src/Gargantext/Text/Corpus/Parsers/Json2Csv.hs

   1 {-|
   2 Module      : Gargantext.Text.Corpus.Parsers.Json2Csv
   3 Description :
   4 Copyright   : (c) CNRS, 2017-Present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Json parser to export towoard CSV GargV3 format.
  11 (Export from the Patent Database.)
  12
  13 -}
  14
  15 {-# LANGUAGE NoImplicitPrelude #-}
  16 {-# LANGUAGE OverloadedStrings #-}
  17 {-# LANGUAGE DeriveGeneric     #-}
  18 {-# LANGUAGE TemplateHaskell   #-}
  19
  20 module Gargantext.Text.Corpus.Parsers.Json2Csv (json2csv, readPatents)
  21   where
  22
  23 import Prelude (read)
  24 import Data.Aeson
  25 import Data.Aeson.TH (deriveJSON)
  26 import Data.ByteString.Lazy (readFile)
  27 import Data.Text (Text, unpack)
  28 import Gargantext.Core.Utils.Prefix (unPrefix)
  29 import Gargantext.Prelude
  30 import System.IO (FilePath)
  31 import Gargantext.Text.Corpus.Parsers.CSV (CsvDoc(..), writeFile, headerCsvGargV3)
  32 import Data.Vector (fromList)
  33
  34 data Patent = Patent { _patent_title :: Text
  35                      , _patent_abstract :: Text
  36                      , _patent_year     :: Text
  37                      , _patent_id       :: Text
  38  } deriving (Show)
  39
  40 $(deriveJSON (unPrefix "_patent_") ''Patent)
  41
  42 readPatents :: FilePath -> IO (Maybe [Patent])
  43 readPatents fp = decode <$> readFile fp
  44
  45 type FilePathIn  = FilePath
  46 type FilePathOut = FilePath
  47
  48 json2csv :: FilePathIn -> FilePathOut -> IO ()
  49 json2csv fin fout = do
  50   patents <- maybe (panic "json2csv error") identity <$> readPatents fin
  51   writeFile fout (headerCsvGargV3, fromList $ map patent2csvDoc patents)
  52
  53 patent2csvDoc :: Patent -> CsvDoc
  54 patent2csvDoc (Patent title abstract year _) =
  55   CsvDoc title "Source" (read (unpack year)) 1 1 abstract "Authors"
  56
  57
  58
  59
  60