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

   1 {-|
   2 Module      : Gargantext.Core.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
  13 module Gargantext.Core.Text.Corpus.API.Istex
  14     where
  15
  16 import Data.List (concat)
  17 import Data.Maybe
  18 import Data.Text (Text, pack)
  19
  20 import Gargantext.Core (Lang(..))
  21 import Gargantext.Database.Admin.Types.Hyperdata (HyperdataDocument(..))
  22 import Gargantext.Prelude
  23 import qualified Gargantext.Core.Text.Corpus.Parsers.Date as Date
  24 import qualified ISTEX        as ISTEX
  25 import qualified ISTEX.Client as ISTEX
  26
  27 get :: Lang -> Text -> Maybe Integer -> IO [HyperdataDocument]
  28 get la q ml = do
  29   docs <- ISTEX.getMetadataWith q (fromIntegral <$> ml)
  30   either (panic . pack . show) (toDoc' la) docs
  31
  32 toDoc' :: Lang -> ISTEX.Documents -> IO [HyperdataDocument]
  33 toDoc' la docs' = do
  34   --printDebug "ISTEX" (ISTEX._documents_total docs')
  35   mapM (toDoc la) (ISTEX._documents_hits docs')
  36
  37 -- | TODO remove dateSplit here
  38 -- TODO current year as default
  39 toDoc :: Lang -> ISTEX.Document -> IO HyperdataDocument
  40 toDoc la (ISTEX.Document i t a ab d s) = do
  41   (utctime, (pub_year, pub_month, pub_day)) <- Date.dateSplit la (maybe (Just "2019") (Just . pack . show) d)
  42   pure $ HyperdataDocument { _hd_bdd = Just "Istex"
  43                            , _hd_doi = Just i
  44                            , _hd_url = Nothing
  45                            , _hd_uniqId = Nothing
  46                            , _hd_uniqIdBdd = Nothing
  47                            , _hd_page = Nothing
  48                            , _hd_title = t
  49                            , _hd_authors = Just $ foldl (\x y -> x <> ", " <> y) "" (map ISTEX._author_name a)
  50                            , _hd_institutes = Just $ foldl (\x y -> x <> ", " <> y) "" (concat $ (map ISTEX._author_affiliations) a)
  51                            , _hd_source = Just $ foldl (\x y -> x <> ", " <> y) "" (catMaybes $ map ISTEX._source_title s)
  52                            , _hd_abstract = ab
  53                            , _hd_publication_date = fmap (pack . show) utctime
  54                            , _hd_publication_year = pub_year
  55                            , _hd_publication_month = pub_month
  56                            , _hd_publication_day = pub_day
  57                            , _hd_publication_hour = Nothing
  58                            , _hd_publication_minute = Nothing
  59                            , _hd_publication_second = Nothing
  60                            , _hd_language_iso2 = Just $ (pack . show) la }
  61