jan
/
punkow
mirror of https://github.com/janLo/punkow
1
0
Fork 0
Code Issues Releases Wiki Activity

Also look for appointments in months not shown

This commit is contained in:
Jan Losinski 2019-01-07 14:27:33 +01:00
parent b2349975e1
commit f98b3d2d8c
1 changed files with 28 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
punkow/scraper.py
View File

@ -85,22 +85,40 @@ class BookingService(object):
def _iter_bookable_day_urls(self, start_url): def _iter_bookable_day_urls(self, start_url):
html = self._fetch_soup(start_url) html = self._fetch_soup(start_url)
months = html.find_all("div", {"class": "calendar-month-table"})
zms = html.find("div", {"class": "zms"}) zms = html.find("div", {"class": "zms"})
self._print_details(zms, 2) self._print_details(zms, 2)
for m_div in months: checked = []
month_name = m_div.find("th", {"class": "month"})
bookable = m_div.find_all('td', {"class": 'buchbar'})
logger.info("Found month %s with %d available days", month_name.text.strip(), len(bookable))
for day in bookable: while html is not None:
day_link = day.find("a") months = html.find_all("div", {"class": "calendar-month-table"})
for m_div in months:
month_name = m_div.find("th", {"class": "month"}).text.strip()
if month_name in checked:
logging.info("Month %s aleady checked - skipping", month_name)
continue
if day_link and day_link != -1: bookable = m_div.find_all('td', {"class": 'buchbar'})
logger.info("Search free slots for day %s. %s", day_link.text.strip(), month_name.text.strip()) logger.info("Found month %s with %d available days", month_name, len(bookable))
yield day_link.attrs["href"]
for day in bookable:
day_link = day.find("a")
if day_link and day_link != -1:
logger.info("Search free slots for day %s. %s", day_link.text.strip(), month_name)
yield day_link.attrs["href"]
checked.append(month_name)
next_field = html.find("th", {"class": "next"})
if next_field is None or next_field == -1:
break
next_link = next_field.find("a")
if next_link is None or next_link == -1:
break
html = self._fetch_soup(next_link["href"])
logger.info("No more days with appointments") logger.info("No more days with appointments")