gramps/gramps2/src/DateParser.py
Don Allingham 7e70532009 * src/DateParser.py: documenation enhancements
* src/GrampsBSDDB.py: Add cursor class
* src/GrampsDbBase.py: Add base cursor class
* src/GrampsGEDDB.py: documenation enhancements
* src/GrampsInMemDB.py: Add cursor class
* src/GrampsXMLDB.py: documenation enhancements


svn: r3781
2004-12-06 04:13:13 +00:00

592 lines
22 KiB
Python

#
# Gramps - a GTK+/GNOME based genealogy program
#
# Copyright (C) 2004 Donald N. Allingham
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
#
# $Id$
"""
Date parsing class. Serves as the base class for any localized
date parsing class. The default, base class provides parsing for
English.
"""
__author__ = "Donald N. Allingham"
__version__ = "$Revision$"
#-------------------------------------------------------------------------
#
# Python modules
#
#-------------------------------------------------------------------------
import re
import locale
import calendar
#-------------------------------------------------------------------------
#
# GRAMPS modules
#
#-------------------------------------------------------------------------
import Date
#-------------------------------------------------------------------------
#
# Top-level module functions
#
#-------------------------------------------------------------------------
_max_days = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ]
_leap_days = [ 31, 29, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ]
def gregorian_valid(date_tuple):
day = date_tuple[0]
month = date_tuple[1]
valid = True
try:
if month > 12:
valid = False
elif calendar.isleap(date_tuple[2]):
if day > _leap_days[month-1]:
valid = False
elif day > _max_days[month-1]:
valid = False
except:
valid = False
return valid
#-------------------------------------------------------------------------
#
# Parser class
#
#-------------------------------------------------------------------------
class DateParser:
"""
Converts a text string into a Date object. If the date cannot be
converted, the text string is assigned.
"""
# determine the code set returned by nl_langinfo
_codeset = locale.nl_langinfo(locale.CODESET)
_fmt_parse = re.compile(".*%(\S).*%(\S).*%(\S).*")
# RFC-2822 only uses capitalized English abbreviated names, no locales.
_rfc_days = ('Sun','Mon','Tue','Wed','Thu','Fri','Sat')
_rfc_mons_to_int = {
'Jan' : 1, 'Feb' : 2, 'Mar' : 3, 'Apr' : 4,
'May' : 5, 'Jun' : 6, 'Jul' : 7, 'Aug' : 8,
'Sep' : 9, 'Oct' : 10, 'Nov' : 11, 'Dec' : 12,
}
month_to_int = {
unicode(locale.nl_langinfo(locale.MON_1),_codeset).lower() : 1,
unicode(locale.nl_langinfo(locale.ABMON_1),_codeset).lower() : 1,
unicode(locale.nl_langinfo(locale.MON_2),_codeset).lower() : 2,
unicode(locale.nl_langinfo(locale.ABMON_2),_codeset).lower() : 2,
unicode(locale.nl_langinfo(locale.MON_3),_codeset).lower() : 3,
unicode(locale.nl_langinfo(locale.ABMON_3),_codeset).lower() : 3,
unicode(locale.nl_langinfo(locale.MON_4),_codeset).lower() : 4,
unicode(locale.nl_langinfo(locale.ABMON_4),_codeset).lower() : 4,
unicode(locale.nl_langinfo(locale.MON_5),_codeset).lower() : 5,
unicode(locale.nl_langinfo(locale.ABMON_5),_codeset).lower() : 5,
unicode(locale.nl_langinfo(locale.MON_6),_codeset).lower() : 6,
unicode(locale.nl_langinfo(locale.ABMON_6),_codeset).lower() : 6,
unicode(locale.nl_langinfo(locale.MON_7),_codeset).lower() : 7,
unicode(locale.nl_langinfo(locale.ABMON_7),_codeset).lower() : 7,
unicode(locale.nl_langinfo(locale.MON_8),_codeset).lower() : 8,
unicode(locale.nl_langinfo(locale.ABMON_8),_codeset).lower() : 8,
unicode(locale.nl_langinfo(locale.MON_9),_codeset).lower() : 9,
unicode(locale.nl_langinfo(locale.ABMON_9),_codeset).lower() : 9,
unicode(locale.nl_langinfo(locale.MON_10),_codeset).lower() : 10,
unicode(locale.nl_langinfo(locale.ABMON_10),_codeset).lower(): 10,
unicode(locale.nl_langinfo(locale.MON_11),_codeset).lower() : 11,
unicode(locale.nl_langinfo(locale.ABMON_11),_codeset).lower(): 11,
unicode(locale.nl_langinfo(locale.MON_12),_codeset).lower() : 12,
unicode(locale.nl_langinfo(locale.ABMON_12),_codeset).lower(): 12,
}
modifier_to_int = {
'before' : Date.MOD_BEFORE, 'bef' : Date.MOD_BEFORE,
'bef.' : Date.MOD_BEFORE, 'after' : Date.MOD_AFTER,
'aft' : Date.MOD_AFTER, 'aft.' : Date.MOD_AFTER,
'about' : Date.MOD_ABOUT, 'abt.' : Date.MOD_ABOUT,
'abt' : Date.MOD_ABOUT, 'circa' : Date.MOD_ABOUT,
'c.' : Date.MOD_ABOUT, 'around' : Date.MOD_ABOUT,
}
hebrew_to_int = {
"tishri" : 1, "heshvan" : 2, "kislev" : 3,
"tevet" : 4, "shevat" : 5, "adari" : 6,
"adarii" : 7, "nisan" : 8, "iyyar" : 9,
"sivan" : 10, "tammuz" : 11, "av" : 12,
"elul" : 13,
}
french_to_int = {
u'vend\xc3\xa9miaire' : 1, 'brumaire' : 2,
'frimaire' : 3, u'niv\xc3\xb4se ': 4,
u'pluvi\xc3\xb4se' : 5, u'vent\xc3\xb4se' : 6,
'germinal' : 7, u'flor\xc3\xa9al' : 8,
'prairial' : 9, 'messidor' : 10,
'thermidor' : 11, 'fructidor' : 12,
'extra' : 13
}
islamic_to_int = {
"muharram" : 1, "muharram ul haram" : 1,
"safar" : 2, "rabi`al-awwal" : 3,
"rabi'l" : 3, "rabi`ul-akhir" : 4,
"rabi`ath-thani" : 4, "rabi` ath-thani" : 4,
"rabi`al-thaany" : 4, "rabi` al-thaany" : 4,
"rabi' ii" : 4, "jumada l-ula" : 5,
"jumaada-ul-awwal" : 5, "jumaada i" : 5,
"jumada t-tania" : 6, "jumaada-ul-akhir" : 6,
"jumaada al-thaany" : 6, "jumaada ii" : 5,
"rajab" : 7, "sha`ban" : 8,
"sha`aban" : 8, "ramadan" : 9,
"ramadhan" : 9, "shawwal" : 10,
"dhu l-qa`da" : 11, "dhu qadah" : 11,
"thw al-qi`dah" : 11, "dhu l-hijja" : 12,
"dhu hijja" : 12, "thw al-hijjah" : 12,
}
persian_to_int = {
"Farvardin" : 1, "Ordibehesht" : 2,
"Khordad" : 3, "Tir" : 4,
"Mordad" : 5, "Shahrivar" : 6,
"Mehr" : 7, "Aban" : 8,
"Azar" : 9, "Dey" : 10,
"Bahman" : 11, "Esfand" : 12,
}
bce = ["BC", "B\.C", "B\.C\.", "BCE", "B\.C\.E", "B\.C\.E"]
calendar_to_int = {
'gregorian' : Date.CAL_GREGORIAN,
'g' : Date.CAL_GREGORIAN,
'julian' : Date.CAL_JULIAN,
'j' : Date.CAL_JULIAN,
'hebrew' : Date.CAL_HEBREW,
'h' : Date.CAL_HEBREW,
'islamic' : Date.CAL_ISLAMIC,
'i' : Date.CAL_ISLAMIC,
'french' : Date.CAL_FRENCH,
'french republican': Date.CAL_FRENCH,
'f' : Date.CAL_FRENCH,
'persian' : Date.CAL_PERSIAN,
'p' : Date.CAL_PERSIAN,
}
quality_to_int = {
'estimated' : Date.QUAL_ESTIMATED,
'est.' : Date.QUAL_ESTIMATED,
'est' : Date.QUAL_ESTIMATED,
'calc.' : Date.QUAL_CALCULATED,
'calc' : Date.QUAL_CALCULATED,
'calculated' : Date.QUAL_CALCULATED,
}
def __init__(self):
self.init_strings()
self.parser = {
Date.CAL_GREGORIAN : self._parse_greg_julian,
Date.CAL_JULIAN : self._parse_greg_julian,
Date.CAL_PERSIAN : self._parse_persian,
Date.CAL_HEBREW : self._parse_hebrew,
Date.CAL_ISLAMIC : self._parse_islamic,
}
fmt = locale.nl_langinfo(locale.D_FMT)
match = self._fmt_parse.match(fmt.lower())
if match:
self.dmy = (match.groups() == ('d','m','y'))
else:
self.dmy = True
def init_strings(self):
"""
This method compiles regular expression strings for matching dates.
Most of the re's in most languages can stay as is. span and range
most likely will need to change. Whatever change is done, this method
may be called first as DateParser.init_strings(self) so that the
invariant expresions don't need to be repeteadly coded. All differences
can be coded after DateParser.init_strings(self) call, that way they
override stuff from this method. See DateParserRU() as an example.
"""
self._rfc_mon_str = '(' + '|'.join(self._rfc_mons_to_int.keys()) + ')'
self._rfc_day_str = '(' + '|'.join(self._rfc_days) + ')'
self._bce_str = '(' + '|'.join(self.bce) + ')'
self._qual_str = '(' + '|'.join(
[ key.replace('.','\.') for key in self.quality_to_int.keys() ]
) + ')'
self._mod_str = '(' + '|'.join(
[ key.replace('.','\.') for key in self.modifier_to_int.keys() ]
) + ')'
# Need to reverse-sort the keys, so that April matches before Apr does.
# Otherwise, 'april 2000' would be matched as 'apr' + garbage ('il 2000')
_month_keys = self.month_to_int.keys()
_month_keys.sort()
_month_keys.reverse()
self._mon_str = '(' + '|'.join(_month_keys) + ')'
self._jmon_str = '(' + '|'.join(self.hebrew_to_int.keys()) + ')'
self._fmon_str = '(' + '|'.join(self.french_to_int.keys()) + ')'
self._pmon_str = '(' + '|'.join(self.persian_to_int.keys()) + ')'
self._cal_str = '(' + '|'.join(self.calendar_to_int.keys()) + ')'
self._imon_str = '(' + '|'.join(self.islamic_to_int.keys()) + ')'
self._bce_re = re.compile("(.+)\s+%s" % self._bce_str)
self._cal = re.compile("(.+)\s\(%s\)" % self._cal_str,
re.IGNORECASE)
self._qual = re.compile("%s\s+(.+)" % self._qual_str,
re.IGNORECASE)
self._span = re.compile("(from)\s+(.+)\s+(to)\s+(.+)",
re.IGNORECASE)
self._range = re.compile("(bet|bet.|between)\s+(.+)\s+(and)\s+(.+)",
re.IGNORECASE)
self._modifier = re.compile('%s\s+(.*)' % self._mod_str,
re.IGNORECASE)
self._text = re.compile('%s\s+(\d+)?\s*,?\s*((\d+)(/\d+)?)?' % self._mon_str,
re.IGNORECASE)
self._text2 = re.compile('(\d+)?\s+?%s\s*((\d+)(/\d+)?)?' % self._mon_str,
re.IGNORECASE)
self._jtext = re.compile('%s\s+(\d+)?\s*,?\s*((\d+)(/\d+)?)?' % self._jmon_str,
re.IGNORECASE)
self._jtext2 = re.compile('(\d+)?\s+?%s\s*((\d+)(/\d+)?)?' % self._jmon_str,
re.IGNORECASE)
self._ftext = re.compile('%s\s+(\d+)?\s*,?\s*((\d+)(/\d+)?)?' % self._fmon_str,
re.IGNORECASE)
self._ftext2 = re.compile('(\d+)?\s+?%s\s*((\d+)(/\d+)?)?' % self._fmon_str,
re.IGNORECASE)
self._ptext = re.compile('%s\s+(\d+)?\s*,?\s*((\d+)(/\d+)?)?' % self._pmon_str,
re.IGNORECASE)
self._ptext2 = re.compile('(\d+)?\s+?%s\s*((\d+)(/\d+)?)?' % self._pmon_str,
re.IGNORECASE)
self._itext = re.compile('%s\s+(\d+)?\s*,?\s*((\d+)(/\d+)?)?' % self._imon_str,
re.IGNORECASE)
self._itext2 = re.compile('(\d+)?\s+?%s\s*((\d+)(/\d+)?)?' % self._imon_str,
re.IGNORECASE)
self._range2 = re.compile('%s\s+(\d+)-(\d+)\s*,?\s*((\d+)(/\d+)?)?' % self._mon_str,
re.IGNORECASE)
self._numeric = re.compile("((\d+)[/\.])?((\d+)[/\.])?(\d+)")
self._iso = re.compile("(\d+)-(\d+)-(\d+)")
self._rfc = re.compile("(%s,)?\s+(\d|\d\d)\s+%s\s+(\d+)\s+\d\d:\d\d(:\d\d)?\s+(\+|-)\d\d\d\d"
% (self._rfc_day_str,self._rfc_mon_str))
def _get_int(self,val):
"""
Converts the string to an integer if the value is not None. If the
value is None, a zero is returned
"""
if val == None:
return 0
else:
return int(val)
def _parse_hebrew(self,text):
return self._parse_calendar(text,self._jtext,self._jtext2,
self.hebrew_to_int)
def _parse_islamic(self,text):
return self._parse_calendar(text,self._itext,self._itext2,
self.islamic_to_int)
def _parse_persian(self,text):
return self._parse_calendar(text,self._ptext,self._ptext2,
self.persian_to_int)
def _parse_french(self,text):
return self._parse_calendar(text,self._ftext,self._ftext2,
self.french_to_int)
def _parse_greg_julian(self,text):
return self._parse_calendar(text,self._text,self._text2,
self.month_to_int,gregorian_valid)
def _parse_calendar(self,text,regex1,regex2,mmap,check=None):
match = regex1.match(text.lower())
if match:
groups = match.groups()
if groups[0] == None:
m = 0
else:
m = mmap[groups[0].lower()]
if groups[2] == None:
y = self._get_int(groups[1])
d = 0
s = None
else:
d = self._get_int(groups[1])
y = int(groups[3])
s = groups[4] != None
value = (d,m,y,s)
if check and not check((d,m,y)):
value = Date.EMPTY
return value
match = regex2.match(text.lower())
if match:
groups = match.groups()
if groups[1] == None:
m = 0
else:
m = mmap[groups[1].lower()]
d = self._get_int(groups[0])
if groups[2] == None:
y = 0
s = None
else:
y = int(groups[3])
s = groups[4] != None
value = (d,m,y,s)
if check and not check((d,m,y)):
value = Date.EMPTY
return value
return Date.EMPTY
def _parse_subdate(self,text,subparser=None):
"""
Converts only the date portion of a date.
"""
if subparser == None:
subparser = self._parse_greg_julian
check = gregorian_valid
else:
check = None
value = subparser(text)
if value != Date.EMPTY:
return value
match = self._iso.match(text)
if match:
groups = match.groups()
y = self._get_int(groups[0])
m = self._get_int(groups[1])
d = self._get_int(groups[2])
if gregorian_valid((d,m,y)):
return (d,m,y,False)
else:
return Date.EMPTY
match = self._rfc.match(text)
if match:
groups = match.groups()
d = self._get_int(groups[2])
m = self._rfc_mons_to_int[groups[3]]
y = self._get_int(groups[4])
if gregorian_valid((d,m,y)):
return (d,m,y,False)
else:
return Date.EMPTY
match = self._numeric.match(text)
if match:
groups = match.groups()
if self.dmy:
m = self._get_int(groups[3])
d = self._get_int(groups[1])
else:
m = self._get_int(groups[1])
d = self._get_int(groups[3])
y = self._get_int(groups[4])
value = (d,m,y,False)
if check and not check((d,m,y)):
value = Date.EMPTY
return value
return Date.EMPTY
def match_calendar(self,text,cal):
"""
Try parsing calendar.
Return calendar index and the remainder of text.
"""
match = self._cal.match(text)
if match:
grps = match.groups()
cal = self.calendar_to_int[grps[1].lower()]
text = grps[0]
return (text,cal)
def match_quality(self,text,qual):
"""
Try matching quality.
Return quality index and the remainder of text.
"""
match = self._qual.match(text)
if match:
grps = match.groups()
qual = self.quality_to_int[grps[0].lower()]
text = grps[1]
return (text,qual)
def match_span(self,text,cal,qual,date):
"""
Try matching span date.
On success, set the date and return 1. On failure return 0.
"""
match = self._span.match(text)
if match:
grps = match.groups()
text_parser = self.parser[cal]
start = self._parse_subdate(grps[1],text_parser)
stop = self._parse_subdate(grps[3],text_parser)
date.set(qual,Date.MOD_SPAN,cal,start + stop)
return 1
return 0
def match_range(self,text,cal,qual,date):
"""
Try matching range date.
On success, set the date and return 1. On failure return 0.
"""
match = self._range.match(text)
if match:
grps = match.groups()
text_parser = self.parser[cal]
start = self._parse_subdate(grps[1],text_parser)
stop = self._parse_subdate(grps[3],text_parser)
date.set(qual,Date.MOD_RANGE,cal,start + stop)
return 1
return 0
def match_range2(self,text,cal,qual,date):
"""
Try matching numerical range date.
On success, set the date and return 1. On failure return 0.
"""
match = self._range2.match(text)
if match:
grps = match.groups()
m = self.month_to_int[grps[0].lower()]
d0 = self._get_int(grps[1])
d1 = self._get_int(grps[2])
if grps[3] == None:
y = 0
s = None
else:
y = int(grps[3])
s = grps[4] != None
date.set(qual,Date.MOD_RANGE,Date.CAL_GREGORIAN,
(d0,m,y,s,d1,m,y,s))
return 1
return 0
def match_bce(self,text):
"""
Try matching BCE qualifier.
Return BCE (True/False) and the remainder of text.
"""
match = self._bce_re.match(text)
bc = False
if match:
text = match.groups()[0]
bc = True
return (text,bc)
def match_modifier(self,text,cal,qual,bc,date):
"""
Try matching date with modifier.
On success, set the date and return 1. On failure return 0.
"""
match = self._modifier.match(text)
if match:
grps = match.groups()
start = self._parse_subdate(grps[1])
mod = self.modifier_to_int.get(grps[0].lower(),Date.MOD_NONE)
if bc:
date.set(qual,mod,cal,self.invert_year(start))
else:
date.set(qual,mod,cal,start)
return 1
return 0
def set_date(self,date,text):
"""
Parses the text and sets the date according to the parsing.
"""
date.set_text_value(text)
qual = Date.QUAL_NONE
cal = Date.CAL_GREGORIAN
(text,cal) = self.match_calendar(text,cal)
(text,qual) = self.match_quality(text,qual)
if self.match_span(text,cal,qual,date):
return
if self.match_range(text,cal,qual,date):
return
if self.match_range2(text,cal,qual,date):
return
(text,bc) = self.match_bce(text)
if self.match_modifier(text,cal,qual,bc,date):
return
subdate = self._parse_subdate(text)
if subdate == Date.EMPTY:
subdate = self._parse_hebrew(text)
if subdate == Date.EMPTY:
subdate = self._parse_persian(text)
if subdate == Date.EMPTY:
subdate = self._parse_islamic(text)
if subdate == Date.EMPTY:
subdate = self._parse_french(text)
if subdate == Date.EMPTY:
date.set_as_text(text)
return
else:
cal = Date.CAL_FRENCH
else:
cal = Date.CAL_ISLAMIC
else:
cal = Date.CAL_PERSIAN
else:
cal = Date.CAL_HEBREW
if bc:
date.set(qual,Date.MOD_NONE,cal,self.invert_year(subdate))
else:
date.set(qual,Date.MOD_NONE,cal,subdate)
def invert_year(self,subdate):
return (subdate[0],subdate[1],-subdate[2],subdate[3])
def parse(self,text):
"""
Parses the text, returning a Date object.
"""
new_date = Date.Date()
self.set_date(new_date,text)
return new_date