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

   1 {-|
   2 Module      : Gargantext.Core.Text.Corpus.API
   3 Description : All crawlers of Gargantext in one file.
   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
  13   ( ExternalAPIs(..)
  14   , Query
  15   , Limit
  16   , get
  17   , externalAPIs
  18   )
  19     where
  20
  21 import Conduit
  22 import Data.Either (Either(..))
  23 import Data.Maybe
  24 import Gargantext.API.Admin.Orchestrator.Types (ExternalAPIs(..), externalAPIs)
  25 import Gargantext.Core (Lang(..))
  26 import Gargantext.Database.Admin.Types.Hyperdata (HyperdataDocument(..))
  27 import Gargantext.Prelude
  28 import qualified Gargantext.Core.Text.Corpus.API.Arxiv   as Arxiv
  29 import qualified Gargantext.Core.Text.Corpus.API.Hal     as HAL
  30 import qualified Gargantext.Core.Text.Corpus.API.Isidore as ISIDORE
  31 import qualified Gargantext.Core.Text.Corpus.API.Istex   as ISTEX
  32 import qualified Gargantext.Core.Text.Corpus.API.Pubmed  as PUBMED
  33 import Servant.Client (ClientError)
  34
  35 -- | Get External API metadata main function
  36 get :: ExternalAPIs
  37     -> Lang
  38     -> Query
  39     -> Maybe Limit
  40     -- -> IO [HyperdataDocument]
  41     -> IO (Either ClientError (Maybe Integer, ConduitT () HyperdataDocument IO ()))
  42 get PubMed  _la q limit = PUBMED.get q limit
  43   --docs <- PUBMED.get   q default_limit -- EN only by default
  44   --pure (Just $ fromIntegral $ length docs, yieldMany docs)
  45 get Arxiv    la q limit = Arxiv.get la q (fromIntegral <$> limit)
  46 get HAL      la q limit = HAL.getC  la q limit
  47 get IsTex    la q limit = do
  48   docs <- ISTEX.get la q limit
  49   pure $ Right (Just $ fromIntegral $ length docs, yieldMany docs)
  50 get Isidore  la q limit = do
  51   docs <- ISIDORE.get la (fromIntegral <$> limit) (Just q) Nothing
  52   pure $ Right (Just $ fromIntegral $ length docs, yieldMany docs)
  53 get _        _  _ _ = undefined
  54
  55 -- | Some Sugar for the documentation
  56 type Query = PUBMED.Query
  57 type Limit = PUBMED.Limit