src/Gargantext/Text/Metrics.hs

   1 {-|
   2 Module      : Gargantext.Text.Metrics
   3 Description : All parsers of Gargantext in one file.
   4 Copyright   : (c) CNRS, 2017 - present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Mainly reexport functions in @Data.Text.Metrics@
  11 -}
  12
  13 {-# LANGUAGE NoImplicitPrelude #-}
  14
  15
  16 module Gargantext.Text.Metrics (levenshtein
  17                                       , levenshteinNorm
  18                                       , damerauLevenshtein
  19                                       , damerauLevenshteinNorm
  20                                       , overlap
  21                                       , jaccard
  22                                       , hamming
  23                                       ) where
  24
  25
  26 import Data.Text (Text)
  27 import GHC.Real (Ratio)
  28 import qualified Data.Text.Metrics as DTM
  29
  30 import Gargantext.Prelude
  31 {- * Example de titre
  32 -}
  33
  34 -- | This module provide metrics to compare Text
  35 -- starting as an API rexporting main functions of the great lib
  36 -- text-metrics of Mark Karpov
  37
  38 -- | Levenshtein Distance
  39 -- In information theory, Linguistics and computer science,
  40 -- the Levenshtein distance is a string metric for measuring
  41 -- the difference between two sequences.
  42 -- See: https://en.wikipedia.org/wiki/Levenshtein_distance
  43 --
  44 levenshtein :: Text -> Text -> Int
  45 levenshtein = DTM.levenshtein
  46
  47 -- | Return normalized Levenshtein distance between two 'Text' values.
  48 -- Result is a non-negative rational number (represented as @'Ratio'
  49 -- 'Data.Numeric.Natural'@), where 0 signifies no similarity between the
  50 -- strings, while 1 means exact match.
  51 --
  52 levenshteinNorm :: Text -> Text -> Ratio Int
  53 levenshteinNorm = DTM.levenshteinNorm
  54
  55 -- | Return Damerau-Levenshtein distance between two 'Text' values. The
  56 -- function works like 'levenshtein', but the collection of allowed
  57 -- operations also includes transposition of two /adjacent/ characters.
  58 -- See also:
  59 -- <https://en.wikipedia.org/wiki/Damerau%E2%80%93Levenshtein_distance>
  60 --
  61 damerauLevenshtein :: Text -> Text -> Int
  62 damerauLevenshtein = DTM.damerauLevenshtein
  63
  64 -- damerau-Levenshtein distance normalized
  65 --
  66 damerauLevenshteinNorm :: Text -> Text -> Ratio Int
  67 damerauLevenshteinNorm = DTM.damerauLevenshteinNorm
  68
  69 -- Treating inputs like sets
  70
  71 -- | Return overlap coefficient for two 'Text' values. Returned value
  72 -- is in the range from 0 (no similarity) to 1 (exact match). Return 1
  73 -- if both 'Text' values are empty.
  74 --
  75 -- See also: <https://en.wikipedia.org/wiki/Overlap_coefficient>.
  76 overlap :: Text -> Text -> Ratio Int
  77 overlap = DTM.overlap
  78
  79
  80 -- | Jaccard distance
  81 -- measures dissimilarity between sample sets
  82 jaccard :: Text -> Text -> Ratio Int
  83 jaccard = DTM.jaccard
  84
  85 -- | Hamming Distance
  86 -- In information theory, the Hamming distance between two strings of
  87 -- equal length is the number of positions at which the corresponding
  88 -- symbols are different. In other words, it measures the minimum number of
  89 -- substitutions required to change one string into the other
  90 -- See:  https://en.wikipedia.org/wiki/Hamming_distance
  91
  92 hamming :: Text -> Text -> Maybe Int
  93 hamming = DTM.hamming
  94
  95