-- 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.


{-# LANGUAGE GADTs #-}
{-# LANGUAGE OverloadedStrings #-}

module Duckling.Numeral.ID.Rules
  ( rules
  ) where

import Data.HashMap.Strict (HashMap)
import Data.Maybe
import Data.Text (Text)
import Prelude
import Data.String
import qualified Data.HashMap.Strict as HashMap
import qualified Data.Text as Text

import Duckling.Dimensions.Types
import Duckling.Numeral.Helpers
import Duckling.Numeral.Types (NumeralData (..))
import Duckling.Regex.Types
import Duckling.Types
import qualified Duckling.Numeral.Types as TNumeral

ruleIntegerMap :: HashMap Text Integer
ruleIntegerMap :: HashMap Text Integer
ruleIntegerMap = [(Text, Integer)] -> HashMap Text Integer
forall k v. (Eq k, Hashable k) => [(k, v)] -> HashMap k v
HashMap.fromList
  [ ( Text
"kosong"  , Integer
0 )
  , ( Text
"nol"     , Integer
0 )
  , ( Text
"satu"    , Integer
1 )
  , ( Text
"dua"     , Integer
2 )
  , ( Text
"tiga"    , Integer
3 )
  , ( Text
"empat"   , Integer
4 )
  , ( Text
"lima"    , Integer
5 )
  , ( Text
"enam"    , Integer
6 )
  , ( Text
"tujuh"   , Integer
7 )
  , ( Text
"delapan" , Integer
8 )
  , ( Text
"sembilan", Integer
9 )
  , ( Text
"sebelas" , Integer
11 )
  ]

ruleTeen :: Rule
ruleTeen :: Rule
ruleTeen = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"teen"
  , pattern :: Pattern
pattern =
    [ Double -> Double -> PatternItem
numberBetween Double
2 Double
10
    , String -> PatternItem
regex String
"belas"
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
Numeral NumeralData{TNumeral.value = v}:[Token]
_) -> Double -> Maybe Token
double (Double -> Maybe Token) -> Double -> Maybe Token
forall a b. (a -> b) -> a -> b
$ Double
v Double -> Double -> Double
forall a. Num a => a -> a -> a
+ Double
10
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleNumeralCommaNumeral :: Rule
ruleNumeralCommaNumeral :: Rule
ruleNumeralCommaNumeral = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"number comma number"
  , pattern :: Pattern
pattern =
    [ Dimension NumeralData -> PatternItem
forall a. Typeable a => Dimension a -> PatternItem
dimension Dimension NumeralData
Numeral
    , String -> PatternItem
regex String
"koma"
    , Predicate -> PatternItem
Predicate (Predicate -> PatternItem) -> Predicate -> PatternItem
forall a b. (a -> b) -> a -> b
$ Bool -> Bool
not (Bool -> Bool) -> Predicate -> Predicate
forall b c a. (b -> c) -> (a -> b) -> a -> c
. Predicate
hasGrain
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
Numeral a
nd1:Token
_:Token Dimension a
Numeral a
nd2:[Token]
_) ->
        Double -> Maybe Token
double (Double -> Maybe Token) -> Double -> Maybe Token
forall a b. (a -> b) -> a -> b
$ NumeralData -> Double
TNumeral.value a
NumeralData
nd1 Double -> Double -> Double
forall a. Num a => a -> a -> a
+ Double -> Double
decimalsToDouble (NumeralData -> Double
TNumeral.value a
NumeralData
nd2)
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleNumeralsPrefixWithNegativeOrMinus :: Rule
ruleNumeralsPrefixWithNegativeOrMinus :: Rule
ruleNumeralsPrefixWithNegativeOrMinus = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"numbers prefix with -, negative or minus"
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"-|minus|negatif"
    , Predicate -> PatternItem
Predicate Predicate
isPositive
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token
_:Token Dimension a
Numeral a
nd:[Token]
_) -> Double -> Maybe Token
double (NumeralData -> Double
TNumeral.value a
NumeralData
nd Double -> Double -> Double
forall a. Num a => a -> a -> a
* (-Double
1))
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleTen :: Rule
ruleTen :: Rule
ruleTen = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"ten"
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"(se)?puluh"
    ]
  , prod :: Production
prod = \[Token]
_ -> Integer -> Maybe Token
integer Integer
10 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Int -> Token -> Maybe Token
withGrain Int
1
  }

ruleDecimalWithThousandsSeparator :: Rule
ruleDecimalWithThousandsSeparator :: Rule
ruleDecimalWithThousandsSeparator = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"decimal with thousands separator"
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"(\\d+(\\.\\d\\d\\d)+,\\d+)"
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
RegexMatch (GroupMatch (match:_)):
       [Token]
_) -> let fmt :: Text
fmt = Text -> Text -> Text -> Text
Text.replace Text
"," Text
"." (Text -> Text) -> Text -> Text
forall a b. (a -> b) -> a -> b
$ Text -> Text -> Text -> Text
Text.replace Text
"." Text
Text.empty Text
match
        in Text -> Maybe Double
parseDouble Text
fmt Maybe Double -> (Double -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Double -> Maybe Token
double
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleMultiply :: Rule
ruleMultiply :: Rule
ruleMultiply = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"compose by multiplication"
  , pattern :: Pattern
pattern =
    [ Dimension NumeralData -> PatternItem
forall a. Typeable a => Dimension a -> PatternItem
dimension Dimension NumeralData
Numeral
    , Predicate -> PatternItem
Predicate Predicate
isMultipliable
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token
token1:Token
token2:[Token]
_) -> Token -> Token -> Maybe Token
multiply Token
token1 Token
token2
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }


ruleDecimalNumeral :: Rule
ruleDecimalNumeral :: Rule
ruleDecimalNumeral = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"decimal number"
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"(\\d*,\\d+)"
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
RegexMatch (GroupMatch (match:_)):[Token]
_) -> Bool -> Text -> Maybe Token
parseDecimal Bool
False Text
match
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleInteger3 :: Rule
ruleInteger3 :: Rule
ruleInteger3 = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"integer 21..99"
  , pattern :: Pattern
pattern =
    [ [Double] -> PatternItem
oneOf [Double
70, Double
20, Double
60, Double
50, Double
40, Double
90, Double
30, Double
80]
    , Double -> Double -> PatternItem
numberBetween Double
1 Double
10
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
Numeral NumeralData{TNumeral.value = v1}:
       Token Dimension a
Numeral NumeralData{TNumeral.value = v2}:
       [Token]
_) -> Double -> Maybe Token
double (Double -> Maybe Token) -> Double -> Maybe Token
forall a b. (a -> b) -> a -> b
$ Double
v1 Double -> Double -> Double
forall a. Num a => a -> a -> a
+ Double
v2
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleIntersect :: Rule
ruleIntersect :: Rule
ruleIntersect = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"intersect"
  , pattern :: Pattern
pattern =
    [ Predicate -> PatternItem
Predicate Predicate
hasGrain
    , Predicate -> PatternItem
Predicate (Predicate -> PatternItem) -> Predicate -> PatternItem
forall a b. (a -> b) -> a -> b
$ [Bool] -> Bool
forall (t :: * -> *). Foldable t => t Bool -> Bool
and ([Bool] -> Bool) -> (Token -> [Bool]) -> Predicate
forall b c a. (b -> c) -> (a -> b) -> a -> c
. [Predicate] -> Token -> [Bool]
forall (t :: * -> *) (m :: * -> *) a.
(Traversable t, Monad m) =>
t (m a) -> m (t a)
sequence [Bool -> Bool
not (Bool -> Bool) -> Predicate -> Predicate
forall b c a. (b -> c) -> (a -> b) -> a -> c
. Predicate
isMultipliable, Predicate
isPositive]
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
Numeral NumeralData{TNumeral.value = val1, TNumeral.grain = Just g}:
       Token Dimension a
Numeral NumeralData{TNumeral.value = val2}:
       [Token]
_) | (Double
10 Double -> Double -> Double
forall a. Floating a => a -> a -> a
** Int -> Double
forall a b. (Integral a, Num b) => a -> b
fromIntegral Int
g) Double -> Double -> Bool
forall a. Ord a => a -> a -> Bool
> Double
val2 -> Double -> Maybe Token
double (Double -> Maybe Token) -> Double -> Maybe Token
forall a b. (a -> b) -> a -> b
$ Double
val1 Double -> Double -> Double
forall a. Num a => a -> a -> a
+ Double
val2
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleSomefewcouple :: Rule
ruleSomefewcouple :: Rule
ruleSomefewcouple = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"some/few/couple"
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"beberapa"
    ]
  , prod :: Production
prod = \[Token]
_ -> Integer -> Maybe Token
integer Integer
3
  }

ruleNumeralsSuffixesKMG :: Rule
ruleNumeralsSuffixesKMG :: Rule
ruleNumeralsSuffixesKMG = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"numbers suffixes (K, M, G)"
  , pattern :: Pattern
pattern =
    [ Dimension NumeralData -> PatternItem
forall a. Typeable a => Dimension a -> PatternItem
dimension Dimension NumeralData
Numeral
    , String -> PatternItem
regex String
"([kmg])(?=[\\W\\$€]|$)"
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
Numeral NumeralData{TNumeral.value = v}:
       Token Dimension a
RegexMatch (GroupMatch (match:_)):
       [Token]
_) -> case Text -> Text
Text.toLower Text
match of
         Text
"k" -> Double -> Maybe Token
double (Double -> Maybe Token) -> Double -> Maybe Token
forall a b. (a -> b) -> a -> b
$ Double
v Double -> Double -> Double
forall a. Num a => a -> a -> a
* Double
1e3
         Text
"m" -> Double -> Maybe Token
double (Double -> Maybe Token) -> Double -> Maybe Token
forall a b. (a -> b) -> a -> b
$ Double
v Double -> Double -> Double
forall a. Num a => a -> a -> a
* Double
1e6
         Text
"g" -> Double -> Maybe Token
double (Double -> Maybe Token) -> Double -> Maybe Token
forall a b. (a -> b) -> a -> b
$ Double
v Double -> Double -> Double
forall a. Num a => a -> a -> a
* Double
1e9
         Text
_   -> Maybe Token
forall a. Maybe a
Nothing
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

rulePowersOfTen :: Rule
rulePowersOfTen :: Rule
rulePowersOfTen = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"powers of tens"
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"(se)?(ratus|ribu|juta)"
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
RegexMatch (GroupMatch (_:match:_)):[Token]
_) -> case Text -> Text
Text.toLower Text
match of
        Text
"ratus"-> Double -> Maybe Token
double Double
1e2 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Int -> Token -> Maybe Token
withGrain Int
2 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Token -> Maybe Token
withMultipliable
        Text
"ribu" -> Double -> Maybe Token
double Double
1e3 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Int -> Token -> Maybe Token
withGrain Int
3 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Token -> Maybe Token
withMultipliable
        Text
"juta" -> Double -> Maybe Token
double Double
1e6 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Int -> Token -> Maybe Token
withGrain Int
6 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Token -> Maybe Token
withMultipliable
        Text
_      -> Maybe Token
forall a. Maybe a
Nothing
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleDozen :: Rule
ruleDozen :: Rule
ruleDozen = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"dozen"
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"(se)?lusin"
    ]
  , prod :: Production
prod = \[Token]
_ -> Integer -> Maybe Token
integer Integer
12 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Int -> Token -> Maybe Token
withGrain Int
1 Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Token -> Maybe Token
withMultipliable
  }

ruleInteger :: Rule
ruleInteger :: Rule
ruleInteger = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"integer (0..9 11)"
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"(kosong|nol|satu|dua|tiga|empat|lima|enam|tujuh|delapan|sembilan|sebelas)"
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
RegexMatch (GroupMatch (match:_)):[Token]
_) ->
        Text -> HashMap Text Integer -> Maybe Integer
forall k v. (Eq k, Hashable k) => k -> HashMap k v -> Maybe v
HashMap.lookup (Text -> Text
Text.toLower Text
match) HashMap Text Integer
ruleIntegerMap Maybe Integer -> (Integer -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Integer -> Maybe Token
integer
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleInteger2 :: Rule
ruleInteger2 :: Rule
ruleInteger2 = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"integer 20..90"
  , pattern :: Pattern
pattern =
    [ Double -> Double -> PatternItem
numberBetween Double
2 Double
10
    , (NumeralData -> Double) -> (Double -> Bool) -> PatternItem
forall t. (NumeralData -> t) -> (t -> Bool) -> PatternItem
numberWith NumeralData -> Double
TNumeral.value (Double -> Double -> Bool
forall a. Eq a => a -> a -> Bool
== Double
10)
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
Numeral NumeralData{TNumeral.value = v1}:
       Token Dimension a
Numeral NumeralData{TNumeral.value = v2, TNumeral.grain = Just g}:
       [Token]
_) -> Double -> Maybe Token
double (Double
v1 Double -> Double -> Double
forall a. Num a => a -> a -> a
* Double
v2) Maybe Token -> (Token -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Int -> Token -> Maybe Token
withGrain Int
g
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

ruleIntegerWithThousandsSeparator :: Rule
ruleIntegerWithThousandsSeparator :: Rule
ruleIntegerWithThousandsSeparator = Rule :: Text -> Pattern -> Production -> Rule
Rule
  { name :: Text
name = Text
"integer with thousands separator ."
  , pattern :: Pattern
pattern =
    [ String -> PatternItem
regex String
"(\\d{1,3}(\\.\\d\\d\\d){1,5})"
    ]
  , prod :: Production
prod = \[Token]
tokens -> case [Token]
tokens of
      (Token Dimension a
RegexMatch (GroupMatch (match:_)):[Token]
_) ->
        Text -> Maybe Double
parseDouble (Text -> Text -> Text -> Text
Text.replace Text
"." Text
Text.empty Text
match) Maybe Double -> (Double -> Maybe Token) -> Maybe Token
forall (m :: * -> *) a b. Monad m => m a -> (a -> m b) -> m b
>>= Double -> Maybe Token
double
      [Token]
_ -> Maybe Token
forall a. Maybe a
Nothing
  }

rules :: [Rule]
rules :: [Rule]
rules =
  [ Rule
ruleDecimalNumeral
  , Rule
ruleDecimalWithThousandsSeparator
  , Rule
ruleDozen
  , Rule
ruleInteger
  , Rule
ruleInteger2
  , Rule
ruleInteger3
  , Rule
ruleIntegerWithThousandsSeparator
  , Rule
ruleIntersect
  , Rule
ruleMultiply
  , Rule
ruleNumeralCommaNumeral
  , Rule
ruleNumeralsPrefixWithNegativeOrMinus
  , Rule
ruleNumeralsSuffixesKMG
  , Rule
rulePowersOfTen
  , Rule
ruleSomefewcouple
  , Rule
ruleTeen
  , Rule
ruleTen
  ]