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.List (concat)
  20 import Data.Maybe
  21 import Data.Maybe (catMaybes)
  22 import Data.Text (Text, pack)
  23 import Gargantext.Core (Lang(..))
  24 import Gargantext.Database.Admin.Types.Node (HyperdataDocument(..))
  25 import Gargantext.Prelude
  26 import qualified Gargantext.Text.Corpus.Parsers.Date as Date
  27 import qualified ISTEX        as ISTEX
  28 import qualified ISTEX.Client as ISTEX
  29
  30 get :: Lang -> Text -> Maybe Integer -> IO [HyperdataDocument]
  31 get la q ml = do
  32   docs <- ISTEX.getMetadataWith q (fromIntegral <$> ml)
  33   either (panic . pack . show) (toDoc' la) docs
  34
  35 toDoc' :: Lang -> ISTEX.Documents -> IO [HyperdataDocument]
  36 toDoc' la docs' = do
  37   --printDebug "ISTEX" (ISTEX._documents_total docs')
  38   mapM (toDoc la) (ISTEX._documents_hits docs')
  39
  40 -- | TODO remove dateSplit here
  41 -- TODO current year as default
  42 toDoc :: Lang -> ISTEX.Document -> IO HyperdataDocument
  43 toDoc la (ISTEX.Document i t a ab d s) = do
  44   (utctime, (pub_year, pub_month, pub_day)) <- Date.dateSplit la (maybe (Just "2019") (Just . pack . show) d)
  45   pure $ HyperdataDocument (Just "Istex")
  46                    (Just i)
  47                    Nothing
  48                    Nothing
  49                    Nothing
  50                    Nothing
  51                    t
  52                    (Just $ foldl (\x y -> x <> ", " <> y) "" (map ISTEX._author_name a))
  53                    (Just $ foldl (\x y -> x <> ", " <> y) "" (concat $ (map ISTEX._author_affiliations) a))
  54                    (Just $ foldl (\x y -> x <> ", " <> y) "" (catMaybes $ map ISTEX._source_title s))
  55                    ab
  56                    (fmap (pack . show) utctime)
  57                    pub_year
  58                    pub_month
  59                    pub_day
  60                    Nothing
  61                    Nothing
  62                    Nothing
  63                   (Just $ (pack . show) la)