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 qualified Gargantext.Defaults as Defaults
  23 import Gargantext.Prelude
  24 import qualified Gargantext.Core.Text.Corpus.Parsers.Date as Date
  25 import qualified ISTEX        as ISTEX
  26 import qualified ISTEX.Client as ISTEX
  27
  28 get :: Lang -> Text -> Maybe Integer -> IO [HyperdataDocument]
  29 get la q ml = do
  30   docs <- ISTEX.getMetadataWith q (fromIntegral <$> ml)
  31   either (panic . pack . show) (toDoc' la) docs
  32
  33 toDoc' :: Lang -> ISTEX.Documents -> IO [HyperdataDocument]
  34 toDoc' la docs' = do
  35   --printDebug "ISTEX" (ISTEX._documents_total docs')
  36   mapM (toDoc la) (ISTEX._documents_hits docs')
  37
  38 -- | TODO remove dateSplit here
  39 -- TODO current year as default
  40 toDoc :: Lang -> ISTEX.Document -> IO HyperdataDocument
  41 toDoc la (ISTEX.Document i t a ab d s) = do
  42   (utctime, (pub_year, pub_month, pub_day)) <- Date.dateSplit la (maybe (Just $ pack $ show Defaults.year) (Just . pack . show) d)
  43   pure $ HyperdataDocument { _hd_bdd = Just "Istex"
  44                            , _hd_doi = Just i
  45                            , _hd_url = Nothing
  46                            , _hd_uniqId = Nothing
  47                            , _hd_uniqIdBdd = Nothing
  48                            , _hd_page = Nothing
  49                            , _hd_title = t
  50                            , _hd_authors = Just $ foldl (\x y -> x <> ", " <> y) "" (map ISTEX._author_name a)
  51                            , _hd_institutes = Just $ foldl (\x y -> x <> ", " <> y) "" (concat $ (map ISTEX._author_affiliations) a)
  52                            , _hd_source = Just $ foldl (\x y -> x <> ", " <> y) "" (catMaybes $ map ISTEX._source_title s)
  53                            , _hd_abstract = ab
  54                            , _hd_publication_date = fmap (pack . show) utctime
  55                            , _hd_publication_year = pub_year
  56                            , _hd_publication_month = pub_month
  57                            , _hd_publication_day = pub_day
  58                            , _hd_publication_hour = Nothing
  59                            , _hd_publication_minute = Nothing
  60                            , _hd_publication_second = Nothing
  61                            , _hd_language_iso2 = Just $ (pack . show) la }
  62