XSLT imports: better date management - refactoring - manage date ISO format

1 files changed, 65 insertions, 15 deletions

diff --git a/chimere/utils.py b/chimere/utils.py
index ba5e558..55fc45c 100644
--- a/chimere/utils.py
+++ b/chimere/utils.py
@@ -1,6 +1,6 @@
 #!/usr/bin/env python
 # -*- coding: utf-8 -*-
-# Copyright (C) 2012-2013  Étienne Loks  <etienne.loks_AT_peacefrogsDOTnet>
+# Copyright (C) 2012-2015  Étienne Loks  <etienne.loks_AT_peacefrogsDOTnet>
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as
@@ -879,7 +879,35 @@ DATE_PARSINGS = {'fr_FR':[
            re.compile(r'(?P<day1>\d{1,2}) '\
                       r'(?P<month1>'+ '|'.join(UNI_MONTH_NAMES['fr_FR']) +') *'\
                       r'(?P<year1>\d{4})?')
-                     ]
+                     ],
+                 'en':[
+           re.compile(r'(?P<year1>\d{4})-'\
+                      r'(?P<month1>\d{2})-'\
+                      r'(?P<day1>\d{2})'\
+                      r'(?:T'\
+                          r'(?P<hour1>\d{2})?:'\
+                          r'(?P<minut1>\d{2})?:'\
+                          r'(?P<second1>\d{2})'\
+                      r')?.*'\
+                      r'(?P<year2>\d{4})-'\
+                      r'(?P<month2>\d{2})-'\
+                      r'(?P<day2>\d{2})'\
+                      r'(?:T'\
+                          r'(?P<hour2>\d{2})?:'\
+                          r'(?P<minut2>\d{2})?:'\
+                          r'(?P<second2>\d{2})'\
+                      r')?.*'
+                      ),
+           re.compile(r'(?P<year1>\d{4})-'\
+                      r'(?P<month1>\d{2})-'\
+                      r'(?P<day1>\d{2})'\
+                      r'(?:T'\
+                          r'(?P<hour1>\d{2})?:'\
+                          r'(?P<minut1>\d{2})?:'\
+                          r'(?P<second1>\d{2})'\
+                      r')?'
+                      )
+                     ],
             }
 
 def clean_field(value):
@@ -979,6 +1007,29 @@ class HtmlXsltManager(ImportManager):
                     u'", "'.join(self.missing_cats))
         return (self.new_item, self.updated_item, msg)
 
+    @classmethod
+    def _internal_parse_date(cls, locale, year, month, day):
+        try:
+            year = datetime.date.today().year if not year else int(year)
+        except ValueError:
+            return
+        month = month.encode('utf-8')
+        if locale in MONTH_NAMES and month in MONTH_NAMES[locale]:
+            month = MONTH_NAMES[locale].index(month) + 1
+        else:
+            try:
+                month = int(month)
+            except ValueError:
+                return
+        try:
+            day = int(day)
+        except ValueError:
+            return
+        try:
+            return datetime.date(year, month, day)
+        except ValueError:
+            return
+
     def parse_date(self, date):
         dct = {}
         has_dates = False
@@ -989,22 +1040,21 @@ class HtmlXsltManager(ImportManager):
                 m = r.search(date)
                 if not m:
                     continue
-                has_dates = True
                 values = m.groupdict()
-                year1 = datetime.date.today().year if 'year1' not in values \
-                        else int(values['year1'])
-                dct['start_date'] = datetime.date(year1,
-                  MONTH_NAMES[locale].index(values['month1'].encode('utf-8')
-                                                                        ) + 1,
-                  int(values['day1']))
+                date = self._internal_parse_date(locale,
+                              'year1' in values and values['year1'],
+                              values['month1'], values['day1'])
+                if not date:
+                    continue
+                dct['start_date'] = date
+                has_dates = True
                 if 'day2' not in values:
                     break
-                year2 = datetime.date.today().year if 'year2' not in values \
-                        else int(values['year2'])
-                dct['end_date'] = datetime.date(year2,
-                  MONTH_NAMES[locale].index(values['month2'].encode('utf-8')
-                                                                        ) + 1,
-                  int(values['day2']))
+                date = self._internal_parse_date(locale,
+                              'year2' in values and values['year2'],
+                              values['month2'], values['day2'])
+                if date:
+                    dct['end_date'] = date
                 break
         return dct