src/Data/Gargantext/Ngrams/Metrics.hs

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