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 Control.Lens ((^.))
  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 Gargantext.Prelude.Config (GargConfig, gc_pubmed_api_key)
  29 import qualified Gargantext.Core.Text.Corpus.API.Arxiv   as Arxiv
  30 import qualified Gargantext.Core.Text.Corpus.API.Hal     as HAL
  31 import qualified Gargantext.Core.Text.Corpus.API.Isidore as ISIDORE
  32 import qualified Gargantext.Core.Text.Corpus.API.Istex   as ISTEX
  33 import qualified Gargantext.Core.Text.Corpus.API.Pubmed  as PUBMED
  34 import Servant.Client (ClientError)
  35
  36 -- | Get External API metadata main function
  37 get :: GargConfig
  38     -> ExternalAPIs
  39     -> Lang
  40     -> Query
  41     -> Maybe Limit
  42     -- -> IO [HyperdataDocument]
  43     -> IO (Either ClientError (Maybe Integer, ConduitT () HyperdataDocument IO ()))
  44 get cfg PubMed _la q limit = PUBMED.get (cfg ^. gc_pubmed_api_key) q limit
  45   --docs <- PUBMED.get   q default_limit -- EN only by default
  46   --pure (Just $ fromIntegral $ length docs, yieldMany docs)
  47 get _ Arxiv    la q limit = Arxiv.get la q (fromIntegral <$> limit)
  48 get _ HAL      la q limit = HAL.getC  la q limit
  49 get _ IsTex    la q limit = do
  50   docs <- ISTEX.get la q limit
  51   pure $ Right (Just $ fromIntegral $ length docs, yieldMany docs)
  52 get _ Isidore  la q limit = do
  53   docs <- ISIDORE.get la (fromIntegral <$> limit) (Just q) Nothing
  54   pure $ Right (Just $ fromIntegral $ length docs, yieldMany docs)
  55 get _ externalApi  _  _ _ = panic $ "[G.C.T.Corpus.API] This options are note taken into account: " <>  (cs $ show externalApi)
  56
  57 -- | Some Sugar for the documentation
  58 type Query = PUBMED.Query
  59 type Limit = PUBMED.Limit