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