src/Gargantext/Text/Corpus/API/Istex.hs

   1 {-|
   2 Module      : Gargantext.Text.Corpus.API.Istex
   3 Description : Pubmed API connection
   4 Copyright   : (c) CNRS, 2017
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 -}
  11
  12 {-# LANGUAGE NoImplicitPrelude #-}
  13 {-# LANGUAGE OverloadedStrings #-}
  14
  15 module Gargantext.Text.Corpus.API.Istex
  16     where
  17
  18 import Data.Either (either)
  19 import Data.Maybe
  20 import Data.List (concat)
  21 import Data.Maybe (catMaybes)
  22 import Data.Text (Text, pack)
  23 import Gargantext.Prelude
  24 import Gargantext.Core (Lang(..))
  25 import qualified Gargantext.Text.Corpus.Parsers.Date as Date
  26 import Gargantext.Database.Types.Node (HyperdataDocument(..))
  27
  28 import qualified ISTEX        as ISTEX
  29 import qualified ISTEX.Client as ISTEX
  30
  31 get :: Lang -> Text -> Maybe Integer -> IO [HyperdataDocument]
  32 get la q ml = do
  33   docs <- ISTEX.getMetadataWith q (fromIntegral <$> ml)
  34   either (panic . pack . show) (toDoc' la) docs
  35
  36 toDoc' :: Lang -> ISTEX.Documents -> IO [HyperdataDocument]
  37 toDoc' la docs' = do
  38   --printDebug "ISTEX" (ISTEX._documents_total docs')
  39   mapM (toDoc la) (ISTEX._documents_hits docs')
  40
  41 -- | TODO remove dateSplit here
  42 -- TODO current year as default
  43 toDoc :: Lang -> ISTEX.Document -> IO HyperdataDocument
  44 toDoc la (ISTEX.Document i t a ab d s) = do
  45   (utctime, (pub_year, pub_month, pub_day)) <- Date.dateSplit la (maybe (Just "2019") (Just . pack . show) d)
  46   pure $ HyperdataDocument (Just "Istex")
  47                    (Just i)
  48                    Nothing
  49                    Nothing
  50                    Nothing
  51                    Nothing
  52                    t
  53                    (Just $ foldl (\x y -> x <> ", " <> y) "" (map ISTEX._author_name a))
  54                    (Just $ foldl (\x y -> x <> ", " <> y) "" (concat $ (map ISTEX._author_affiliations) a))
  55                    (Just $ foldl (\x y -> x <> ", " <> y) "" (catMaybes $ map ISTEX._source_title s))
  56                    ab
  57                    (fmap (pack . show) utctime)
  58                    pub_year
  59                    pub_month
  60                    pub_day
  61                    Nothing
  62                    Nothing
  63                    Nothing
  64                   (Just $ (pack . show) la)