{-|
Module      : Gargantext.API.Node.DocumentsFromWriteNodes
Description :
Copyright   : (c) CNRS, 2017-Present
License     : AGPL + CECILL v3
Maintainer  : team@gargantext.org
Stability   : experimental
Portability : POSIX

-}

{-# LANGUAGE MonoLocalBinds     #-}
{-# LANGUAGE TemplateHaskell    #-}
{-# LANGUAGE TypeOperators      #-}

module Gargantext.API.Node.DocumentsFromWriteNodes
      where

import Conduit
import Control.Lens ((^.))
import Data.Aeson
import Data.Either (Either(..), rights)
import Data.Swagger
import qualified Data.Text as T
import Gargantext.API.Admin.EnvTypes (Env, GargJob(..))
import Gargantext.API.Admin.Orchestrator.Types (JobLog(..), AsyncJobs)
import Gargantext.API.Admin.Types (HasSettings)
import Gargantext.API.Job (jobLogSuccess, jobLogFailTotalWithMessage)
import Gargantext.API.Prelude (GargM, GargError)
import Gargantext.Core (Lang(..))
import Gargantext.Core.Text.Corpus.Parsers.FrameWrite
import Gargantext.Core.Text.Terms (TermType(..))
import Gargantext.Core.Types.Individu (User(..))
import Gargantext.Database.Action.Flow (flowDataText, DataText(..))
import Gargantext.Database.Action.Flow.Types (FlowCmdM)
import Gargantext.Database.Admin.Types.Hyperdata.Document (HyperdataDocument(..))
import Gargantext.Database.Admin.Types.Hyperdata.Frame
import Gargantext.Database.Admin.Types.Node
import Gargantext.Database.Query.Table.Node (getChildrenByType, getClosestParentIdByType', getNodeWith)
import Gargantext.Database.Schema.Node (node_hyperdata)
import qualified Data.List as List
import qualified Gargantext.Defaults as Defaults
import Gargantext.Prelude
import Gargantext.Utils.Jobs (serveJobsAPI)
import GHC.Generics (Generic)
import Servant
import Data.Text (Text)
import Gargantext.Core.Text.List.Social (FlowSocialListWith)
import Text.Read (readMaybe)
import Data.Maybe (fromMaybe)

------------------------------------------------------------------------
type API = Summary " Documents from Write nodes."
         :> AsyncJobs JobLog '[JSON] Params JobLog
------------------------------------------------------------------------
data Params = Params 
  { id         :: Int 
  , paragraphs :: Text
  , lang       :: Lang
  , selection  :: FlowSocialListWith
  }
  deriving (Generic, Show)
instance FromJSON Params where
  parseJSON = genericParseJSON defaultOptions
instance ToJSON Params where
  toJSON = genericToJSON defaultOptions
instance ToSchema Params
------------------------------------------------------------------------
api :: UserId -> NodeId -> ServerT API (GargM Env GargError)
api uId nId =
  serveJobsAPI DocumentFromWriteNodeJob $ \p log'' ->
      let
        log' x = do
          liftBase $ log'' x
      in documentsFromWriteNodes uId nId p (liftBase . log')

documentsFromWriteNodes :: (HasSettings env, FlowCmdM env err m)
    => UserId
    -> NodeId
    -> Params
    -> (JobLog -> m ())
    -> m JobLog
documentsFromWriteNodes uId nId Params { selection, lang, paragraphs } logStatus = do
  let jobLog = JobLog { _scst_succeeded = Just 1
                      , _scst_failed    = Just 0
                      , _scst_remaining = Just 1
                      , _scst_events    = Just []
                      }
  logStatus jobLog

  mcId <- getClosestParentIdByType' nId NodeCorpus
  cId <- case mcId of
    Just cId -> pure cId
    Nothing -> do
      let msg = T.pack $ "[G.A.N.DFWN] Node has no corpus parent: " <> show nId
      logStatus $ jobLogFailTotalWithMessage msg jobLog
      panic msg

  frameWriteIds <- getChildrenByType nId NodeFrameWrite

  -- https://write.frame.gargantext.org/<frame_id>/download
  frameWrites <- mapM (\id -> getNodeWith id (Proxy :: Proxy HyperdataFrame)) frameWriteIds

  frameWritesWithContents <- liftBase $
    mapM (\node -> do
             contents <- getHyperdataFrameContents (node ^. node_hyperdata)
             pure (node, contents)
         ) frameWrites
  
  let paragraphs' = readMaybe $ T.unpack paragraphs :: Maybe Int

  let parsedE = (\(node, contents) -> hyperdataDocumentFromFrameWrite (fromMaybe 7 paragraphs') (node ^. node_hyperdata, contents)) <$> frameWritesWithContents
                                              -- TODO hard coded param should be take
  let parsed = List.concat $ rights parsedE

  _ <- flowDataText (RootId (NodeId uId)) (DataNew (Just $ fromIntegral $ length parsed, yieldMany parsed)) (Multi lang) cId (Just selection) logStatus

  pure $ jobLogSuccess jobLog
------------------------------------------------------------------------
hyperdataDocumentFromFrameWrite :: Int -> (HyperdataFrame, T.Text) -> Either T.Text [HyperdataDocument]
hyperdataDocumentFromFrameWrite paragraphSize (HyperdataFrame { _hf_base, _hf_frame_id }, contents) =
  case parseLines contents of
    Left _ -> Left "Error parsing node"
    Right (Parsed { authors, contents = ctxts, date, source, title = t }) ->
      let authorJoinSingle (Author { firstName, lastName }) = T.concat [ lastName, ", ", firstName ]
          authors' = T.concat $ authorJoinSingle <$> authors 
          date' = (\(Date { year, month, day }) -> T.concat [ T.pack $ show year, "-"
                                                            , T.pack $ show month, "-"
                                                            , T.pack $ show day ]) <$> date
          year' = fromIntegral $ maybe Defaults.year (\(Date { year }) -> year) date
          month' = maybe Defaults.month (\(Date { month }) -> fromIntegral month) date
          day' = maybe Defaults.day (\(Date { day }) -> fromIntegral day) date in
      Right (List.map (\ctxt ->  HyperdataDocument { _hd_bdd = Just "FrameWrite"
                              , _hd_doi = Nothing
                              , _hd_url = Nothing
                              , _hd_uniqId = Nothing
                              , _hd_uniqIdBdd = Nothing
                              , _hd_page = Nothing
                              , _hd_title = Just t
                              , _hd_authors = Just authors'
                              , _hd_institutes = Nothing
                              , _hd_source = source
                              , _hd_abstract = Just ctxt
                              , _hd_publication_date = date'
                              , _hd_publication_year = Just year'
                              , _hd_publication_month = Just month'
                              , _hd_publication_day = Just day'
                              , _hd_publication_hour = Nothing
                              , _hd_publication_minute = Nothing
                              , _hd_publication_second = Nothing
                              , _hd_language_iso2 = Just $ T.pack $ show EN }
                      ) (text2paragraphs paragraphSize ctxts)
                  )