en_CA + fix Canadian Thanksgiving

patapizza · facebook-github-bot · commit 1ab5f447d263 · 2017-10-16T10:04:43.000-07:00
Summary:
* `en_CA` locale
* In Canada, Thanksgiving Day is the second Monday of October.
* Black Friday is the same as the US.
* However Canada observes both DDMM and MMDD formats. Defer to later, falling back to US.

Reviewed By: blandinw

Differential Revision: D6058909

fbshipit-source-id: 3d4e05e
diff --git a/Duckling/Locale.hs b/Duckling/Locale.hs
@@ -69,7 +69,8 @@ instance TextShow Lang where
 -- | ISO 3166-1 alpha-2 Country code (includes regions and territories).
 -- See https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2
 data Region
-  = CN
+  = CA
+  | CN
   | GB
   | HK
   | MO
@@ -100,6 +101,6 @@ makeLocale lang (Just region)
 
 allLocales :: HashMap Lang (HashSet Region)
 allLocales = HashMap.fromList
-  [ (EN, HashSet.fromList [GB, US])
+  [ (EN, HashSet.fromList [CA, GB, US])
   , (ZH, HashSet.fromList [CN, HK, MO, TW])
   ]
diff --git a/Duckling/Ranking/Classifiers/EN_CA.hs b/Duckling/Ranking/Classifiers/EN_CA.hs
diff --git a/Duckling/Rules/EN.hs b/Duckling/Rules/EN.hs
@@ -29,6 +29,7 @@ import qualified Duckling.Ordinal.EN.Rules as Ordinal
 import qualified Duckling.Quantity.EN.Rules as Quantity
 import qualified Duckling.Temperature.EN.Rules as Temperature
 import qualified Duckling.Time.EN.Rules as Time
+import qualified Duckling.Time.EN.CA.Rules as TimeCA
 import qualified Duckling.Time.EN.GB.Rules as TimeGB
 import qualified Duckling.Time.EN.US.Rules as TimeUS
 import qualified Duckling.TimeGrain.EN.Rules as TimeGrain
@@ -38,10 +39,12 @@ defaultRules :: Some Dimension -> [Rule]
 defaultRules dim@(This Time) =
   [ TimeUS.ruleMMDD
   , TimeUS.ruleMMDDYYYY
+  , TimeUS.ruleThanksgiving
   ] ++ langRules dim
 defaultRules dim = langRules dim
 
 localeRules :: Region -> Some Dimension -> [Rule]
+localeRules CA (This Time) = TimeCA.rules
 localeRules GB (This Time) = TimeGB.rules
 localeRules US (This Time) = TimeUS.rules
 localeRules _ _            = []
diff --git a/Duckling/Time/EN/CA/Corpus.hs b/Duckling/Time/EN/CA/Corpus.hs
@@ -0,0 +1,60 @@
+-- Copyright (c) 2016-present, Facebook, Inc.
+-- All rights reserved.
+--
+-- This source code is licensed under the BSD-style license found in the
+-- LICENSE file in the root directory of this source tree. An additional grant
+-- of patent rights can be found in the PATENTS file in the same directory.
+
+{-# LANGUAGE OverloadedStrings #-}
+
+module Duckling.Time.EN.CA.Corpus
+  ( allExamples
+  ) where
+
+import Data.String
+import Prelude
+
+import Duckling.Testing.Types hiding (examples)
+import Duckling.Time.Corpus
+import Duckling.Time.Types hiding (Month)
+import Duckling.TimeGrain.Types hiding (add)
+
+allExamples :: [Example]
+allExamples = concat
+  [ examples (datetime (2013, 2, 15, 0, 0, 0) Day)
+             [ "2/15"
+             , "on 2/15"
+             , "2 / 15"
+             , "2-15"
+             , "2 - 15"
+             ]
+  , examples (datetime (1974, 10, 31, 0, 0, 0) Day)
+             [ "10/31/1974"
+             , "10/31/74"
+             , "10-31-74"
+             ]
+  , examples (datetime (2013, 4, 25, 16, 0, 0) Minute)
+             [ "4/25 at 4:00pm"
+             ]
+  , examples (datetime (2013, 10, 14, 0, 0, 0) Day)
+             [ "thanksgiving day"
+             , "thanksgiving"
+             , "thanksgiving 2013"
+             , "this thanksgiving"
+             , "next thanksgiving day"
+             ]
+  , examples (datetime (2014, 10, 13, 0, 0, 0) Day)
+             [ "thanksgiving of next year"
+             , "thanksgiving 2014"
+             ]
+  , examples (datetime (2012, 10, 8, 0, 0, 0) Day)
+             [ "last thanksgiving"
+             , "thanksgiving day 2012"
+             ]
+  , examples (datetime (2016, 10, 10, 0, 0, 0) Day)
+             [ "thanksgiving 2016"
+             ]
+  , examples (datetime (2017, 10, 9, 0, 0, 0) Day)
+             [ "thanksgiving 2017"
+             ]
+  ]
diff --git a/Duckling/Time/EN/CA/Rules.hs b/Duckling/Time/EN/CA/Rules.hs
@@ -0,0 +1,69 @@
+-- Copyright (c) 2016-present, Facebook, Inc.
+-- All rights reserved.
+--
+-- This source code is licensed under the BSD-style license found in the
+-- LICENSE file in the root directory of this source tree. An additional grant
+-- of patent rights can be found in the PATENTS file in the same directory.
+
+
+{-# LANGUAGE GADTs #-}
+{-# LANGUAGE NoRebindableSyntax #-}
+{-# LANGUAGE OverloadedStrings #-}
+
+module Duckling.Time.EN.CA.Rules
+  ( rules
+  ) where
+
+import Data.Maybe
+import Prelude
+
+import Duckling.Dimensions.Types
+import Duckling.Numeral.Helpers (parseInt)
+import Duckling.Regex.Types
+import Duckling.Time.Helpers
+import Duckling.Time.Types (TimeData (..))
+import Duckling.Types
+
+-- Although one can see both MMDD and DDMM in Canada,
+-- there is no direct way to implement this today. Let's fallback to MMDD (US).
+ruleMMDD :: Rule
+ruleMMDD = Rule
+  { name = "mm/dd"
+  , pattern = [regex "(0?[1-9]|1[0-2])\\s?[/-]\\s?(3[01]|[12]\\d|0?[1-9])"]
+  , prod = \tokens -> case tokens of
+      (Token RegexMatch (GroupMatch (mm:dd:_)):_) -> do
+        m <- parseInt mm
+        d <- parseInt dd
+        tt $ monthDay m d
+      _ -> Nothing
+  }
+
+ruleMMDDYYYY :: Rule
+ruleMMDDYYYY = Rule
+  { name = "mm/dd/yyyy"
+  , pattern =
+    [regex "(0?[1-9]|1[0-2])[/-](3[01]|[12]\\d|0?[1-9])[-/](\\d{2,4})"]
+  , prod = \tokens -> case tokens of
+      (Token RegexMatch (GroupMatch (mm:dd:yy:_)):_) -> do
+        y <- parseInt yy
+        m <- parseInt mm
+        d <- parseInt dd
+        tt $ yearMonthDay y m d
+      _ -> Nothing
+  }
+
+ruleThanksgiving :: Rule
+ruleThanksgiving = Rule
+  { name = "Thanksgiving Day"
+  , pattern =
+    [ regex "thanks?giving( day)?"
+    ]
+  , prod = \_ -> tt $ nthDOWOfMonth 2 1 10 -- Second Monday of October
+  }
+
+rules :: [Rule]
+rules =
+  [ ruleMMDD
+  , ruleMMDDYYYY
+  , ruleThanksgiving
+  ]
diff --git a/Duckling/Time/EN/Corpus.hs b/Duckling/Time/EN/Corpus.hs
@@ -44,6 +44,27 @@ defaultCorpus = (testContext, allExamples ++ custom)
       , examples (datetime (2013, 4, 25, 16, 0, 0) Minute)
                  [ "4/25 at 4:00pm"
                  ]
+      , examples (datetime (2013, 11, 28, 0, 0, 0) Day)
+                 [ "thanksgiving day"
+                 , "thanksgiving"
+                 , "thanksgiving 2013"
+                 , "this thanksgiving"
+                 , "next thanksgiving day"
+                 ]
+      , examples (datetime (2014, 11, 27, 0, 0, 0) Day)
+                 [ "thanksgiving of next year"
+                 , "thanksgiving 2014"
+                 ]
+      , examples (datetime (2012, 11, 22, 0, 0, 0) Day)
+                 [ "last thanksgiving"
+                 , "thanksgiving day 2012"
+                 ]
+      , examples (datetime (2016, 11, 24, 0, 0, 0) Day)
+                 [ "thanksgiving 2016"
+                 ]
+      , examples (datetime (2017, 11, 23, 0, 0, 0) Day)
+                 [ "thanksgiving 2017"
+                 ]
       ]
 
 negativeCorpus :: NegativeCorpus
@@ -613,27 +634,6 @@ allExamples = concat
              , "next halloween"
              , "Halloween 2013"
              ]
-  , examples (datetime (2013, 11, 28, 0, 0, 0) Day)
-             [ "thanksgiving day"
-             , "thanksgiving"
-             , "thanksgiving 2013"
-             , "this thanksgiving"
-             , "next thanksgiving day"
-             ]
-  , examples (datetime (2014, 11, 27, 0, 0, 0) Day)
-             [ "thanksgiving of next year"
-             , "thanksgiving 2014"
-             ]
-  , examples (datetime (2012, 11, 22, 0, 0, 0) Day)
-             [ "last thanksgiving"
-             , "thanksgiving day 2012"
-             ]
-  , examples (datetime (2016, 11, 24, 0, 0, 0) Day)
-             [ "thanksgiving 2016"
-             ]
-  , examples (datetime (2017, 11, 23, 0, 0, 0) Day)
-             [ "thanksgiving 2017"
-             ]
   , examples (datetime (2013, 11, 29, 0, 0, 0) Day)
              [ "black friday"
              , "black friday of this year"
diff --git a/Duckling/Time/EN/GB/Corpus.hs b/Duckling/Time/EN/GB/Corpus.hs
@@ -36,4 +36,25 @@ allExamples = concat
   , examples (datetime (2013, 4, 25, 16, 0, 0) Minute)
              [ "25/4 at 4:00pm"
              ]
+  , examples (datetime (2013, 11, 28, 0, 0, 0) Day)
+             [ "thanksgiving day"
+             , "thanksgiving"
+             , "thanksgiving 2013"
+             , "this thanksgiving"
+             , "next thanksgiving day"
+             ]
+  , examples (datetime (2014, 11, 27, 0, 0, 0) Day)
+             [ "thanksgiving of next year"
+             , "thanksgiving 2014"
+             ]
+  , examples (datetime (2012, 11, 22, 0, 0, 0) Day)
+             [ "last thanksgiving"
+             , "thanksgiving day 2012"
+             ]
+  , examples (datetime (2016, 11, 24, 0, 0, 0) Day)
+             [ "thanksgiving 2016"
+             ]
+  , examples (datetime (2017, 11, 23, 0, 0, 0) Day)
+             [ "thanksgiving 2017"
+             ]
   ]
diff --git a/Duckling/Time/EN/GB/Rules.hs b/Duckling/Time/EN/GB/Rules.hs
@@ -53,8 +53,18 @@ ruleDDMMYYYY = Rule
       _ -> Nothing
   }
 
+ruleThanksgiving :: Rule
+ruleThanksgiving = Rule
+  { name = "Thanksgiving Day"
+  , pattern =
+    [ regex "thanks?giving( day)?"
+    ]
+  , prod = \_ -> tt $ nthDOWOfMonth 4 4 11 -- Fourth Thursday of November
+  }
+
 rules :: [Rule]
 rules =
   [ ruleDDMM
   , ruleDDMMYYYY
+  , ruleThanksgiving
   ]
diff --git a/Duckling/Time/EN/Rules.hs b/Duckling/Time/EN/Rules.hs
@@ -1264,10 +1264,6 @@ moreUSHolidays =
     , "mother'?s?'? day"
     , 2, 7, 5
     )
-  , ( "Thanksgiving Day" -- Fourth Thursday of November
-    , "thanks?giving( day)?"
-    , 4, 4, 11
-    )
   ,  ( "Labor Day" -- First Monday of September
      , "labor day"
      , 1, 1, 9
diff --git a/Duckling/Time/EN/US/Corpus.hs b/Duckling/Time/EN/US/Corpus.hs
@@ -36,4 +36,25 @@ allExamples = concat
   , examples (datetime (2013, 4, 25, 16, 0, 0) Minute)
              [ "4/25 at 4:00pm"
              ]
+  , examples (datetime (2013, 11, 28, 0, 0, 0) Day)
+             [ "thanksgiving day"
+             , "thanksgiving"
+             , "thanksgiving 2013"
+             , "this thanksgiving"
+             , "next thanksgiving day"
+             ]
+  , examples (datetime (2014, 11, 27, 0, 0, 0) Day)
+             [ "thanksgiving of next year"
+             , "thanksgiving 2014"
+             ]
+  , examples (datetime (2012, 11, 22, 0, 0, 0) Day)
+             [ "last thanksgiving"
+             , "thanksgiving day 2012"
+             ]
+  , examples (datetime (2016, 11, 24, 0, 0, 0) Day)
+             [ "thanksgiving 2016"
+             ]
+  , examples (datetime (2017, 11, 23, 0, 0, 0) Day)
+             [ "thanksgiving 2017"
+             ]
   ]
diff --git a/Duckling/Time/EN/US/Rules.hs b/Duckling/Time/EN/US/Rules.hs
@@ -14,6 +14,7 @@ module Duckling.Time.EN.US.Rules
   ( rules
   , ruleMMDD
   , ruleMMDDYYYY
+  , ruleThanksgiving
   ) where
 
 import Data.Maybe
@@ -52,8 +53,18 @@ ruleMMDDYYYY = Rule
       _ -> Nothing
   }
 
+ruleThanksgiving :: Rule
+ruleThanksgiving = Rule
+  { name = "Thanksgiving Day"
+  , pattern =
+    [ regex "thanks?giving( day)?"
+    ]
+  , prod = \_ -> tt $ nthDOWOfMonth 4 4 11 -- Fourth Thursday of November
+  }
+
 rules :: [Rule]
 rules =
   [ ruleMMDD
   , ruleMMDDYYYY
+  , ruleThanksgiving
   ]
diff --git a/duckling.cabal b/duckling.cabal
@@ -82,6 +82,7 @@ library
                      , Duckling.Ranking.Classifiers.CS_XX
                      , Duckling.Ranking.Classifiers.DA_XX
                      , Duckling.Ranking.Classifiers.DE_XX
+                     , Duckling.Ranking.Classifiers.EN_CA
                      , Duckling.Ranking.Classifiers.EN_GB
                      , Duckling.Ranking.Classifiers.EN_US
                      , Duckling.Ranking.Classifiers.EN_XX
@@ -436,6 +437,8 @@ library
                      , Duckling.Time.DE.Rules
                      , Duckling.Time.EN.Corpus
                      , Duckling.Time.EN.Rules
+                     , Duckling.Time.EN.CA.Corpus
+                     , Duckling.Time.EN.CA.Rules
                      , Duckling.Time.EN.GB.Corpus
                      , Duckling.Time.EN.GB.Rules
                      , Duckling.Time.EN.US.Corpus
diff --git a/exe/Duckling/Ranking/Generate.hs b/exe/Duckling/Ranking/Generate.hs
@@ -30,6 +30,7 @@ import Duckling.Testing.Types
 import qualified Duckling.Time.DA.Corpus as DATime
 import qualified Duckling.Time.DE.Corpus as DETime
 import qualified Duckling.Time.EN.Corpus as ENTime
+import qualified Duckling.Time.EN.CA.Corpus as EN_CATime
 import qualified Duckling.Time.EN.GB.Corpus as EN_GBTime
 import qualified Duckling.Time.EN.US.Corpus as EN_USTime
 import qualified Duckling.Time.ES.Corpus as ESTime
@@ -188,6 +189,7 @@ getCorpusForLang VI = VITime.corpus
 getCorpusForLang ZH = ZHTime.corpus
 
 getExamplesForLocale :: Lang -> Region -> [Example]
+getExamplesForLocale EN CA = EN_CATime.allExamples
 getExamplesForLocale EN GB = EN_GBTime.allExamples
 getExamplesForLocale EN US = EN_USTime.allExamples
 getExamplesForLocale ZH CN = ZH_CNTime.allExamples
diff --git a/tests/Duckling/Time/EN/GB/Tests.hs b/tests/Duckling/Time/EN/GB/Tests.hs
diff --git a/tests/Duckling/Time/EN/Tests.hs b/tests/Duckling/Time/EN/Tests.hs
diff --git a/tests/Duckling/Time/EN/US/Tests.hs b/tests/Duckling/Time/EN/US/Tests.hs
