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.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 -- | TODO put in gargantext.init
  35 default_limit :: Maybe Integer
  36 default_limit = Just 10000
  37
  38 -- | Get External API metadata main function
  39 get :: ExternalAPIs
  40     -> Lang
  41     -> Query
  42     -> Maybe Limit
  43     -- -> IO [HyperdataDocument]
  44     -> IO (Either ClientError (Maybe Integer, ConduitT () HyperdataDocument IO ()))
  45 get PubMed  _la q _l = PUBMED.get q Nothing
  46   --docs <- PUBMED.get   q default_limit -- EN only by default
  47   --pure (Just $ fromIntegral $ length docs, yieldMany docs)
  48 get HAL      la q _l = HAL.getC  la q Nothing
  49 get IsTex    la q _l = do
  50   docs <- ISTEX.get la q default_limit
  51   pure $ Right (Just $ fromIntegral $ length docs, yieldMany docs)
  52 get Isidore  la q _l = do
  53   docs <- ISIDORE.get la (fromIntegral <$> default_limit) (Just q) Nothing
  54   pure $ Right (Just $ fromIntegral $ length docs, yieldMany docs)
  55 get _        _  _ _ = undefined
  56
  57 -- | Some Sugar for the documentation
  58 type Query = PUBMED.Query
  59 type Limit = PUBMED.Limit