diff options
| author | Petter Reinholdtsen <pere@hungry.com> | 2015-01-17 21:02:14 +0100 |
|---|---|---|
| committer | Petter Reinholdtsen <pere@hungry.com> | 2015-01-17 21:03:52 +0100 |
| commit | 0bc6a3f36cc86e7f89cb10219558c464a58bfc0c (patch) | |
| tree | 36d68d6100f887b1c672e32f347c03edfcf8892f | |
| parent | 1db83ecb49ccf5abac641409adbfda5eb1d86d34 (diff) | |
Fix parser.
| -rw-r--r-- | scrapersources/postliste-bergen-kommune | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/scrapersources/postliste-bergen-kommune b/scrapersources/postliste-bergen-kommune index 096b06a..98a9ef6 100644 --- a/scrapersources/postliste-bergen-kommune +++ b/scrapersources/postliste-bergen-kommune @@ -80,14 +80,14 @@ def fetch_postjournal_day(parser, url, html, saver): raise ValueError("Something is strange, missing case ID") matchObj = \ - re.match("^Dokumentdato:\s+(\d{2}).(\d{2}).(\d{4})$", + re.match("^Dokumentdato:[^\d]+(\d{2}).(\d{2}).(\d{4})$", line, re.M|re.I) if matchObj: entry['docdate'] = "%s-%s-%s" % (matchObj.group(3), matchObj.group(2), matchObj.group(1)) matchObj = \ - re.match("^Journal Dato:\s+(\d{2}).(\d{2}).(\d{4})$", + re.match("^Journal Dato:[^\d]+(\d{2}).(\d{2}).(\d{4})$", line, re.M|re.I) if matchObj: entry['recorddate'] = "%s-%s-%s" % (matchObj.group(3), @@ -143,7 +143,7 @@ print "Fetching public journal!" parser = postlistelib.JournalParser(agency=agency) -parsedays = 46 +parsedays = 48 today = datetime.date.today() i = 1 |