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

   1 {-|
   2 Module      : Gargantext.Core.Text.Corpus.API.OpenAlex
   3 Description : OpenAlex API connection
   4 Copyright   : (c) CNRS, 2023
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9 -}
  10 module Gargantext.Core.Text.Corpus.API.OpenAlex where
  11
  12 import Conduit
  13 import qualified Data.Text as T
  14 import Gargantext.Core.Text.Corpus.Query as Corpus
  15 import Gargantext.Database.Admin.Types.Hyperdata (HyperdataDocument(..))
  16 import Protolude
  17 import qualified OpenAlex as OA
  18 import qualified OpenAlex.Types as OA
  19 import Servant.Client (ClientError)
  20
  21
  22 get :: Text
  23     -> Corpus.RawQuery
  24     -> Maybe Limit
  25     -> IO (Either ClientError (Maybe Integer, ConduitT () HyperdataDocument IO ()))
  26 get _email q _l = do
  27   eRes <- OA.fetchWorksC Nothing $ Just $ Corpus.getRawQuery q
  28   pure $ (\(len, docsC) -> (len, docsC .| takeC 1000 .| mapC toDoc)) <$> eRes
  29
  30 toDoc :: OA.Work -> HyperdataDocument
  31 toDoc (OA.Work { .. } ) =
  32   HyperdataDocument { _hd_bdd = Just "OpenAlex"
  33                     , _hd_doi = doi
  34                     , _hd_url = url
  35                     , _hd_uniqId = Just id
  36                     , _hd_uniqIdBdd = Just id
  37                     , _hd_page = firstPage biblio
  38                     , _hd_title = title
  39                     , _hd_authors = authors authorships
  40                     , _hd_institutes = institutes authorships
  41                     , _hd_source = source
  42                     , _hd_abstract = Just abstract_reconstructed
  43                     , _hd_publication_date = Just $ show publication_date
  44                     , _hd_publication_year = Just $ publication_year
  45                     , _hd_publication_month = Nothing  -- TODO
  46                     , _hd_publication_day = Nothing  -- TODO
  47                     , _hd_publication_hour = Nothing  -- TODO
  48                     , _hd_publication_minute = Nothing  -- TODO
  49                     , _hd_publication_second = Nothing  -- TODO
  50                     , _hd_language_iso2 = language }
  51       where
  52         firstPage :: OA.Biblio -> Maybe Int
  53         firstPage OA.Biblio { first_page } = maybe Nothing readMaybe $ T.unpack <$> first_page
  54
  55         authors :: [OA.Authorship] -> Maybe Text
  56         authors [] = Nothing
  57         authors aus = Just $ T.intercalate ", " (getDisplayName <$> aus)
  58           where
  59             getDisplayName :: OA.Authorship -> Text
  60             getDisplayName OA.Authorship { author = OA.DehydratedAuthor { display_name = dn } } = dn
  61
  62         institutes :: [OA.Authorship] -> Maybe Text
  63         institutes [] = Nothing
  64         institutes aus = Just $ T.intercalate ", " ((T.replace ", " " - ") . getInstitutesNames <$> aus)
  65           where
  66             getInstitutesNames OA.Authorship { institutions } = T.intercalate ", " $ getDisplayName <$> institutions
  67             getDisplayName :: OA.DehydratedInstitution -> Text
  68             getDisplayName OA.DehydratedInstitution { display_name = dn } = dn
  69
  70         source :: Maybe Text
  71         source = maybe Nothing getSource primary_location
  72           where
  73             getSource OA.Location { source = s } = getSourceDisplayName <$> s
  74             getSourceDisplayName OA.DehydratedSource { display_name = dn }  = dn