add french language support

README.rst

@@ -48,14 +48,15 @@ How to Use
 
     In [2]: import datefinder
 
-    In [3]: matches = datefinder.find_dates(string_with_dates)
+    In [3]: matches = datefinder.find_dates(string_with_dates, locale="en_US")
 
     In [4]: for match in matches:
        ...:     print match
        ...:
     2017-01-04 20:00:00
     2005-01-15 00:00:00
 
+**Note: The `locale` parameter is optional. If you do not specify a locale, the default is `en_US`.**
 
 Demo
 ----

datefinder/__init__.py

@@ -1,4 +1,6 @@
 import copy
+import datetime
+from locale import setlocale, LC_ALL
 import logging
 import regex as re
 from dateutil import tz, parser
@@ -21,7 +23,7 @@ class DateFinder(object):
     Locates dates in a text
     """
 
-    def __init__(self, base_date=None, first="month"):
+    def __init__(self, base_date=None, first="month", locale="en_US"):
         self.base_date = base_date
         self.dayfirst = False
         self.yearfirst = False
@@ -30,6 +32,12 @@ def __init__(self, base_date=None, first="month"):
         if first == "year":
             self.yearfirst = True
 
+
+        setlocale(LC_ALL, locale) 
+        from datefinder.local_parser_info import LocaleParserInfo
+        self.parserinfo = LocaleParserInfo()
+
+
     def find_dates(self, text, source=False, index=False, strict=False):
 
         for date_string, indices, captures in self.extract_date_strings(
@@ -117,6 +125,7 @@ def parse_date_string(self, date_string, captures):
                 default=self.base_date,
                 dayfirst=self.dayfirst,
                 yearfirst=self.yearfirst,
+                parserinfo=self.parserinfo
             )
         except (ValueError, OverflowError):
             # replace tokens that are problematic for dateutil
@@ -136,6 +145,7 @@ def parse_date_string(self, date_string, captures):
                     default=self.base_date,
                     dayfirst=self.dayfirst,
                     yearfirst=self.yearfirst,
+                    parserinfo=self.parserinfo
                 )
             except Exception as e:
                 logger.debug(e)
@@ -318,7 +328,7 @@ def split_date_range(text):
 
 
 def find_dates(
-    text, source=False, index=False, strict=False, base_date=None, first="month"
+    text, source=False, index=False, strict=False, base_date=None, first="month", locale=None
 ):
     """
     Extract datetime strings from text
@@ -351,5 +361,5 @@ def find_dates(
     :return: Returns a generator that produces :mod:`datetime.datetime` objects,
         or a tuple with the source text and index, if requested
     """
-    date_finder = DateFinder(base_date=base_date, first=first)
+    date_finder = DateFinder(base_date=base_date, first=first, locale=locale)
     return date_finder.find_dates(text, source=source, index=index, strict=strict)

datefinder/constants.py

@@ -4,8 +4,8 @@
 POSITIONNAL_TOKENS = r"next|last"
 DIGITS_PATTERN = r"\d+"
 DIGITS_SUFFIXES = r"st|th|rd|nd"
-DAYS_PATTERN = "monday|tuesday|wednesday|thursday|friday|saturday|sunday|mandag|tirsdag|onsdag|torsdag|fredag|lørdag|søndag|mon|tue|tues|wed|thu|thur|thurs|fri|sat|sun|man|tir|tirs|ons|tor|tors|fre|lør|søn"
-MONTHS_PATTERN = r"january|february|march|april|may|june|july|august|september|october|november|december|enero|febrero|marzo|abril|mayo|junio|julio|agosto|septiembre|octubre|noviembre|diciembre|januar|februar|marts|april|maj|juni|juli|august|september|oktober|november|december|jan[\.\s]|ene[\.\s]|feb[\.\s]|mar[\.\s]|apr[\.\s]|abr[\.\s]|may[\.\s]|maj[\.\s]|jun[\.\s]|jul[\.\s]|aug[\.\s]|ago[\.\s]|sep[^A-Za-z]|sept[\.\s]|oct[\.\s]|okt[\.\s]|nov[\.\s]|dec[\.\s]|dic[\.\s]"
+DAYS_PATTERN = "lundi|mardi|mercredi|jeudi|vendredi|samedi|dimanche|monday|tuesday|wednesday|thursday|friday|saturday|sunday|mandag|tirsdag|onsdag|torsdag|fredag|lørdag|søndag|mon|tue|tues|wed|thu|thur|thurs|fri|sat|sun|man|tir|tirs|ons|tor|tors|fre|lør|søn"
+MONTHS_PATTERN = r"janvier|février|mars|avril|mai|juin|juillet|août|septembre|octobre|novembre|décembre|january|february|march|april|may|june|july|august|september|october|november|december|enero|febrero|marzo|abril|mayo|junio|julio|agosto|septiembre|octubre|noviembre|diciembre|januar|februar|marts|april|maj|juni|juli|august|september|oktober|november|december|jan[\.\s]|ene[\.\s]|feb[\.\s]|mar[\.\s]|apr[\.\s]|abr[\.\s]|may[\.\s]|maj[\.\s]|jun[\.\s]|jul[\.\s]|aug[\.\s]|ago[\.\s]|sep[^A-Za-z]|sept[\.\s]|oct[\.\s]|okt[\.\s]|nov[\.\s]|dec[\.\s]|dic[\.\s]"
 TIMEZONES_PATTERN = "ACDT|ACST|ACT|ACWDT|ACWST|ADDT|ADMT|ADT|AEDT|AEST|AFT|AHDT|AHST|AKDT|AKST|AKTST|AKTT|ALMST|ALMT|AMST|AMT|ANAST|ANAT|ANT|APT|AQTST|AQTT|ARST|ART|ASHST|ASHT|AST|AWDT|AWST|AWT|AZOMT|AZOST|AZOT|AZST|AZT|BAKST|BAKT|BDST|BDT|BEAT|BEAUT|BIOT|BMT|BNT|BORT|BOST|BOT|BRST|BRT|BST|BTT|BURT|CANT|CAPT|CAST|CAT|CAWT|CCT|CDDT|CDT|CEDT|CEMT|CEST|CET|CGST|CGT|CHADT|CHAST|CHDT|CHOST|CHOT|CIST|CKHST|CKT|CLST|CLT|CMT|COST|COT|CPT|CST|CUT|CVST|CVT|CWT|CXT|ChST|DACT|DAVT|DDUT|DFT|DMT|DUSST|DUST|EASST|EAST|EAT|ECT|EDDT|EDT|EEDT|EEST|EET|EGST|EGT|EHDT|EMT|EPT|EST|ET|EWT|FET|FFMT|FJST|FJT|FKST|FKT|FMT|FNST|FNT|FORT|FRUST|FRUT|GALT|GAMT|GBGT|GEST|GET|GFT|GHST|GILT|GIT|GMT|GST|GYT|HAA|HAC|HADT|HAE|HAP|HAR|HAST|HAT|HAY|HDT|HKST|HKT|HLV|HMT|HNA|HNC|HNE|HNP|HNR|HNT|HNY|HOVST|HOVT|HST|ICT|IDDT|IDT|IHST|IMT|IOT|IRDT|IRKST|IRKT|IRST|ISST|IST|JAVT|JCST|JDT|JMT|JST|JWST|KART|KDT|KGST|KGT|KIZST|KIZT|KMT|KOST|KRAST|KRAT|KST|KUYST|KUYT|KWAT|LHDT|LHST|LINT|LKT|LMT|LMT|LMT|LMT|LRT|LST|MADMT|MADST|MADT|MAGST|MAGT|MALST|MALT|MART|MAWT|MDDT|MDST|MDT|MEST|MET|MHT|MIST|MIT|MMT|MOST|MOT|MPT|MSD|MSK|MSM|MST|MUST|MUT|MVT|MWT|MYT|NCST|NCT|NDDT|NDT|NEGT|NEST|NET|NFT|NMT|NOVST|NOVT|NPT|NRT|NST|NT|NUT|NWT|NZDT|NZMT|NZST|OMSST|OMST|ORAST|ORAT|PDDT|PDT|PEST|PET|PETST|PETT|PGT|PHOT|PHST|PHT|PKST|PKT|PLMT|PMDT|PMMT|PMST|PMT|PNT|PONT|PPMT|PPT|PST|PT|PWT|PYST|PYT|QMT|QYZST|QYZT|RET|RMT|ROTT|SAKST|SAKT|SAMT|SAST|SBT|SCT|SDMT|SDT|SET|SGT|SHEST|SHET|SJMT|SLT|SMT|SRET|SRT|SST|STAT|SVEST|SVET|SWAT|SYOT|TAHT|TASST|TAST|TBIST|TBIT|TBMT|TFT|THA|TJT|TKT|TLT|TMT|TOST|TOT|TRST|TRT|TSAT|TVT|ULAST|ULAT|URAST|URAT|UTC|UYHST|UYST|UYT|UZST|UZT|VET|VLAST|VLAT|VOLST|VOLT|VOST|VUST|VUT|WARST|WART|WAST|WAT|WDT|WEDT|WEMT|WEST|WET|WFT|WGST|WGT|WIB|WIT|WITA|WMT|WSDT|WSST|WST|WT|XJT|YAKST|YAKT|YAPT|YDDT|YDT|YEKST|YEKST|YEKT|YEKT|YERST|YERT|YPT|YST|YWT|zzz"
 ## explicit north american timezones that get replaced
 NA_TIMEZONES_PATTERN = "pacific|eastern|mountain|central"

datefinder/local_parser_info.py

@@ -0,0 +1,6 @@
+import calendar
+from dateutil import parser
+
+class LocaleParserInfo(parser.parserinfo):
+    WEEKDAYS = list(zip(calendar.day_abbr, calendar.day_name))
+    MONTHS = list(zip(calendar.month_abbr, calendar.month_name))[1:]