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

   1 {-|
   2 Module      : Gargantext.Core.Text.Corpus.API.Pubmed
   3 Description : Pubmed API connection
   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
  13 module Gargantext.Core.Text.Corpus.API.Pubmed
  14     where
  15
  16 import Conduit
  17 import Control.Monad.Reader (runReaderT)
  18 import Data.Either (Either)
  19 import Data.Maybe
  20 import Data.Text (Text)
  21 import qualified Data.Text as Text
  22 import Servant.Client (ClientError)
  23
  24 import Gargantext.Prelude
  25 import Gargantext.Core (Lang(..))
  26 import Gargantext.Database.Admin.Types.Hyperdata (HyperdataDocument(..))
  27
  28 import qualified PUBMED as PubMed
  29 import qualified PUBMED.Parser as PubMedDoc
  30 import PUBMED.Types (Config(..))
  31
  32
  33 type Query = Text
  34 type Limit = Integer
  35
  36
  37 -- | TODO put default pubmed query in gargantext.ini
  38 -- by default: 10K docs
  39 get :: Maybe Text
  40     -> Query
  41     -> Maybe Limit
  42     -> IO (Either ClientError (Maybe Integer, ConduitT () HyperdataDocument IO ()))
  43 get mAPIKey q l = do
  44   eRes <- runReaderT PubMed.getMetadataWithC (Config { apiKey = mAPIKey
  45                                                      , query = q
  46                                                      , perPage = Nothing })
  47   let takeLimit = case l of
  48         Nothing -> mapC identity
  49         Just l' -> takeC $ fromIntegral l'
  50   pure $ (\(len, docsC) -> (len, docsC .| takeLimit .| mapC (toDoc EN))) <$> eRes
  51   --either (\e -> panic $ "CRAWL: PubMed" <> e) (map (toDoc EN))
  52   --      <$> PubMed.getMetadataWithC q l
  53
  54 toDoc :: Lang -> PubMedDoc.PubMed -> HyperdataDocument
  55 toDoc l (PubMedDoc.PubMed { pubmed_id
  56                           , pubmed_article = PubMedDoc.PubMedArticle t j as aus
  57                           , pubmed_date = PubMedDoc.PubMedDate a y m d }
  58           ) = HyperdataDocument { _hd_bdd = Just "PubMed"
  59                                 , _hd_doi = Nothing
  60                                 , _hd_url = Nothing
  61                                 , _hd_uniqId = Just $ Text.pack $ show pubmed_id
  62                                 , _hd_uniqIdBdd = Nothing
  63                                 , _hd_page = Nothing
  64                                 , _hd_title = t
  65                                 , _hd_authors = authors aus
  66                                 , _hd_institutes = institutes aus
  67                                 , _hd_source = j
  68                                 , _hd_abstract = abstract as
  69                                 , _hd_publication_date = Just $ Text.pack $ show a
  70                                 , _hd_publication_year = Just $ fromIntegral y
  71                                 , _hd_publication_month = Just m
  72                                 , _hd_publication_day = Just d
  73                                 , _hd_publication_hour = Nothing
  74                                 , _hd_publication_minute = Nothing
  75                                 , _hd_publication_second = Nothing
  76                                 , _hd_language_iso2 = Just $ (Text.pack . show) l }
  77       where
  78         authors :: [PubMedDoc.Author] -> Maybe Text
  79         authors [] = Nothing
  80         authors au = Just $ (Text.intercalate ", ")
  81                           $ catMaybes
  82                           $ map (\n -> PubMedDoc.foreName n <> Just " " <> PubMedDoc.lastName n) au
  83
  84         institutes :: [PubMedDoc.Author] -> Maybe Text
  85         institutes [] = Nothing
  86         institutes au = Just $ (Text.intercalate ", ")
  87                              $ (map (Text.replace ", " " - "))
  88                              $ catMaybes
  89                              $ map PubMedDoc.affiliation au
  90
  91
  92         abstract :: [Text] -> Maybe Text
  93         abstract [] = Nothing
  94         abstract as' = Just $ Text.intercalate ", " as'