Files
scrapers/Working/workshop/mn.mpls.dakota.py

80 lines
2.5 KiB
Python
Raw Normal View History

2025-11-30 16:29:50 -05:00
import os, sys
from datetime import datetime
from dateutil import relativedelta
2026-01-09 22:49:19 -05:00
sys.path.append('/var/www/digisnaxx.ado/scrapers')
import dtss
dtss.getReady()
2025-11-30 16:29:50 -05:00
from time import sleep
from pprint import pprint as ppr
import pytz
from events.models import Organization, Scraper
import events.digitools as digitools
count = 0
venue, created = Organization.objects.get_or_create(
name="Dakota",
city="Minneapolis",
website="https://www.dakotacooks.com/events/",
is_venue=True
)
2026-01-09 22:49:19 -05:00
scraper,item_count_start,virtcal = digitools.getScraper(venue, 'msp')
2025-11-30 16:29:50 -05:00
# Time Signatures
tz = pytz.timezone("US/Central")
DATETIME_FORMAT = '%a %b %d, %Y • %I:%M%p'
DATETIME_FORMAT_2 = '%a %b %d, %Y • %I%p'
def get_events(ps, event_type):
links = ps.xpath('.//*/div[@class="wicked-event-title tooltipstered"]/a/@href')
links = set(links)
for l in links:
pse = getSource(br, l)
event = {}
event['scraper'] = scraper
event['calendar'] = scraper.calendar
event['link'] = l
try:
event['time'] = pse.xpath('.//*/span[@class="text-uppercase"]/text()')[0].strip()
event['dateStamp'] = datetime.strptime(event['time'], DATETIME_FORMAT)
event['title'] = pse.xpath('.//*/div[@class="sidebar-group"]/h1/text()')[0]
# event['detail-headers'] = pse.xpath('.//*/ul[@class="eo-event-meta"]/li/strong/text()')
# event['details'] = pse.xpath('.//*/ul[@class="eo-event-meta"]/li/text()')
except:
try:
event['title'] = pse.xpath('.//*/div[@class="sidebar-group"]/h1/text()')[0]
event['dateStamp'] = datetime.strptime(event['time'], DATETIME_FORMAT_2)
event['time'] = pse.xpath('.//*/span[@class="text-uppercase"]/text()')[0]
except Exception as e:
print(e)
print("failed event: ", event)
ppr(event)
try:
digitools.createBasicEvent(event, "Mu", venue)
scraper.items+=1
except Exception as e:
print(e)
print('failed to create: ', event)
if len(sys.argv) >= 2:
arg1 = sys.argv[1]
br = digitools.getBrowser(arg1)
else:
print("No run_env")
quit()
2026-01-09 22:49:19 -05:00
ps = digitools.getSource(br, "https://www.dakotacooks.com/events/?__cf_chl_tk=pnk3Xrc76yMA_SU9ADPYcnvIvjODoxJnhbnM7YcePaM-1764701728-1.0.1.1-1SD9BF5cmpUbVJVna52BCyFUZkVtXdExNBglo9nxRU8")
sleep(15)
2025-11-30 16:29:50 -05:00
get_events(ps, "Mu")
2026-01-09 22:49:19 -05:00
sleep(15)
2025-11-30 16:29:50 -05:00
digitools.updateScraper(scraper, item_count_start)
br.close()