Merge pull request #1816 from eikek/1775-edit-extracted-text

Allow to change extracted text of attachments

modules/restapi/src/main/resources/docspell-openapi.yml

@@ -4823,6 +4823,72 @@ paths:
             application/json:
               schema:
                 $ref: "#/components/schemas/BasicResult"
+  /sec/attachment/{id}/extracted-text:
+    get:
+      operationId: "sec-attachment-get-extracted-text"
+      tags: [ Attachment ]
+      summary: Get the extracted text of the given attachment.
+      description: |
+        Returns the extracted text of the attachment with the given
+        id.
+      security:
+        - authTokenHeader: []
+      parameters:
+        - $ref: "#/components/parameters/id"
+      responses:
+        422:
+          description: BadRequest
+        200:
+          description: Ok
+          content:
+            application/json:
+              schema:
+                $ref: "#/components/schemas/OptionalText"
+    delete:
+      operationId: "sec-attachment-delete-extracted-text"
+      summary: Removes extracted text for the given attachment.
+      description: |
+        Removes any extracted text for the given attachment.
+      security:
+        - authTokenHeader: []
+      parameters:
+        - $ref: "#/components/parameters/id"
+      responses:
+        422:
+          description: BadRequest
+        200:
+          description: Ok
+          content:
+            application/json:
+              schema:
+                $ref: "#/components/schemas/BasicResult"    
+    post:
+      operationId: "sec-attachment-post-extracted-text"
+      tags: [ Attachment ]
+      summary: Changes the extracted text for an attachment
+      description: |
+        Changes the extracted text of the attachment with the given
+        id. The attachment must be part of an item that belongs to the
+        collective of the current user. This can be used to correct
+        poor ocr-ed files.
+      security:
+        - authTokenHeader: []
+      parameters:
+        - $ref: "#/components/parameters/id"
+      requestBody:
+        content:
+          application/json:
+            schema:
+              $ref: "#/components/schemas/OptionalText"
+      responses:
+        422:
+          description: BadRequest
+        200:
+          description: Ok
+          content:
+            application/json:
+              schema:
+                $ref: "#/components/schemas/BasicResult"
 
   /sec/attachments/delete:
     post:

modules/restserver/src/main/scala/docspell/restserver/routes/AttachmentRoutes.scala

@@ -6,8 +6,9 @@
 
 package docspell.restserver.routes
 
+import cats.data.OptionT
 import cats.effect._
-import cats.implicits._
+import cats.syntax.all._
 
 import docspell.backend.BackendApp
 import docspell.backend.auth.AuthToken
@@ -157,6 +158,40 @@ object AttachmentRoutes {
           resp <- Ok(Conversions.basicResult(res, "Name updated."))
         } yield resp
 
+      case req @ POST -> Root / Ident(id) / "extracted-text" =>
+        (for {
+          itemId <- OptionT(
+            backend.itemSearch.findAttachment(id, user.account.collectiveId)
+          ).map(_.ra.itemId)
+          nn <- OptionT.liftF(req.as[OptionalText])
+          newText = nn.text.getOrElse("").pure[F]
+          _ <- OptionT.liftF(
+            backend.attachment
+              .setExtractedText(user.account.collectiveId, itemId, id, newText)
+          )
+          resp <- OptionT.liftF(Ok(BasicResult(true, "Extracted text updated.")))
+        } yield resp).getOrElseF(NotFound(BasicResult(false, "Attachment not found")))
+
+      case DELETE -> Root / Ident(id) / "extracted-text" =>
+        (for {
+          itemId <- OptionT(
+            backend.itemSearch.findAttachment(id, user.account.collectiveId)
+          ).map(_.ra.itemId)
+          _ <- OptionT.liftF(
+            backend.attachment
+              .setExtractedText(user.account.collectiveId, itemId, id, "".pure[F])
+          )
+          resp <- OptionT.liftF(Ok(BasicResult(true, "Extracted text cleared.")))
+        } yield resp).getOrElseF(NotFound())
+
+      case GET -> Root / Ident(id) / "extracted-text" =>
+        (for {
+          meta <- OptionT(
+            backend.itemSearch.findAttachmentMeta(id, user.account.collectiveId)
+          )
+          resp <- OptionT.liftF(Ok(OptionalText(meta.content)))
+        } yield resp).getOrElseF(NotFound(BasicResult(false, "Attachment not found")))
+
       case DELETE -> Root / Ident(id) =>
         for {
           n <- backend.item.deleteAttachment(id, user.account.collectiveId)

modules/webapp/src/main/elm/Api.elm

@@ -159,6 +159,7 @@ module Api exposing
     , searchShare
     , searchShareStats
     , sendMail
+    , setAttachmentExtractedText
     , setAttachmentName
     , setCollectiveSettings
     , setConcEquip
@@ -2049,6 +2050,21 @@ setAttachmentName flags attachId newName receive =
         }
 
 
+setAttachmentExtractedText :
+    Flags
+    -> String
+    -> Maybe String
+    -> (Result Http.Error BasicResult -> msg)
+    -> Cmd msg
+setAttachmentExtractedText flags attachId newName receive =
+    Http2.authPost
+        { url = flags.config.baseUrl ++ "/api/v1/sec/attachment/" ++ attachId ++ "/extracted-text"
+        , account = getAccount flags
+        , body = Http.jsonBody (Api.Model.OptionalText.encode (OptionalText newName))
+        , expect = Http.expectJson receive Api.Model.BasicResult.decoder
+        }
+
+
 moveAttachmentBefore :
     Flags
     -> String

modules/webapp/src/main/elm/Comp/AttachmentMeta.elm

@@ -15,12 +15,14 @@ module Comp.AttachmentMeta exposing
 
 import Api
 import Api.Model.AttachmentMeta exposing (AttachmentMeta)
+import Api.Model.BasicResult exposing (BasicResult)
 import Api.Model.ItemProposals exposing (ItemProposals)
 import Api.Model.Label exposing (Label)
 import Comp.Basic as B
 import Data.Flags exposing (Flags)
 import Html exposing (..)
 import Html.Attributes exposing (..)
+import Html.Events exposing (onClick, onInput)
 import Http
 import Messages.Comp.AttachmentMeta exposing (Texts)
 import Styles as S
@@ -28,13 +30,20 @@ import Styles as S
 
 type alias Model =
     { id : String
-    , meta : DataResult AttachmentMeta
+    , meta : DataResult
     }
 
 
-type DataResult a
+type alias EditModel =
+    { meta : AttachmentMeta
+    , text : String
+    }
+
+
+type DataResult
     = NotAvailable
-    | Success a
+    | Success AttachmentMeta
+    | Editing EditModel
     | HttpFailure Http.Error
 
 
@@ -54,16 +63,64 @@ init flags id =
 
 type Msg
     = MetaResp (Result Http.Error AttachmentMeta)
+    | SaveResp String (Result Http.Error BasicResult)
+    | ToggleEdit
+    | SaveExtractedText
+    | SetExtractedText String
 
 
-update : Msg -> Model -> Model
-update msg model =
+update : Flags -> Msg -> Model -> ( Model, Cmd Msg )
+update flags msg model =
     case msg of
         MetaResp (Ok am) ->
-            { model | meta = Success am }
+            ( { model | meta = Success am }, Cmd.none )
 
         MetaResp (Err err) ->
-            { model | meta = HttpFailure err }
+            ( { model | meta = HttpFailure err }, Cmd.none )
+
+        SaveResp newText (Ok result) ->
+            if result.success then
+                case model.meta of
+                    Editing { meta } ->
+                        ( { model | meta = Success { meta | content = newText } }, Cmd.none )
+
+                    _ ->
+                        ( model, Cmd.none )
+
+            else
+                ( model, Cmd.none )
+
+        SaveResp _ (Err err) ->
+            ( { model | meta = HttpFailure err }, Cmd.none )
+
+        ToggleEdit ->
+            case model.meta of
+                Editing m ->
+                    ( { model | meta = Success m.meta }, Cmd.none )
+
+                Success m ->
+                    ( { model | meta = Editing { meta = m, text = m.content } }, Cmd.none )
+
+                _ ->
+                    ( model, Cmd.none )
+
+        SaveExtractedText ->
+            case model.meta of
+                Editing em ->
+                    ( model
+                    , Api.setAttachmentExtractedText flags model.id (Just em.text) (SaveResp em.text)
+                    )
+
+                _ ->
+                    ( model, Cmd.none )
+
+        SetExtractedText txt ->
+            case model.meta of
+                Editing em ->
+                    ( { model | meta = Editing { em | text = txt } }, Cmd.none )
+
+                _ ->
+                    ( model, Cmd.none )
 
 
 
@@ -89,19 +146,55 @@ view2 texts attrs model =
                     ]
 
             Success data ->
-                viewData2 texts data
+                viewData2 texts data Nothing
+
+            Editing em ->
+                viewData2 texts em.meta (Just em.text)
         ]
 
 
-viewData2 : Texts -> AttachmentMeta -> Html Msg
-viewData2 texts meta =
+viewData2 : Texts -> AttachmentMeta -> Maybe String -> Html Msg
+viewData2 texts meta maybeText =
     div [ class "flex flex-col" ]
-        [ div [ class "text-lg font-bold" ]
-            [ text texts.content
-            ]
-        , div [ class "px-2 py-2 text-sm bg-yellow-50 dark:bg-stone-800 break-words whitespace-pre max-h-80 overflow-auto" ]
-            [ text meta.content
+        [ div [ class "flex flex-row items-center" ]
+            [ div [ class "text-lg font-bold flex flex-grow" ]
+                [ text texts.content
+                ]
+            , case maybeText of
+                Nothing ->
+                    div [ class "flex text-sm" ]
+                        [ a [ href "#", class S.link, onClick ToggleEdit ]
+                            [ i [ class "fa fa-edit pr-1" ] []
+                            , text texts.basics.edit
+                            ]
+                        ]
+
+                Just _ ->
+                    div [ class "flex text-sm" ]
+                        [ a [ href "#", class S.link, onClick ToggleEdit ]
+                            [ text texts.basics.cancel
+                            ]
+                        , span [ class "px-2" ] [ text "•" ]
+                        , a [ href "#", class S.link, onClick SaveExtractedText ]
+                            [ i [ class "fa fa-save pr-1" ] []
+                            , text texts.basics.submit
+                            ]
+                        ]
             ]
+        , case maybeText of
+            Nothing ->
+                div [ class "px-2 py-2 text-sm bg-yellow-50 dark:bg-stone-800 break-words whitespace-pre max-h-80 overflow-auto" ]
+                    [ text meta.content
+                    ]
+
+            Just em ->
+                textarea
+                    [ class "px-2 py-2 text-sm bg-yellow-50 dark:bg-stone-800 break-words whitespace-pre h-80 overflow-auto"
+                    , value em
+                    , onInput SetExtractedText
+                    , rows 10
+                    ]
+                    []
         , div [ class "text-lg font-bold mt-2" ]
             [ text texts.labels
             ]

modules/webapp/src/main/elm/Comp/ItemDetail/Update.elm

@@ -991,11 +991,13 @@ update inav env msg model =
             case Dict.get id model.attachMeta of
                 Just cm ->
                     let
-                        am =
-                            Comp.AttachmentMeta.update lmsg cm
+                        ( am, ac ) =
+                            Comp.AttachmentMeta.update env.flags lmsg cm
                     in
-                    resultModel
-                        { model | attachMeta = Dict.insert id am model.attachMeta }
+                    resultModelCmd
+                        ( { model | attachMeta = Dict.insert id am model.attachMeta }
+                        , Cmd.map (AttachMetaMsg id) ac
+                        )
 
                 Nothing ->
                     resultModel model