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