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 qualified Gargantext.Defaults as Defaults
  21 import Gargantext.Prelude
  22 import qualified Gargantext.Core.Text.Corpus.Parsers.Date as Date
  23 import qualified HAL            as HAL
  24 import qualified HAL.Client     as HAL
  25 import qualified HAL.Doc.Corpus as HAL
  26
  27 get :: Lang -> Text -> Maybe Integer -> IO [HyperdataDocument]
  28 get la q ml = do
  29   docs <- HAL.getMetadataWith q (Just 0) (fromIntegral <$> ml)
  30   pure $ either (panic . pack . show) (\d -> map (toDoc' la) $ HAL._docs d) docs
  31
  32 toDoc' :: Lang -> HAL.Corpus -> HyperdataDocument
  33 toDoc' la (HAL.Corpus i t ab d s aus affs struct_id) = do
  34   let (utctime, (pub_year, pub_month, pub_day)) =
  35         Date.dateSplit la (maybe (Just $ pack $ show Defaults.year) Just d)
  36   HyperdataDocument { _hd_bdd = Just "Hal"
  37                     , _hd_doi = Just $ pack $ show i
  38                     , _hd_url = Nothing
  39                     , _hd_uniqId = Nothing
  40                     , _hd_uniqIdBdd = Nothing
  41                     , _hd_page = Nothing
  42                     , _hd_title = Just $ intercalate " " t
  43                     , _hd_authors = Just $ foldl (\x y -> x <> ", " <> y) "" aus
  44                     , _hd_institutes = Just $ foldl (\x y -> x <> ", " <> y) "" $ affs <> map (cs . show) struct_id
  45                     , _hd_source = Just $ maybe "Nothing" identity s
  46                     , _hd_abstract = Just $ intercalate " " ab
  47                     , _hd_publication_date = fmap (pack . show) utctime
  48                     , _hd_publication_year = pub_year
  49                     , _hd_publication_month = pub_month
  50                     , _hd_publication_day = pub_day
  51                     , _hd_publication_hour = Nothing
  52                     , _hd_publication_minute = Nothing
  53                     , _hd_publication_second = Nothing
  54                     , _hd_language_iso2 = Just $ (pack . show) la }
  55