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 Conduit
  16 import Data.Either
  17 import Data.Maybe
  18 import Data.Text (Text, pack, intercalate)
  19 import Servant.Client (ClientError)
  20
  21 import Gargantext.Core (Lang(..))
  22 import Gargantext.Database.Admin.Types.Hyperdata (HyperdataDocument(..))
  23 import qualified Gargantext.Defaults as Defaults
  24 import Gargantext.Prelude
  25 import qualified Gargantext.Core.Text.Corpus.Parsers.Date as Date
  26 import qualified HAL            as HAL
  27 import qualified HAL.Client     as HAL
  28 import qualified HAL.Doc.Corpus as HAL
  29
  30 get :: Lang -> Text -> Maybe Int -> IO [HyperdataDocument]
  31 get la q ml = do
  32   eDocs <- HAL.getMetadataWith q (Just 0) (fromIntegral <$> ml)
  33   either (panic . pack . show) (\d -> mapM (toDoc' la) $ HAL._docs d) eDocs
  34
  35 getC :: Lang -> Text -> Maybe Int -> IO (Either ClientError (Maybe Integer, ConduitT () HyperdataDocument IO ()))
  36 getC la q ml = do
  37   eRes <- HAL.getMetadataWithC q (Just 0) (fromIntegral <$> ml)
  38   pure $ (\(len, docsC) -> (len, docsC .| mapMC (toDoc' la))) <$> eRes
  39 --  case eRes of
  40 --    Left err -> panic $ pack $ show err
  41 --    Right (len, docsC) -> pure (len, docsC .| mapMC (toDoc' la))
  42
  43 toDoc' :: Lang -> HAL.Corpus -> IO HyperdataDocument
  44 toDoc' la (HAL.Corpus i t ab d s aus affs struct_id) = do
  45   (utctime, (pub_year, pub_month, pub_day)) <-
  46         Date.dateSplit la (maybe (Just $ pack $ show Defaults.year) Just d)
  47   pure HyperdataDocument { _hd_bdd = Just "Hal"
  48                          , _hd_doi = Just $ pack $ show i
  49                          , _hd_url = Nothing
  50                          , _hd_uniqId = Nothing
  51                          , _hd_uniqIdBdd = Nothing
  52                          , _hd_page = Nothing
  53                          , _hd_title = Just $ intercalate " " t
  54                          , _hd_authors = Just $ foldl (\x y -> x <> ", " <> y) "" aus
  55                          , _hd_institutes = Just $ foldl (\x y -> x <> ", " <> y) "" $ affs <> map (cs . show) struct_id
  56                          , _hd_source = Just $ maybe "Nothing" identity s
  57                          , _hd_abstract = Just $ intercalate " " ab
  58                          , _hd_publication_date = fmap (pack . show) utctime
  59                          , _hd_publication_year = pub_year
  60                          , _hd_publication_month = pub_month
  61                          , _hd_publication_day = pub_day
  62                          , _hd_publication_hour = Nothing
  63                          , _hd_publication_minute = Nothing
  64                          , _hd_publication_second = Nothing
  65                          , _hd_language_iso2 = Just $ (pack . show) la }