src/Gargantext/Text/Corpus/API/Istex.hs

   1 {-|
   2 Module      : Gargantext.Text.Corpus.API.Istex
   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.Text.Corpus.API.Istex
  14     where
  15
  16 import Data.Either (either)
  17 import Data.List (concat)
  18 import Data.Maybe
  19 import Data.Maybe (catMaybes)
  20 import Data.Text (Text, pack)
  21 import Gargantext.Core (Lang(..))
  22 import Gargantext.Database.Admin.Types.Node (HyperdataDocument(..))
  23 import Gargantext.Prelude
  24 import qualified Gargantext.Text.Corpus.Parsers.Date as Date
  25 import qualified ISTEX        as ISTEX
  26 import qualified ISTEX.Client as ISTEX
  27
  28 get :: Lang -> Text -> Maybe Integer -> IO [HyperdataDocument]
  29 get la q ml = do
  30   docs <- ISTEX.getMetadataWith q (fromIntegral <$> ml)
  31   either (panic . pack . show) (toDoc' la) docs
  32
  33 toDoc' :: Lang -> ISTEX.Documents -> IO [HyperdataDocument]
  34 toDoc' la docs' = do
  35   --printDebug "ISTEX" (ISTEX._documents_total docs')
  36   mapM (toDoc la) (ISTEX._documents_hits docs')
  37
  38 -- | TODO remove dateSplit here
  39 -- TODO current year as default
  40 toDoc :: Lang -> ISTEX.Document -> IO HyperdataDocument
  41 toDoc la (ISTEX.Document i t a ab d s) = do
  42   (utctime, (pub_year, pub_month, pub_day)) <- Date.dateSplit la (maybe (Just "2019") (Just . pack . show) d)
  43   pure $ HyperdataDocument (Just "Istex")
  44                    (Just i)
  45                    Nothing
  46                    Nothing
  47                    Nothing
  48                    Nothing
  49                    t
  50                    (Just $ foldl (\x y -> x <> ", " <> y) "" (map ISTEX._author_name a))
  51                    (Just $ foldl (\x y -> x <> ", " <> y) "" (concat $ (map ISTEX._author_affiliations) a))
  52                    (Just $ foldl (\x y -> x <> ", " <> y) "" (catMaybes $ map ISTEX._source_title s))
  53                    ab
  54                    (fmap (pack . show) utctime)
  55                    pub_year
  56                    pub_month
  57                    pub_day
  58                    Nothing
  59                    Nothing
  60                    Nothing
  61                   (Just $ (pack . show) la)