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