diff options
author | Alban Gruin | 2018-10-09 20:30:07 +0200 |
---|---|---|
committer | Alban Gruin | 2018-10-09 20:30:07 +0200 |
commit | 11f340b4c3adb4200ff41e7e4587392b10b13e47 (patch) | |
tree | 38e55db95168f24a1f287b643850cd16cf32b621 /management | |
parent | 536061a041b5ea0ea0b64a8f4e5fa3b99a49861a (diff) |
ups2018: vérification de la présence des événements dans la pageag/ups2018-correctifs-apres-tests
Il arrive que les pages retournées par la source sont incomplètes et
ne contiennent pas d’événement, et parfois des mois entiers sont
vides.
On tente donc de récupérer une page trois fois au maximum, et, si cela
échoue toujours, on abandonne en renvoyant une exception.
Signed-off-by: Alban Gruin <alban at pa1ch dot fr>
Diffstat (limited to 'management')
-rw-r--r-- | management/parsers/ups2018.py | 67 |
1 files changed, 37 insertions, 30 deletions
diff --git a/management/parsers/ups2018.py b/management/parsers/ups2018.py index 522a26a..f1da5bf 100644 --- a/management/parsers/ups2018.py +++ b/management/parsers/ups2018.py @@ -28,20 +28,22 @@ import requests from ...models import Course, Group, Room from ...utils import get_current_week, get_week -from .abstractparser import AbstractParser +from .abstractparser import AbstractParser, ParserError VARNAME = "v.events.list = " def find_events_list(soup): res = [] + found = False for script in soup.xpath("//script/text()"): if VARNAME in script: for var in script.split('\n'): if var.startswith(VARNAME): res = json.loads(var[len(VARNAME):-2]) + found = True - return res + return res, found def get_next_month(dt): @@ -52,21 +54,40 @@ def get_next_month(dt): class Parser(AbstractParser): def __init__(self, source): super(Parser, self).__init__(source) + self.events = [self._make_request(source.url)] + self.source = source + + def _make_request(self, url, date=None): + events, found = [], False + attempts = 0 + params = {} + + if date is not None: + params["Date"] = date + + while not found: + if attempts == 3: + raise ParserError("Failed to retrieve {0}".format(url)) + attempts += 1 + + # En-tête tiré de mon Firefox… + req = super(Parser, self)._make_request( + url, params=params, + headers={"Accept-Language": "en-US,en;q=0.5"}, + ) + req.raise_for_status() - # En-tête tiré de mon Firefox… - base_req = self._make_request( - source.url, headers={"Accept-Language": "en-US,en;q=0.5"} - ) + parser = lxml.html.HTMLParser(encoding="utf8") + soup = lxml.html.document_fromstring(req.content, parser=parser) + events, found = find_events_list(soup) - parser = lxml.html.HTMLParser(encoding="utf-8") - self.soup = lxml.html.document_fromstring( - base_req.content, parser=parser - ) + if date is None: + self.months = [] + for option in soup.xpath("//option"): + if option.get("selected") is not None or len(self.months) > 0: + self.months.append(option.text) - self.months = [] - for option in self.soup.xpath("//option"): - if option.get("selected") is not None or len(self.months) > 0: - self.months.append(option.text) + return events def __get_event(self, event, today, beginning_of_month, end_of_month, @@ -179,19 +200,7 @@ class Parser(AbstractParser): ) month_str = month.replace(day=first_monday).strftime("%Y%m%d") - req = self._make_request( - self.source.url, - headers={ - "Accept-Language": "en-US,en;q=0.5", - }, - params={"Date": month_str}, - ) - req.raise_for_status() - - parser = lxml.html.HTMLParser(encoding="utf8") - soup = lxml.html.document_fromstring(req.content, parser=parser) - - return find_events_list(soup) + return self._make_request(self.source.url, month_str) @asyncio.coroutine def get_months_async(self): @@ -217,7 +226,5 @@ class Parser(AbstractParser): return events def get_source(self): - self.events = [ - find_events_list(self.soup) - ] + self.get_source_from_months() + self.events += self.get_source_from_months() return self.events |