Initial revision

1 files changed, 328 insertions, 0 deletions

diff --git a/dateparse/dateparse_utils.g4 b/dateparse/dateparse_utils.g4
new file mode 100644
index 0000000..86a9cd4
--- /dev/null
+++ b/dateparse/dateparse_utils.g4
@@ -0,0 +1,328 @@
+// antlr4 -Dlanguage=Python3 ./dateparse_utils.g4
+
+// Hi, self.  In ANTLR grammars, there are two separate types of symbols: those
+// for the lexer and those for the parser.  The former begin with a CAPITAL
+// whereas the latter begin with lowercase.  The order of the lexer symbols
+// is the order that the lexer will recognize them in.  There's a good tutorial
+// on this shit at:
+//
+//    https://tomassetti.me/antlr-mega-tutorial/
+//
+// There are also a zillion premade grammars at:
+//
+//    https://github.com/antlr/grammars-v4
+
+grammar dateparse_utils;
+
+parse: dateExpr ;
+
+dateExpr
+    : singleDateExpr
+    | baseAndOffsetDateExpr
+    ;
+
+singleDateExpr
+    : monthDayMaybeYearExpr
+    | dayMonthMaybeYearExpr
+    | yearMonthDayExpr
+    | specialDateMaybeYearExpr
+    | nthWeekdayInMonthMaybeYearExpr
+    | firstLastWeekdayInMonthMaybeYearExpr
+    ;
+
+monthDayMaybeYearExpr
+    : monthExpr DIV* dayOfMonth (DIV* year)?
+    ;
+
+dayMonthMaybeYearExpr
+    : dayOfMonth DIV* monthName (DIV* year)?
+    ;
+
+yearMonthDayExpr
+    : year DIV* monthName DIV* dayOfMonth
+    ;
+
+nthWeekdayInMonthMaybeYearExpr
+    : nth dayName ('in'|'of') monthName (DIV* year)?
+    ;
+
+firstLastWeekdayInMonthMaybeYearExpr
+    : firstOrLast dayName ('in'|'of'|DIV)? monthName (DIV* year)?
+    ;
+
+specialDateMaybeYearExpr
+    : specialDate (DIV* year)?
+    ;
+
+baseAndOffsetDateExpr
+    : baseDate deltaPlusMinusExpr
+    | deltaPlusMinusExpr baseDate
+    ;
+
+baseDate: singleDateExpr ;
+
+deltaPlusMinusExpr: deltaInt deltaUnit deltaBeforeAfter? ;
+
+deltaUnit: (WEEK|DAY|SUN|WEEKDAY) ;
+
+deltaBeforeAfter: (BEFORE|AFTER) ;
+
+monthExpr
+    : monthName
+    | monthNumber
+    ;
+
+year: DIGIT DIGIT DIGIT DIGIT ;
+
+specialDate: SPECIAL_DATE ;
+
+dayOfMonth: DIGIT? DIGIT ('st'|'nd'|'rd'|'th')? ;
+
+firstOrLast: (FIRST|LAST) ;
+
+nth: DIGIT ('st'|'nd'|'rd'|'th')? ;
+
+deltaInt: ('+'|'-')? DIGIT+ ;
+
+dayName: WEEKDAY ;
+
+monthName: MONTH ;
+
+monthNumber: DIGIT? DIGIT ;
+
+// ----------------------------------
+
+COMMENT: '#' ~[\r\n]* -> skip ;
+
+SPACE: [ \t\r\n] -> skip ;
+
+THE: 'the' -> skip ;
+
+DIV: ('/'|','|'.') ;
+
+MONTH: (JAN|FEB|MAR|APR|MAY|JUN|JUL|AUG|SEP|OCT|NOV|DEC) ;
+
+JAN : 'jan'
+    | 'january'
+    ;
+
+FEB : 'feb'
+    | 'february'
+    ;
+
+MAR : 'mar'
+    | 'march'
+    ;
+
+APR : 'apr'
+    | 'april'
+    ;
+
+MAY : 'may'
+    ;
+
+JUN : 'jun'
+    | 'june'
+    ;
+
+JUL : 'jul'
+    | 'july'
+    ;
+
+AUG : 'aug'
+    | 'august'
+    ;
+
+SEP : 'sep'
+    | 'sept'
+    | 'september'
+    ;
+
+OCT : 'oct'
+    | 'october'
+    ;
+
+NOV : 'nov'
+    | 'november'
+    ;
+
+DEC : 'dec'
+    | 'december'
+    ;
+
+WEEKDAY: (SUN|MON|TUE|WED|THU|FRI|SAT) ;
+
+SUN : 'sun'
+    | 'suns'
+    | 'sunday'
+    | 'sundays'
+    ;
+
+MON : 'mon'
+    | 'mons'
+    | 'monday'
+    | 'mondays'
+    ;
+
+TUE : 'tue'
+    | 'tues'
+    | 'tuesday'
+    | 'tuesdays'
+    ;
+
+WED : 'wed'
+    | 'weds'
+    | 'wednesday'
+    | 'wednesdays'
+    ;
+
+THU : 'thu'
+    | 'thur'
+    | 'thurs'
+    | 'thursday'
+    | 'thursdays'
+    ;
+
+FRI : 'fri'
+    | 'fris'
+    | 'friday'
+    | 'fridays'
+    ;
+
+SAT : 'sat'
+    | 'sats'
+    | 'saturday'
+    | 'saturdays'
+    ;
+
+WEEK
+    : 'week'
+    | 'weeks'
+    ;
+
+DAY
+    : 'day'
+    | 'days'
+    ;
+
+SPECIAL_DATE
+    : TODAY
+    | NEW_YEARS_EVE
+    | NEW_YEARS_DAY
+    | MARTIN_LUTHER_KING_DAY
+    | PRESIDENTS_DAY
+    | EASTER
+    | MEMORIAL_DAY
+    | INDEPENDENCE_DAY
+    | LABOR_DAY
+    | COLUMBUS_DAY
+    | VETERANS_DAY
+    | THANKSGIVING_DAY
+    | CHRISTMAS_EVE
+    | CHRISTMAS
+    ;
+
+// today
+TODAY
+    : 'today'
+    ;
+
+// easte
+EASTER
+    : 'easter'
+    | 'easter sunday'
+    ;
+
+// newye
+NEW_YEARS_DAY
+    : 'new years'
+    | 'new years day'
+    | 'new year\'s'
+    | 'new year\'s day'
+    ;
+
+// newyeeve
+NEW_YEARS_EVE
+    : 'nye'
+    | 'new years eve'
+    | 'new year\'s eve'
+    ;
+
+// chris
+CHRISTMAS
+    : 'christmas'
+    | 'christmas day'
+    | 'xmas'
+    | 'xmas day'
+    ;
+
+// chriseve
+CHRISTMAS_EVE
+    : 'christmas eve'
+    | 'xmas eve'
+    ;
+
+// mlk
+MARTIN_LUTHER_KING_DAY
+    : 'martin luther king day'
+    | 'mlk day'
+    | 'mlk'
+    ;
+
+// memor
+MEMORIAL_DAY
+    : 'memorial'
+    | 'memorial day'
+    ;
+
+// indep
+INDEPENDENCE_DAY
+    : 'independence day'
+    ;
+
+// labor
+LABOR_DAY
+    : 'labor'
+    | 'labor day'
+    ;
+
+// presi
+PRESIDENTS_DAY
+    : 'presidents\' day'
+    | 'president\'s day'
+    | 'presidents day'
+    | 'presidents'
+    | 'president\'s'
+    | 'presidents\''
+    ;
+
+// colum
+COLUMBUS_DAY
+    : 'columbus'
+    | 'columbus day'
+    | 'indiginous peoples day'
+    | 'indiginous peoples\' day'
+    ;
+
+// veter
+VETERANS_DAY
+    : 'veterans'
+    | 'veterans day'
+    | 'veterans\' day'
+    ;
+
+// thank
+THANKSGIVING_DAY
+    : 'thanksgiving'
+    | 'thanksgiving day'
+    ;
+
+FIRST: 'first' ;
+
+LAST: 'last' ;
+
+BEFORE: 'before' ;
+
+AFTER: ('after'|'from') ;
+
+DIGIT: ('0'..'9') ;