scraper, Calendar, Event import events.digitools as digitools tz = pytz.timezone("US/Central") tz_str = "-0600 UTC" DATETIME_FORMAT = '%a, %b %d, %Y %I:%M%p %z %Z' DATETIME_FORMAT_2 = '%a, %b %d, %Y %I %p %z %Z' venue, created = Organization.objects.get_or_create( name="Hideout - Chicago", city="Chicago", website="https://hideoutchicago.com/", is_venue = True ) scraper,item_count_start,virtcal = digitools.getScraper(venue, venue.website, 'chi') scraper.items = 0 scraper.save() def get_events(ps, event_type): contents = ps.xpath('.//*/div[@class="col-12 eventWrapper rhpSingleEvent py-4 px-0 rhp-event__single-event--list"]') for c in contents: try: event = {} date = c.xpath('.//*/div[@id="eventDate"]/text()')[0].replace("\n", "").replace("\t", "") time = c.xpath('.//*/span[@class="font0by75 fontWeight500 lineHeight15 rhp-event__time-text--list"]/text()')[0].split("//")[1].replace("Doors: ", "").replace("\n", "").replace("\t", "") event['date'] = (' ').join([date, time, tz_str]) event['scraper'] = scraper event['calendars'] = [scraper.calendar] event['title'] = c.xpath('.//*/a[@id="eventTitle"]/h2/text()')[0].replace("\n", "").replace("\t", "") try: event['dateStamp'] = datetime.strptime(event['date'], DATETIME_FORMAT) except: event['dateStamp'] = datetime.strptime(event['date'], DATETIME_FORMAT_2) event['link'] = c.xpath('.//*/a[@id="eventTitle"]/@href')[0] digitools.createBasicEvent(event, "Mu", venue) scraper.items+=1 except Exception as e: print("\nError: ", e) # ppr(event) pass if len(sys.argv) >= 2: arg1 = sys.argv[1] br = digitools.getBrowser(arg1) else: print("No run_env") br.close() quit() ps = digitools.getSource(br, venue.website) get_events(ps, "Mu") sleep(3) digitools.updateScraper(scraper, item_count_start) br.close()