Add more properties to upload meta data

build.sbt

@@ -124,9 +124,8 @@ val buildInfoSettings = Seq(
 val openapiScalaSettings = Seq(
   openapiScalaConfig := ScalaConfig()
     .withJson(ScalaJson.circeSemiauto)
-    .addMapping(CustomMapping.forType({
-      case TypeDef("LocalDateTime", _) =>
-        TypeDef("Timestamp", Imports("docspell.common.Timestamp"))
+    .addMapping(CustomMapping.forType({ case TypeDef("LocalDateTime", _) =>
+      TypeDef("Timestamp", Imports("docspell.common.Timestamp"))
     }))
     .addMapping(CustomMapping.forFormatType({
       case "ident" =>
@@ -182,6 +181,8 @@ val openapiScalaSettings = Seq(
               )
             )
           )
+      case "glob" =>
+        field => field.copy(typeDef = TypeDef("Glob", Imports("docspell.common.Glob")))
     }))
 )
 
@@ -595,11 +596,10 @@ def copyWebjarResources(
   src.flatMap { dir =>
     if (dir.isDirectory) {
       val files = (dir ** "*").filter(_.isFile).get.pair(Path.relativeTo(dir))
-      files.flatMap {
-        case (f, name) =>
-          val target = targetDir / name
-          IO.createDirectories(Seq(target.getParentFile))
-          copyWithGZ(f, target)
+      files.flatMap { case (f, name) =>
+        val target = targetDir / name
+        IO.createDirectories(Seq(target.getParentFile))
+        copyWithGZ(f, target)
       }
     } else {
       val target = targetDir / dir.name
@@ -633,11 +633,13 @@ def compileElm(
 }
 
 def createWebjarSource(wj: Seq[ModuleID], out: File): Seq[File] = {
-  val target = out / "Webjars.scala"
+  val target   = out / "Webjars.scala"
   val badChars = "-.".toSet
   val fields = wj
     .map(m =>
-      s"""val ${m.name.toLowerCase.filter(c => !badChars.contains(c))} = "/${m.name}/${m.revision}" """
+      s"""val ${m.name.toLowerCase.filter(c =>
+        !badChars.contains(c)
+      )} = "/${m.name}/${m.revision}" """
     )
     .mkString("\n\n")
   val content = s"""package docspell.restserver.webapp

modules/backend/src/main/scala/docspell/backend/ops/OUpload.scala

@@ -60,7 +60,9 @@ object OUpload {
       sourceAbbrev: String,
       folderId: Option[Ident],
       validFileTypes: Seq[MimeType],
-      skipDuplicates: Boolean
+      skipDuplicates: Boolean,
+      fileFilter: Option[Glob],
+      tags: List[String]
   )
 
   case class UploadData[F[_]](
@@ -127,7 +129,9 @@ object OUpload {
             data.meta.sourceAbbrev,
             data.meta.folderId,
             data.meta.validFileTypes,
-            data.meta.skipDuplicates
+            data.meta.skipDuplicates,
+            data.meta.fileFilter,
+            data.meta.tags.some
           )
           args =
             if (data.multiple) files.map(f => ProcessItemArgs(meta, List(f)))

modules/common/src/main/scala/docspell/common/ProcessItemArgs.scala

@@ -38,7 +38,9 @@ object ProcessItemArgs {
       sourceAbbrev: String,
       folderId: Option[Ident],
       validFileTypes: Seq[MimeType],
-      skipDuplicate: Boolean
+      skipDuplicate: Boolean,
+      fileFilter: Option[Glob],
+      tags: Option[List[String]]
   )
 
   object ProcessMeta {

modules/common/src/main/scala/docspell/common/ScanMailboxArgs.scala

@@ -29,7 +29,11 @@ case class ScanMailboxArgs(
     // set the direction when submitting
     direction: Option[Direction],
     // set a folder for items
-    itemFolder: Option[Ident]
+    itemFolder: Option[Ident],
+    // set a filter for files when importing archives
+    fileFilter: Option[Glob],
+    // set a list of tags to apply to new item
+    tags: Option[List[String]]
 )
 
 object ScanMailboxArgs {

modules/joex/src/main/scala/docspell/joex/process/ReProcessItem.scala

@@ -91,7 +91,9 @@ object ReProcessItem {
               "",   //source-id
               None, //folder
               Seq.empty,
-              false
+              false,
+              None,
+              None
             ),
             Nil
           ).pure[F]

modules/joex/src/main/scala/docspell/joex/process/TextExtraction.scala

@@ -47,7 +47,9 @@ object TextExtraction {
         _   <- fts.indexData(ctx.logger, (idxItem +: txt.map(_.td)).toSeq: _*)
         dur <- start
         _   <- ctx.logger.info(s"Text extraction finished in ${dur.formatExact}")
-      } yield item.copy(metas = txt.map(_.am), tags = txt.flatMap(_.tags).distinct.toList)
+      } yield item
+        .copy(metas = txt.map(_.am))
+        .appendTags(txt.flatMap(_.tags).distinct.toList)
     }
 
   // --  helpers

modules/joex/src/main/scala/docspell/joex/scanmailbox/ScanMailboxTask.scala

@@ -255,7 +255,9 @@ object ScanMailboxTask {
           s"mailbox-${ctx.args.account.user.id}",
           args.itemFolder,
           Seq.empty,
-          true
+          true,
+          args.fileFilter,
+          args.tags.getOrElse(Nil)
         )
         data = OUpload.UploadData(
           multiple = false,

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

@@ -3499,6 +3499,14 @@ components:
             The folder id that is applied to items resulting from
             importing mails. If the folder id is not valid when the
             task executes, items have no folder set.
+        tags:
+          $ref: "#/components/schemas/StringList"
+        fileFilter:
+          description: |
+            A glob to filter attachments to import.
+          type: string
+          format: glob
+
     ImapSettingsList:
       description: |
         A list of user email settings.
@@ -4256,6 +4264,18 @@ components:
         A folderId can be given, the item is placed into this folder
         after creation.
 
+        The `fileFilter` is an optional glob for filtering files to
+        import. Only applicable if archive files are uploaded. It
+        applies to all of them. For example, to only import pdf files
+        when uploading e-mails, use `*.pdf`. If the pattern doesn't
+        contain a slash `/`, then it is applied to all file names.
+        Otherwise it is applied to the complete path in the archive
+        (useful for zip files). Note that the archive file itself is
+        always saved completely, too.
+
+        The `tags` input allows to provide tags that should be applied
+        to the item being created. This only works if the tags already
+        exist. It is possible to specify their ids or names.
       required:
         - multiple
       properties:
@@ -4271,6 +4291,11 @@ components:
         skipDuplicates:
           type: boolean
           default: false
+        tags:
+          $ref: "#/components/schemas/StringList"
+        fileFilter:
+          type: string
+          format: glob
 
     Collective:
       description: |

modules/restserver/src/main/scala/docspell/restserver/conv/Conversions.scala

@@ -310,13 +310,15 @@ trait Conversions {
               sourceName,
               m.folder,
               validFileTypes,
-              m.skipDuplicates.getOrElse(false)
+              m.skipDuplicates.getOrElse(false),
+              m.fileFilter,
+              m.tags.map(_.items).getOrElse(Nil)
             )
           )
         )
       )
       .getOrElse(
-        (true, UploadMeta(None, sourceName, None, validFileTypes, false)).pure[F]
+        (true, UploadMeta(None, sourceName, None, validFileTypes, false, None, Nil)).pure[F]
       )
 
     val files = mp.parts

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

@@ -113,7 +113,9 @@ object ScanMailboxRoutes {
           settings.targetFolder,
           settings.deleteMail,
           settings.direction,
-          settings.itemFolder
+          settings.itemFolder,
+          settings.fileFilter,
+          settings.tags.map(_.items)
         )
       )
     )
@@ -141,6 +143,8 @@ object ScanMailboxRoutes {
       task.args.targetFolder,
       task.args.deleteMail,
       task.args.direction,
-      task.args.itemFolder
+      task.args.itemFolder,
+      task.args.tags.map(StringList.apply),
+      task.args.fileFilter
     )
 }