src/Gargantext/Text/Corpus/API.hs

   1 {-|
   2 Module      : Gargantext.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 {-# LANGUAGE NoImplicitPrelude #-}
  13 {-# LANGUAGE OverloadedStrings #-}
  14 {-# LANGUAGE DeriveGeneric     #-}
  15
  16 module Gargantext.Text.Corpus.API
  17     where
  18
  19 import GHC.Generics (Generic)
  20 import Data.Aeson
  21 import Data.Maybe
  22 import Data.Text (Text)
  23 import Gargantext.Prelude
  24 import Gargantext.Core (Lang(..))
  25 import Gargantext.Database.Types.Node (HyperdataDocument(..))
  26 import Test.QuickCheck.Arbitrary
  27 import Test.QuickCheck (elements)
  28 import Data.Swagger
  29 import qualified Data.Text as Text
  30
  31 import qualified PUBMED as PubMed
  32 import qualified PUBMED.Parser as Doc
  33 --import qualified Gargantext.Text.Corpus.API.Isidore as Isidore
  34
  35 data ExternalAPIs = All
  36                   | PubMed
  37                   | HAL
  38                   -- | IsTex
  39                   | IsidoreQuery | IsidoreAuth
  40   deriving (Show, Eq, Enum, Bounded, Generic)
  41
  42 instance FromJSON ExternalAPIs
  43 instance ToJSON ExternalAPIs
  44
  45 externalAPIs :: [ExternalAPIs]
  46 externalAPIs = [minBound..maxBound]
  47
  48 instance Arbitrary ExternalAPIs
  49   where
  50     arbitrary = elements externalAPIs
  51
  52 instance ToSchema ExternalAPIs
  53
  54 type Query = Text
  55 type Limit = PubMed.Limit
  56
  57 get :: ExternalAPIs -> Query -> Maybe Limit -> IO [HyperdataDocument]
  58 get PubMed q l = either (\e -> panic $ "CRAWL: PubMed" <> e) (map (toDoc EN)) <$> PubMed.getMetadataWith q l
  59 get _ _ _ = undefined
  60
  61 toDoc :: Lang -> Doc.PubMed -> HyperdataDocument
  62 toDoc l (Doc.PubMed (Doc.PubMedArticle t j as aus)
  63                     (Doc.PubMedDate a y m d)
  64           ) = HyperdataDocument (Just "PubMed")
  65                                  Nothing
  66                                  Nothing
  67                                  Nothing
  68                                  Nothing
  69                                  Nothing
  70                                  t
  71                                  (authors aus)
  72                                  Nothing
  73                                  j
  74                                  (abstract as)
  75                                  (Just $ Text.pack $ show a)
  76                                  (Just $ fromIntegral y)
  77                                  (Just m)
  78                                  (Just d)
  79                                  Nothing
  80                                  Nothing
  81                                  Nothing
  82                                  (Just $ (Text.pack . show) l)
  83       where
  84         authors :: Maybe [Doc.Author] -> Maybe Text
  85         authors aus' = case aus' of
  86             Nothing -> Nothing
  87             Just au -> Just $ (Text.intercalate ", ") $ catMaybes $ map Doc.foreName au
  88
  89         abstract :: Maybe [Text] -> Maybe Text
  90         abstract as' = fmap (Text.intercalate ", ") as'
  91