Add Quantity/ZH

Summary:
Fixes #181.
Closes https://github.com/facebook/duckling/pull/182

Reviewed By: patapizza

Differential Revision: D7891288

Pulled By: haoxuany

fbshipit-source-id: faa505dbf149c728545a92579cb2e92a89cd5cb8

Duckling/Dimensions/ZH.hs

@@ -18,6 +18,7 @@ allDimensions =
   , This Duration
   , This Numeral
   , This Ordinal
+  , This Quantity
   , This Temperature
   , This Time
   ]

Duckling/Quantity/ZH/Corpus.hs

@@ -0,0 +1,64 @@
+-- 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.Quantity.ZH.Corpus
+  ( corpus
+  ) where
+
+import Data.String
+import Prelude
+
+import Duckling.Locale
+import Duckling.Quantity.Types
+import Duckling.Resolve
+import Duckling.Testing.Types
+
+corpus :: Corpus
+corpus = (testContext {locale = makeLocale ZH Nothing},
+  testOptions, allExamples)
+
+allExamples :: [Example]
+allExamples = concat
+  [ examples (simple Gram 2 Nothing)
+             [ "2克"
+             , "二克"
+             , "两克"
+             , "2g"
+             , "2.0g"
+             ]
+  , examples (simple Gram 1500 Nothing)
+             [ "3斤"
+             , "三斤"
+             , "1.5公斤"
+             , "1.5千克"
+             , "1.5kg"
+             , "1500g"
+             ]
+  , examples (simple Gram 0.035 Nothing)
+             [ "35毫克"
+             , "三十五毫克"
+             , "35mg"
+             ]
+  , examples (simple Gram 1750 Nothing)
+             [ "三斤半"
+             , "3斤半"
+             , "3斤5两"
+             , "1.75千克"
+             ]
+  , examples (simple Gram 475 Nothing)
+             [ "9两半"
+             , "九兩半"
+             ]
+  , examples (simple Gram 3400 Nothing)
+             [ "六斤八两"
+             , "六斤8兩"
+             , "3.4公斤"
+             ]
+  ]

Duckling/Quantity/ZH/Rules.hs

@@ -0,0 +1,122 @@
+-- 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 LambdaCase #-}
+{-# LANGUAGE OverloadedStrings #-}
+
+module Duckling.Quantity.ZH.Rules
+  ( rules
+  ) where
+
+import Data.HashMap.Strict (HashMap)
+import Data.String
+import Data.Text (Text)
+import Prelude
+import qualified Data.HashMap.Strict as HashMap
+import qualified Data.Text as Text
+
+import Duckling.Dimensions.Types
+import Duckling.Numeral.Helpers
+import Duckling.Quantity.Helpers
+import Duckling.Regex.Types
+import Duckling.Types
+import qualified Duckling.Numeral.Types as TNumeral
+import qualified Duckling.Quantity.Types as TQuantity
+
+quantities :: [(Text, String, TQuantity.Unit)]
+quantities =
+  [ ("<quantity> grams", "(千克|公斤|斤|两|兩|克|毫克|kg|g|mg)", TQuantity.Gram)
+  ]
+
+opsMap :: HashMap Text (Double -> Double)
+opsMap = HashMap.fromList
+  [ ( "千克" , (* 1000))
+  , ( "公斤",  (* 1000))
+  , ( "kg",   (* 1000))
+  , ( "斤",    (* 500))
+  , ( "两",    (* 50))
+  , ( "兩",    (* 50))
+  , ( "毫克",  (/ 1000))
+  , ( "mg",   (/ 1000))
+  ]
+
+ruleNumeralQuantities :: [Rule]
+ruleNumeralQuantities = map go quantities
+  where
+    getValue :: Text -> Double -> Double
+    getValue match = HashMap.lookupDefault id (Text.toLower match) opsMap
+
+    go :: (Text, String, TQuantity.Unit) -> Rule
+    go (name, regexPattern, u) = Rule
+      { name = name
+      , pattern =
+        [ Predicate isPositive
+        , regex regexPattern
+        ]
+      , prod = \case
+        (Token Numeral nd:
+         Token RegexMatch (GroupMatch (match:_)):
+         _) -> Just . Token Quantity $ quantity u value
+          where value = getValue match $ TNumeral.value nd
+        _ -> Nothing
+      }
+
+ruleCattyTael :: Rule
+ruleCattyTael = Rule
+  { name = "<quantity> catty <quantity> tael"
+  , pattern =
+    [ Predicate isPositive
+    , regex "斤"
+    , numberBetween 1 10
+    , regex "两|兩"
+    ]
+  , prod = \case
+    (Token Numeral TNumeral.NumeralData{TNumeral.value = x}:
+     Token RegexMatch (GroupMatch _):
+     Token Numeral TNumeral.NumeralData{TNumeral.value = y}:
+     Token RegexMatch (GroupMatch _):
+     _) -> Just . Token Quantity $ quantity TQuantity.Gram (x * 500 + y * 50)
+    _ -> Nothing
+  }
+
+ruleCattyHalf :: Rule
+ruleCattyHalf = Rule
+  { name = "<quantity> catty half"
+  , pattern =
+    [ Predicate isPositive
+    , regex "斤半"
+    ]
+  , prod = \case
+    (Token Numeral TNumeral.NumeralData{TNumeral.value = x}:
+     Token RegexMatch (GroupMatch _):
+     _) -> Just . Token Quantity $ quantity TQuantity.Gram (x * 500 + 250)
+    _ -> Nothing
+  }
+
+ruleTaelHalf :: Rule
+ruleTaelHalf = Rule
+  { name = "<quantity> tael half"
+  , pattern =
+    [ Predicate isPositive
+    , regex "(两|兩)半"
+    ]
+  , prod = \case
+    (Token Numeral TNumeral.NumeralData{TNumeral.value = x}:
+     Token RegexMatch (GroupMatch _):
+     _) -> Just . Token Quantity $ quantity TQuantity.Gram (x * 50 + 25)
+    _ -> Nothing
+  }
+
+rules :: [Rule]
+rules =
+  [ ruleCattyTael
+  , ruleCattyHalf
+  , ruleTaelHalf
+  ]
+  ++ ruleNumeralQuantities

Duckling/Rules/ZH.hs

@@ -22,6 +22,7 @@ import qualified Duckling.AmountOfMoney.ZH.Rules as AmountOfMoney
 import qualified Duckling.Distance.ZH.Rules as Distance
 import qualified Duckling.Numeral.ZH.Rules as Numeral
 import qualified Duckling.Ordinal.ZH.Rules as Ordinal
+import qualified Duckling.Quantity.ZH.Rules as Quantity
 import qualified Duckling.Temperature.ZH.Rules as Temperature
 import qualified Duckling.Time.ZH.Rules as Time
 import qualified Duckling.Time.ZH.CN.Rules as TimeCN
@@ -49,7 +50,7 @@ langRules (This Email) = []
 langRules (This Numeral) = Numeral.rules
 langRules (This Ordinal) = Ordinal.rules
 langRules (This PhoneNumber) = []
-langRules (This Quantity) = []
+langRules (This Quantity) = Quantity.rules
 langRules (This RegexMatch) = []
 langRules (This Temperature) = Temperature.rules
 langRules (This Time) = Time.rules

duckling.cabal

@@ -452,6 +452,8 @@ library
                      , Duckling.Quantity.RO.Rules
                      , Duckling.Quantity.RU.Corpus
                      , Duckling.Quantity.RU.Rules
+                     , Duckling.Quantity.ZH.Corpus
+                     , Duckling.Quantity.ZH.Rules
                      , Duckling.Quantity.Helpers
                      , Duckling.Quantity.Types
 
@@ -824,6 +826,7 @@ test-suite duckling-test
                      , Duckling.Quantity.PT.Tests
                      , Duckling.Quantity.RO.Tests
                      , Duckling.Quantity.RU.Tests
+                     , Duckling.Quantity.ZH.Tests
                      , Duckling.Quantity.Tests
 
                      -- Temperature

tests/Duckling/Quantity/Tests.hs

@@ -22,6 +22,7 @@ import qualified Duckling.Quantity.KO.Tests as KO
 import qualified Duckling.Quantity.PT.Tests as PT
 import qualified Duckling.Quantity.RO.Tests as RO
 import qualified Duckling.Quantity.RU.Tests as RU
+import qualified Duckling.Quantity.ZH.Tests as ZH
 
 tests :: TestTree
 tests = testGroup "Quantity Tests"
@@ -33,4 +34,5 @@ tests = testGroup "Quantity Tests"
   , PT.tests
   , RO.tests
   , RU.tests
+  , ZH.tests
   ]

tests/Duckling/Quantity/ZH/Tests.hs

@@ -0,0 +1,23 @@
+-- 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.
+
+
+module Duckling.Quantity.ZH.Tests
+  ( tests ) where
+
+import Data.String
+import Prelude
+import Test.Tasty
+
+import Duckling.Dimensions.Types
+import Duckling.Quantity.ZH.Corpus
+import Duckling.Testing.Asserts
+
+tests :: TestTree
+tests = testGroup "ZH Tests"
+  [ makeCorpusTest [This Quantity] corpus
+  ]