src/Gargantext/Ngrams/Metrics.hs

   1 {-|
   2 Module      : Gargantext.Ngrams.Metrics
   3 Description : Short description
   4 Copyright   : (c) Some Guy, 2013
   5                   Someone Else, 2014
   6 License     : GPL-3
   7 Maintainer  : sample@email.com
   8 Stability   : experimental
   9 Portability : POSIX
  10
  11 Here is a longer description of this module, containing some
  12 commentary with @some markup@.
  13 -}
  14
  15 module Gargantext.Ngrams.Metrics (levenshtein
  16                                       , levenshteinNorm
  17                                       , damerauLevenshtein
  18                                       , damerauLevenshteinNorm
  19                                       , overlap
  20                                       , jaccard
  21                                       , hamming
  22                                       ) where
  23
  24 import Gargantext.Prelude
  25
  26 import Data.Text (Text)
  27 import GHC.Real (Ratio)
  28 import qualified Data.Text.Metrics as DTM
  29
  30 {- * Example de titre
  31 -}
  32
  33 -- | This module provide metrics to compare Text
  34 -- starting as an API rexporting main functions of the great lib
  35 -- text-metrics of Mark Karpov
  36
  37 -- | Levenshtein Distance
  38 -- In information theory, Linguistics and computer science,
  39 -- the Levenshtein distance is a string metric for measuring
  40 -- the difference between two sequences.
  41 -- See: https://en.wikipedia.org/wiki/Levenshtein_distance
  42 --
  43 levenshtein :: Text -> Text -> Int
  44 levenshtein = DTM.levenshtein
  45
  46 -- | Return normalized Levenshtein distance between two 'Text' values.
  47 -- Result is a non-negative rational number (represented as @'Ratio'
  48 -- 'Data.Numeric.Natural'@), where 0 signifies no similarity between the
  49 -- strings, while 1 means exact match.
  50 --
  51 levenshteinNorm :: Text -> Text -> Ratio Int
  52 levenshteinNorm = DTM.levenshteinNorm
  53
  54 -- | Return Damerau-Levenshtein distance between two 'Text' values. The
  55 -- function works like 'levenshtein', but the collection of allowed
  56 -- operations also includes transposition of two /adjacent/ characters.
  57 -- See also:
  58 -- <https://en.wikipedia.org/wiki/Damerau%E2%80%93Levenshtein_distance>
  59 --
  60 damerauLevenshtein :: Text -> Text -> Int
  61 damerauLevenshtein = DTM.damerauLevenshtein
  62
  63 -- damerau-Levenshtein distance normalized
  64 --
  65 damerauLevenshteinNorm :: Text -> Text -> Ratio Int
  66 damerauLevenshteinNorm = DTM.damerauLevenshteinNorm
  67
  68 -- Treating inputs like sets
  69
  70 -- | Return overlap coefficient for two 'Text' values. Returned value
  71 -- is in the range from 0 (no similarity) to 1 (exact match). Return 1
  72 -- if both 'Text' values are empty.
  73 --
  74 -- See also: <https://en.wikipedia.org/wiki/Overlap_coefficient>.
  75 overlap :: Text -> Text -> Ratio Int
  76 overlap = DTM.overlap
  77
  78
  79 -- | Jaccard distance
  80 -- measures dissimilarity between sample sets
  81 jaccard :: Text -> Text -> Ratio Int
  82 jaccard = DTM.jaccard
  83
  84 -- | Hamming Distance
  85 -- In information theory, the Hamming distance between two strings of
  86 -- equal length is the number of positions at which the corresponding
  87 -- symbols are different. In other words, it measures the minimum number of
  88 -- substitutions required to change one string into the other
  89 -- See:  https://en.wikipedia.org/wiki/Hamming_distance
  90
  91 hamming :: Text -> Text -> Maybe Int
  92 hamming = DTM.hamming
  93
  94