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

   1 {-|
   2 Module      : Gargantext.Core.Text.Corpus.API.Hal
   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 module Gargantext.Core.Text.Corpus.API.Hal
  13     where
  14
  15 import Data.Maybe
  16 import Data.Text (Text, pack, intercalate)
  17
  18 import Gargantext.Core (Lang(..))
  19 import Gargantext.Database.Admin.Types.Hyperdata (HyperdataDocument(..))
  20 import Gargantext.Prelude
  21 import qualified Gargantext.Core.Text.Corpus.Parsers.Date as Date
  22 import qualified HAL            as HAL
  23 import qualified HAL.Client     as HAL
  24 import qualified HAL.Doc.Corpus as HAL
  25
  26 get :: Lang -> Text -> Maybe Integer -> IO [HyperdataDocument]
  27 get la q ml = do
  28   docs <- HAL.getMetadataWith q (Just 0) (fromIntegral <$> ml)
  29   either (panic . pack . show) (\d -> mapM (toDoc' la) $ HAL._docs d) docs
  30
  31 toDoc' :: Lang -> HAL.Corpus -> IO HyperdataDocument
  32 toDoc' la (HAL.Corpus i t ab d s aus affs struct_id) = do
  33   (utctime, (pub_year, pub_month, pub_day)) <- Date.dateSplit la (maybe (Just "2019") Just d)
  34   pure $ HyperdataDocument { _hd_bdd = Just "Hal"
  35                            , _hd_doi = Just $ pack $ show i
  36                            , _hd_url = Nothing
  37                            , _hd_uniqId = Nothing
  38                            , _hd_uniqIdBdd = Nothing
  39                            , _hd_page = Nothing
  40                            , _hd_title = Just $ intercalate " " t
  41                            , _hd_authors = Just $ foldl (\x y -> x <> ", " <> y) "" aus
  42                            , _hd_institutes = Just $ foldl (\x y -> x <> ", " <> y) "" $ affs <> map (cs . show) struct_id
  43                            , _hd_source = Just $ maybe "Nothing" identity s
  44                            , _hd_abstract = Just $ intercalate " " ab
  45                            , _hd_publication_date = fmap (pack . show) utctime
  46                            , _hd_publication_year = pub_year
  47                            , _hd_publication_month = pub_month
  48                            , _hd_publication_day = pub_day
  49                            , _hd_publication_hour = Nothing
  50                            , _hd_publication_minute = Nothing
  51                            , _hd_publication_second = Nothing
  52                            , _hd_language_iso2 = Just $ (pack . show) la }
  53