acute 1.0.0

data/data/words.yml

@@ -0,0 +1,699 @@
+# Words the lexer knows, grouped by token kind. bin/generate turns this into lib/acute/tables/words.rb.
+# Matching ignores ASCII case. A word may hold letters, apostrophes and inner dots ("p.m").
+# A value is an Integer, or a name that becomes a Symbol.
+
+day:            # value: days from today
+  today: 0
+  tomorrow: 1
+  tmrw: 1
+  tmr: 1
+  yesterday: -1
+  today's: 0
+  todays: 0
+  tomorrow's: 1
+  tomorrows: 1
+  yesterday's: -1
+
+meridiem:       # value: hours added to 1-12
+  am: 0
+  a.m: 0
+  pm: 12
+  p.m: 12
+  p: 12
+
+named:          # value: the hour it names
+  noon: 12
+  midday: 12
+  midnight: 0
+
+now:
+  now: 0
+
+at:
+  at: 0
+
+weekday:        # value: 0 = Sunday
+  sunday: 0
+  sun: 0
+  monday: 1
+  mon: 1
+  tuesday: 2
+  tue: 2
+  tues: 2
+  tu: 2
+  wednesday: 3
+  wed: 3
+  midweek: 3      # D2
+  thursday: 4
+  thu: 4
+  thur: 4
+  thurs: 4
+  friday: 5
+  fri: 5
+  fr: 5
+  frid: 5
+  saturday: 6
+  sat: 6
+  sundays: 0
+  mondays: 1
+  tuesdays: 2
+  wednesdays: 3
+  thursdays: 4
+  fridays: 5
+  saturdays: 6
+
+month:
+  january: 1
+  jan: 1
+  february: 2
+  feb: 2
+  march: 3
+  mar: 3
+  april: 4
+  apr: 4
+  may: 5
+  june: 6
+  jun: 6
+  july: 7
+  jul: 7
+  august: 8
+  aug: 8
+  september: 9
+  sep: 9
+  sept: 9
+  october: 10
+  oct: 10
+  november: 11
+  nov: 11
+  december: 12
+  dec: 12
+
+unit:
+  s: second
+  sec: second
+  secs: second
+  seconds: second
+  m: minute
+  min: minute
+  mins: minute
+  minute: minute
+  minutes: minute
+  h: hour
+  hr: hour
+  hrs: hour
+  hour: hour
+  hours: hour
+  d: day
+  day: day
+  days: day
+  w: week
+  wk: week
+  wks: week
+  week: week
+  week's: week
+  weeks: week
+  fortnight: fortnight
+  fortnight's: fortnight
+  fortnights: fortnight
+  mo: month
+  month: month
+  month's: month
+  months: month
+  quarter: quarter
+  quarters: quarter
+  qtr: quarter
+  qtrs: quarter
+  y: year
+  yr: year
+  yrs: year
+  year: year
+  year's: year
+  years: year
+  decade: decade
+  decades: decade
+  century: century
+  season: season    # D48: three months from December, March, June or September
+  seasons: season
+  centuries: century
+  weekday: business_day
+  weekdays: business_day
+
+relative:       # "past" is also the clock word in "twenty past four"
+  next: next
+  this: this
+  current: this
+  last: last
+  past: past
+  coming: next
+  upcoming: next
+  following: next
+  previous: last
+  prev: last
+
+number:
+  one: 1
+  two: 2
+  three: 3
+  four: 4
+  five: 5
+  six: 6
+  seven: 7
+  eight: 8
+  nine: 9
+  ten: 10
+  eleven: 11
+  twelve: 12
+  thirteen: 13
+  fourteen: 14
+  fifteen: 15
+  sixteen: 16
+  seventeen: 17
+  eighteen: 18
+  nineteen: 19
+  twenty: 20
+  thirty: 30
+  forty: 40
+  fifty: 50
+  sixty: 60
+  seventy: 70
+  eighty: 80
+  ninety: 90
+
+decade:         # D12: value: the decade's last two digits
+  twenties: 20
+  thirties: 30
+  forties: 40
+  fifties: 50
+  sixties: 60
+  seventies: 70
+  eighties: 80
+  nineties: 90
+
+ordinal:        # "second" is also the unit, by context
+  first: 1
+  final: -1      # only before what it picks, "final thursday in april"
+  second: 2
+  third: 3
+  fourth: 4
+  fifth: 5
+  sixth: 6
+  seventh: 7
+  eighth: 8
+  ninth: 9
+  tenth: 10
+  eleventh: 11
+  twelfth: 12
+  thirteenth: 13
+  fourteenth: 14
+  fifteenth: 15
+  sixteenth: 16
+  seventeenth: 17
+  eighteenth: 18
+  nineteenth: 19
+  twentieth: 20
+  thirtieth: 30
+
+part:           # parts of the day (D1, D35)
+  morning: morning
+  afternoon: afternoon
+  evening: evening
+  mornings: morning
+  afternoons: afternoon
+  evenings: evening
+  nights: night
+  night: night
+  tonight: tonight
+  tonight's: tonight
+  lunch: lunch
+  lunchtime: lunch
+
+season:         # D3
+  spring: spring
+  summer: summer
+  autumn: autumn
+  fall: autumn
+  winter: winter
+  springs: spring
+  summers: summer
+  autumns: autumn
+  falls: autumn
+  winters: winter
+
+holiday:        # D15, D28: formulas only
+  christmas: christmas
+  xmas: christmas
+  thanksgiving: thanksgiving
+  halloween: halloween
+  easter: easter
+  juneteenth: juneteenth
+  pentecost: pentecost
+  whitsunday: pentecost
+  whitsun: pentecost
+  thanksgivings: thanksgiving
+  kwanzaa: kwanzaa
+
+weekend:
+  weekend: 0
+  weekends: 0
+  wkend: 0
+
+zone:           # value: the UTC offset in seconds (D18, D40: abbreviations are fixed offsets), or an IANA name
+                # D38: an ambiguous one lists its offsets in order; a leading country is where the first alone holds
+  utc: 0
+  gmt: 0
+  z: 0
+  est: -18000
+  edt: -14400
+  cst: -21600
+  cdt: -18000
+  mst: -25200
+  mdt: -21600
+  pst: -28800
+  pdt: -25200
+  cet: 3600
+  cest: 7200
+  bst: [GB, 3600, 21600]
+  ist: [19800, 3600, 7200]
+  aest: 36000
+  aedt: 39600
+  jst: 32400
+  msk: 10800
+  hst: -36000
+  akst: -32400
+  akdt: -28800
+  eet: 7200
+  eest: 10800
+  nzst: 43200
+  nzdt: 46800
+  acst: 34200
+  acdt: 37800
+  awst: 28800
+  sgt: 28800
+  hkt: 28800
+  kst: 32400
+  # D44: a US zone named without standard or daylight follows its clock changes
+  et: America/New_York
+  eastern: America/New_York
+  ct: America/Chicago
+  central: America/Chicago
+  mt: America/Denver
+  mountain: America/Denver
+  pt: America/Los_Angeles
+  pacific: America/Los_Angeles
+
+fuzzy:          # D2
+  early: early
+  mid: mid
+  middle: mid
+  late: late
+  beginning: start
+  start: start
+  end: end
+  eod: eod
+  eom: eom
+  eoy: eoy
+  bom: bom
+  boy: boy
+
+article:        # "a" is also the meridiem in "8a"
+  a: 1
+  an: 1
+
+word:           # function words; the grammar gives them their meaning
+  workweek: work_week
+  during: during
+  all: all
+  rest: rest
+  remainder: rest
+  anytime: anytime
+  sometime: anytime
+  sometimes: anytime
+  in: in
+  within: within
+  ago: ago
+  back: ago
+  hence: hence
+  out: hence
+  later: later
+  earlier: ago
+  from: from
+  to: to
+  closest: closest  # D48
+  nearest: closest
+  and: and
+  or: or
+  of: of
+  the: the
+  "on": "on"
+  between: between
+  after: after
+  before: before
+  since: since
+  until: until
+  till: until
+  til: until
+  untill: until
+  through: through
+  thru: through
+  by: by
+  around: around
+  about: around
+  approximately: around
+  roughly: around
+  approx: around
+  circa: around
+  ish: around
+  o'clock: oclock
+  thousand: thousand
+  hundred: hundred
+  million: million
+  oh: oh            # "twelve oh three"
+  o: oh
+  zero: oh
+  oclock: oclock
+  half: half
+  for: for
+  more: more
+  extra: more
+  additional: more
+  t: t
+  q: q
+  fy: fy
+  ad: ad
+  be: be            # D48: the Buddhist era
+  b.e: be
+  a.d: ad
+  ce: ad
+  c.e: ad
+  starting: starting
+  onwards: orlater
+  onward: orlater
+  commencing: starting
+  every: every      # R5, R6: a rule; only Acute.recurrence reads these
+  each: every
+  other: other
+  alternate: other
+  once: once
+  times: times
+  occurrences: times
+  occurrence: times
+  except: except
+  excluding: except
+  everyday: daily
+  daily: daily
+  hourly: hourly
+  weekly: weekly
+  monthly: monthly
+  yearly: yearly
+  annually: yearly
+  annual: yearly
+  quarterly: quarterly
+  semiannually: semiannual
+  semiannual: semiannual
+  biweekly: biweekly
+  fortnightly: biweekly
+  bimonthly: bimonthly
+
+# Several words read as one token, longest first. Spaces in a key stand for any run of spaces or hyphens.
+phrases:
+  o clock: [word, oclock]
+  eastern time: [zone, America/New_York]
+  central time: [zone, America/Chicago]
+  mountain time: [zone, America/Denver]
+  pacific time: [zone, America/Los_Angeles]
+  alaska time: [zone, America/Anchorage]
+  hawaii time: [zone, Pacific/Honolulu]
+  week end: [weekend, 0]
+  lunch time: [part, lunch]
+  christmas eve: [holiday, christmas_eve]
+  christmas day: [holiday, christmas]
+  thanksgiving day: [holiday, thanksgiving]
+  new year's eve: [holiday, new_years_eve]
+  new years eve: [holiday, new_years_eve]
+  new year's day: [holiday, new_years_day]
+  new years day: [holiday, new_years_day]
+  valentine's day: [holiday, valentines]
+  valentines day: [holiday, valentines]
+  valentine day: [holiday, valentines]
+  st patrick's day: [holiday, st_patricks]
+  st patricks day: [holiday, st_patricks]
+  saint patrick's day: [holiday, st_patricks]
+  saint patricks day: [holiday, st_patricks]
+  saint patrick: [holiday, st_patricks]
+  april fools day: [holiday, april_fools]
+  april fool's day: [holiday, april_fools]
+  april fools' day: [holiday, april_fools]
+  independence day: [holiday, independence_day]
+  veterans day: [holiday, veterans_day]
+  veteran's day: [holiday, veterans_day]
+  veterans' day: [holiday, veterans_day]
+  veteran day: [holiday, veterans_day]
+  boxing day: [holiday, boxing_day]
+  groundhog day: [holiday, groundhog_day]
+  groundhogs day: [holiday, groundhog_day]
+  groundhog's day: [holiday, groundhog_day]
+  earth day: [holiday, earth_day]
+  flag day: [holiday, flag_day]
+  eastern standard time: [zone, -18000]
+  eastern daylight time: [zone, -14400]
+  central standard time: [zone, -21600]
+  central daylight time: [zone, -18000]
+  mountain standard time: [zone, -25200]
+  mountain daylight time: [zone, -21600]
+  pacific standard time: [zone, -28800]
+  pacific daylight time: [zone, -25200]
+  patriot day: [holiday, patriot_day]
+  boss's day: [holiday, bosses_day]
+  boss day: [holiday, bosses_day]
+  bosses day: [holiday, bosses_day]
+  bosses' day: [holiday, bosses_day]
+  australia day: [holiday, australia_day]
+  anzac day: [holiday, anzac_day]
+  canada day: [holiday, canada_day]
+  victoria day: [holiday, victoria_day]
+  melbourne cup day: [holiday, melbourne_cup]
+  melbourne cup: [holiday, melbourne_cup]
+  mlk day: [holiday, mlk_day]
+  mlk jr day: [holiday, mlk_day]
+  martin luther king day: [holiday, mlk_day]
+  martin luther king's day: [holiday, mlk_day]
+  martin luther kings day: [holiday, mlk_day]
+  martin luther king jr day: [holiday, mlk_day]
+  presidents day: [holiday, presidents_day]
+  president day: [holiday, presidents_day]
+  presidents' day: [holiday, presidents_day]
+  president's day: [holiday, presidents_day]
+  mother's day: [holiday, mothers_day]
+  mothers day: [holiday, mothers_day]
+  mothers' day: [holiday, mothers_day]
+  mothering sunday: [holiday, mothering_sunday]
+  memorial day: [holiday, memorial_day]
+  memorial days: [holiday, memorial_day]
+  father's day: [holiday, fathers_day]
+  fathers day: [holiday, fathers_day]
+  fathers' day: [holiday, fathers_day]
+  labor day: [holiday, labor_day]
+  labour day: [holiday, labor_day]
+  columbus day: [holiday, columbus_day]
+  black friday: [holiday, black_friday]
+  black fridays: [holiday, black_friday]
+  cyber monday: [holiday, cyber_monday]
+  easter sunday: [holiday, easter]
+  easter day: [holiday, easter]
+  good friday: [holiday, good_friday]
+  easter monday: [holiday, easter_monday]
+  easter mon: [holiday, easter_monday]
+  palm sunday: [holiday, palm_sunday]
+  maundy thursday: [holiday, maundy_thursday]
+  holy thursday: [holiday, maundy_thursday]
+  ash wednesday: [holiday, ash_wednesday]
+  mardi gras: [holiday, mardi_gras]
+  shrove tuesday: [holiday, mardi_gras]
+  whit sunday: [holiday, pentecost]
+  white sunday: [holiday, pentecost]
+  whit monday: [holiday, whit_monday]
+  trinity sunday: [holiday, trinity_sunday]
+  new year eve: [holiday, new_years_eve]
+  new year day: [holiday, new_years_day]
+  fourth of july: [holiday, independence_day]
+  st paddy's day: [holiday, st_patricks]
+  st paddys day: [holiday, st_patricks]
+  saint paddy's day: [holiday, st_patricks]
+  saint paddys day: [holiday, st_patricks]
+  st patty's day: [holiday, st_patricks]
+  st pattys day: [holiday, st_patricks]
+  saint patty's day: [holiday, st_patricks]
+  saint pattys day: [holiday, st_patricks]
+  thanks giving day: [holiday, thanksgiving]
+  thanks giving: [holiday, thanksgiving]
+  civil rights day: [holiday, mlk_day]
+  washington's birthday: [holiday, presidents_day]
+  washingtons birthday: [holiday, presidents_day]
+  washington birthday: [holiday, presidents_day]
+  george washington's birthday: [holiday, presidents_day]
+  george washington day: [holiday, presidents_day]
+  decoration day: [holiday, memorial_day]
+  covenant thursday: [holiday, maundy_thursday]
+  covenant thu: [holiday, maundy_thursday]
+  thursday of mysteries: [holiday, maundy_thursday]
+  thu of mysteries: [holiday, maundy_thursday]
+  monday of the holy spirit: [holiday, whit_monday]
+  branch sunday: [holiday, palm_sunday]
+  pancake day: [holiday, mardi_gras]
+  election day: [holiday, election_day]
+  lincoln's birthday: [holiday, lincolns_birthday]
+  lincolns birthday: [holiday, lincolns_birthday]
+  lincolns' birthday: [holiday, lincolns_birthday]
+  lincoln birthday: [holiday, lincolns_birthday]
+  abraham lincoln's birthday: [holiday, lincolns_birthday]
+  abraham lincolns birthday: [holiday, lincolns_birthday]
+  abraham lincoln birthday: [holiday, lincolns_birthday]
+  law day: [holiday, law_day]
+  loyalty day: [holiday, loyalty_day]
+  may day: [holiday, may_day]
+  workers day: [holiday, workers_day]
+  workers' day: [holiday, workers_day]
+  international workers day: [holiday, workers_day]
+  international workers' day: [holiday, workers_day]
+  int'l workers day: [holiday, workers_day]
+  int'l workers' day: [holiday, workers_day]
+  siblings day: [holiday, siblings_day]
+  sibling day: [holiday, siblings_day]
+  national siblings day: [holiday, siblings_day]
+  national sibling day: [holiday, siblings_day]
+  grandparents day: [holiday, grandparents_day]
+  grandparents' day: [holiday, grandparents_day]
+  grandparent's day: [holiday, grandparents_day]
+  national grandparents day: [holiday, grandparents_day]
+  national grandparents' day: [holiday, grandparents_day]
+  military spouse day: [holiday, military_spouse_day]
+  military spouses day: [holiday, military_spouse_day]
+  military spouse appreciation day: [holiday, military_spouse_day]
+  administrative professionals' day: [holiday, admin_day]
+  administrative professionals day: [holiday, admin_day]
+  administrative professional's day: [holiday, admin_day]
+  administrative professional day: [holiday, admin_day]
+  admin professionals day: [holiday, admin_day]
+  secretaries' day: [holiday, admin_day]
+  secretaries day: [holiday, admin_day]
+  secretary's day: [holiday, admin_day]
+  admins day: [holiday, admin_day]
+  admins' day: [holiday, admin_day]
+  admin day: [holiday, admin_day]
+  world vegan day: [holiday, world_vegan_day]
+  ides of march: [holiday, ides_of_march]
+  leap day: [holiday, leap_day]
+  leap year: [holiday, leap_year]
+  guy fawkes day: [holiday, guy_fawkes]
+  guy fawkes night: [holiday, guy_fawkes]
+  bonfire night: [holiday, guy_fawkes]
+  dominion day: [holiday, canada_day]
+  vimy ridge day: [holiday, vimy_ridge_day]
+  sovereign's birthday: [holiday, victoria_day]
+  rizal day: [holiday, rizal_day]
+  indian arrival day: [holiday, indian_arrival_day]
+  day of reconciliation: [holiday, reconciliation_day]
+  reconciliation day: [holiday, reconciliation_day]
+  day of the vow: [holiday, reconciliation_day]
+  day of the covenant: [holiday, reconciliation_day]
+  kruger day: [holiday, kruger_day]
+  garifuna settlement day: [holiday, garifuna_day]
+  business day: [unit, business_day]
+  business days: [unit, business_day]
+  working day: [unit, business_day]
+  working days: [unit, business_day]
+  working week: [word, work_week]   # D48
+  in between: [word, between]
+  work week: [word, work_week]
+  business week: [word, work_week]
+  early hours: [part, early_hours]
+  small hours: [part, early_hours]
+  a m: [meridiem, 0]
+  p m: [meridiem, 12]
+  prior to: [word, before]
+  later than or at: [word, onafter]
+  later than or in: [word, onafter]
+  later than or on: [word, onafter]
+  after or equal to: [word, onafter]
+  on or after: [word, onafter]
+  at or after: [word, onafter]
+  after or on: [word, onafter]
+  on or earlier than: [word, onbefore]
+  on or before: [word, onbefore]
+  at or before: [word, onbefore]
+  before or on: [word, onbefore]
+  before or in: [word, onbefore]
+  before or at: [word, onbefore]
+  no later than: [word, by]
+  later than: [word, after]
+  but before: [word, butbefore]
+  greater than: [word, after]
+  year greater than: [word, after]
+  earlier than: [word, before]
+  sooner than: [word, before]
+  in advance of: [word, before]
+  or later: [word, orlater]
+  and later: [word, orlater]
+  or greater: [word, orlater]
+  and greater: [word, orlater]
+  and beyond: [word, orlater]
+  or earlier: [word, orearlier]
+  up to: [word, upto]
+  right now: [now, 0]
+  just now: [now, 0]
+  at the moment: [now, 0]
+  at the minute: [now, 0]
+  at present: [now, 0]
+  at the present time: [now, 0]
+  at this time: [now, 0]
+  this time: [word, sametime]
+  same time: [word, sametime]
+  minutes time: [unit, minute]
+  minutes' time: [unit, minute]
+  hours time: [unit, hour]
+  hours' time: [unit, hour]
+  days time: [unit, day]
+  days' time: [unit, day]
+  weeks time: [unit, week]
+  weeks' time: [unit, week]
+  months time: [unit, month]
+  months' time: [unit, month]
+  years time: [unit, year]
+  years' time: [unit, year]
+  minute's time: [unit, minute]
+  hour's time: [unit, hour]
+  day's time: [unit, day]
+  week's time: [unit, week]
+  fortnight's time: [unit, fortnight]
+  month's time: [unit, month]
+  year's time: [unit, year]
+  a couple of: [word, couple]
+  a couple: [word, couple]
+  couple of: [word, couple]
+  a pair of: [word, couple]
+  fiscal year: [word, fy]
+  calendar year: [unit, year]
+  calendar month: [unit, month]
+  calendar week: [unit, week]
+  calendar quarter: [unit, quarter]
+  financial year: [word, fy]
+  semi annually: [word, semiannual]
+  semi annual: [word, semiannual]
+  bi weekly: [word, biweekly]
+
+# Words that are also plain English. In extraction they need a neighbour that confirms the date reading.
+english: [may, march, following, sat, sun, tu, fr, second, fall, a, mon, wed, start, end, last, past, this, next, d, m, s, h, w, y, lunch,
+          boy]
+
+# D44: plurals, which only count ("3 fridays from now"); alone they recur, which is not a time. Plural units
+# let "2 upcoming weeks" be "next 2 weeks" while "3 next week" is not.
+plural: [thanksgivings, springs, summers, autumns, falls, winters, seasons, sundays, mondays, tuesdays, wednesdays, thursdays, fridays, saturdays, weekends, weekdays, black fridays, memorial days,
+         secs, seconds, mins, minutes, hrs, hours, days, wks, weeks, fortnights, months, quarters, yrs, years,
+         decades, centuries, mornings, afternoons, evenings, nights]
+
+# D18: a fixed table, only misspellings the corpus has. Each maps to the word it stands for.
+misspellings:
+  tommorow: tomorrow
+  tonite: tonight
+  febuary: february
+  wensday: wednesday

data/exe/acute

@@ -0,0 +1,4 @@
+#!/usr/bin/env ruby
+require_relative "../lib/acute/cli"
+
+exit Acute::CLI.run(ARGV)

data/lib/acute/assumption.rb

@@ -0,0 +1,13 @@
+require_relative "tables/assumptions"
+
+module Acute
+  # A rule a parse applied: frozen and Ractor-shareable
+  Assumption = Data.define(:code, :rule, :message, :short)
+
+  class Assumption
+    # @api private
+    ALL = Tables::ASSUMPTIONS.to_h { |code, (rule, message, short)| [code, new(code, rule, message, short)] }.freeze
+
+    def self.for(code) = ALL.fetch(code)
+  end
+end