src/Data/Gargantext/Ngrams/Metrics.hs

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