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_str = "-0800 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="Mississippi Studios", city="Portland", website="https://mississippistudios.com/full-view/", is_venue = True ) scraper,item_count_start,virtcal = digitools.getScraper(venue, venue.website, 'pdx') scraper.items = 0 scraper.save() def get_events(contents, event_type): contents = ps.xpath('.//*/div[@class="event"]/div/div[@class="event__inner"]') i = 0 for c in contents: i += 1 try: event = {} date = c.xpath('.//*/h5/text()')[0].replace("\n", "").replace("\t", "")#.replace("rd", "").replace("nd", "").strip() time = c.xpath('//div/div/div[4]/p[1]/span/text()')[i].split("/")[0].replace("Doors: ", "").strip() event['date'] = (' ').join([date, time, tz_str]) event['date'] = (' ').join([date, time, tz_str]) event['scraper'] = scraper event['calendars'] = [scraper.calendar] event['title'] = c.xpath('.//*/h2/a/text()')[0] try: event['dateStamp'] = datetime.strptime(event['date'], DATETIME_FORMAT) except: event['dateStamp'] = datetime.strptime(event['date'], DATETIME_FORMAT_2) event['link'] = c.xpath('.//*/h2/a/@href')[0] digitools.createBasicEvent(event, "Mu", venue) scraper.items+=1 except Exception as e: print("\nError: ", e) 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) link = venue.website + "2" print(link) ps = digitools.getSource(br, link) get_events(ps, "Mu") sleep(3) digitools.updateScraper(scraper, item_count_start) br.close()