Fix: If a file remains stable but has zero size after the stability window has passed, drop it from tracking (#14047 )

A slow writer might create a zero byte file, then take longer than the window to finish the write. We would then queue a zero byte file for consumption and race the writer to most likely fail due to still being empty. Instead, drop zero size files at yeild time. The slow writer may or may not finish, but if it does, there will be a Change.modified event fired again
Auto translate strings
2026-09-09 19:27:59 +00:00 · 2026-09-09 16:24:17 +00:00 · 2026-09-09 16:04:18 +00:00 · 2026-09-09 16:02:43 +00:00 · 2026-09-09 15:50:33 +00:00 · 2026-09-09 08:48:59 -07:00
33 changed files with 1393 additions and 888 deletions
@@ -112,22 +112,6 @@

                <pngx-input-check i18n-title title="Use 'slim' sidebar (icons only)" formControlName="slimSidebarEnabled"></pngx-input-check>

-                <p class="mb-2 mt-3" i18n>Sidebar items to show:</p>
-                @for (option of sidebarItemOptions; track option.id) {
-                  <div class="form-check">
-                    <input
-                      class="form-check-input"
-                      type="checkbox"
-                      [id]="'sidebar-item-setting-' + option.id"
-                      [checked]="isSidebarItemShown(option.id)"
-                      (change)="toggleSidebarItem(option.id, $event.target.checked)"
-                    />
-                    <label class="form-check-label" [for]="'sidebar-item-setting-' + option.id">
-                      {{ option.label }}
-                    </label>
-                  </div>
-                }
-
              </div>
            </div>

@@ -24,7 +24,7 @@ import {
  SystemStatus,
  SystemStatusItemStatus,
 } from 'src/app/data/system-status'
-import { HideableSidebarItemID, SETTINGS_KEYS } from 'src/app/data/ui-settings'
+import { SETTINGS_KEYS } from 'src/app/data/ui-settings'
 import { IfOwnerDirective } from 'src/app/directives/if-owner.directive'
 import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { PermissionsGuard } from 'src/app/guards/permissions.guard'
@@ -209,45 +209,6 @@ describe('SettingsComponent', () => {
    fixture.detectChanges()
  }

-  it('supports configuring sidebar items and canceling changes', () => {
-    completeSetup()
-
-    component.toggleSidebarItem(HideableSidebarItemID.Workflows, false)
-    fixture.detectChanges()
-
-    expect(component.settingsForm.value.sidebarHiddenItems).toContain(
-      HideableSidebarItemID.Workflows
-    )
-
-    settingsService.updateSidebarItemVisibility(
-      HideableSidebarItemID.Mail,
-      false
-    )
-
-    expect(component.settingsForm.value.sidebarHiddenItems).toContain(
-      HideableSidebarItemID.Mail
-    )
-
-    component.reset()
-
-    expect(component.settingsForm.value.sidebarHiddenItems).not.toContain(
-      HideableSidebarItemID.Workflows
-    )
-    expect(component.settingsForm.value.sidebarHiddenItems).not.toContain(
-      HideableSidebarItemID.Mail
-    )
-  })
-
-  it('enables sidebar item controls on general settings until destroyed', () => {
-    completeSetup()
-
-    expect(settingsService.organizingSidebarItems()).toBe(true)
-
-    component.ngOnDestroy()
-
-    expect(settingsService.organizingSidebarItems()).toBe(false)
-  })
-
  it('should support tabbed settings & change URL, prevent navigation if dirty confirmation rejected', async () => {
    completeSetup()
    const navigateSpy = jest.spyOn(router, 'navigate')
@@ -288,7 +249,6 @@ describe('SettingsComponent', () => {

  it('should support save local settings updating appearance settings and calling API, show error', () => {
    completeSetup()
-    component.toggleSidebarItem(HideableSidebarItemID.Workflows, false)
    const toastErrorSpy = jest.spyOn(toastService, 'showError')
    const toastSpy = jest.spyOn(toastService, 'show')
    const storeSpy = jest.spyOn(settingsService, 'storeSettings')
@@ -307,10 +267,7 @@ describe('SettingsComponent', () => {
    expect(toastErrorSpy).toHaveBeenCalled()
    expect(storeSpy).toHaveBeenCalled()
    expect(appearanceSettingsSpy).not.toHaveBeenCalled()
-    expect(setSpy).toHaveBeenCalledTimes(34)
-    expect(setSpy).toHaveBeenCalledWith(SETTINGS_KEYS.SIDEBAR_HIDDEN_ITEMS, [
-      HideableSidebarItemID.Workflows,
-    ])
+    expect(setSpy).toHaveBeenCalledTimes(33)

    // succeed
    storeSpy.mockReturnValueOnce(of(true))
@@ -39,12 +39,7 @@ import {
  SystemStatus,
  SystemStatusItemStatus,
 } from 'src/app/data/system-status'
-import {
-  GlobalSearchType,
-  HIDEABLE_SIDEBAR_ITEM_IDS,
-  HideableSidebarItemID,
-  SETTINGS_KEYS,
-} from 'src/app/data/ui-settings'
+import { GlobalSearchType, SETTINGS_KEYS } from 'src/app/data/ui-settings'
 import { User } from 'src/app/data/user'
 import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { CustomDatePipe } from 'src/app/pipes/custom-date.pipe'
@@ -107,14 +102,6 @@ const documentDetailFieldOptions = [
  { id: DocumentDetailFieldID.Tags, label: $localize`Tags` },
 ]

-const sidebarItemLabels: Record<HideableSidebarItemID, string> = {
-  [HideableSidebarItemID.Dashboard]: $localize`Dashboard`,
-  [HideableSidebarItemID.SavedViews]: $localize`Saved Views`,
-  [HideableSidebarItemID.Workflows]: $localize`Workflows`,
-  [HideableSidebarItemID.Mail]: $localize`Mail`,
-  [HideableSidebarItemID.Documentation]: $localize`Documentation`,
-}
-
@Component({
  selector: 'pngx-settings',
  templateUrl: './settings.component.html',
@@ -162,7 +149,6 @@ export class SettingsComponent
    bulkEditApplyOnClose: new FormControl(null),
    documentListItemPerPage: new FormControl(null),
    slimSidebarEnabled: new FormControl(null),
-    sidebarHiddenItems: new FormControl<HideableSidebarItemID[]>([]),
    darkModeUseSystem: new FormControl(null),
    darkModeEnabled: new FormControl(null),
    darkModeInvertThumbs: new FormControl(null),
@@ -200,7 +186,6 @@ export class SettingsComponent

  store: BehaviorSubject<any>
  storeSub: Subscription
-  sidebarItemsSub: Subscription
  isDirty$: Observable<boolean>
  isDirty: boolean = false
  unsubscribeNotifier: Subject<any> = new Subject()
@@ -218,10 +203,6 @@ export class SettingsComponent
  public readonly PdfEditorEditMode = PdfEditorEditMode

  public readonly documentDetailFieldOptions = documentDetailFieldOptions
-  public readonly sidebarItemOptions = HIDEABLE_SIDEBAR_ITEM_IDS.map((id) => ({
-    id,
-    label: sidebarItemLabels[id],
-  }))

  get systemStatusHasErrors(): boolean {
    const status = this.systemStatus()
@@ -249,10 +230,6 @@ export class SettingsComponent

  constructor() {
    super()
-    this.sidebarItemsSub =
-      this.settings.sidebarHiddenItemsEditingChanged.subscribe((hiddenItems) =>
-        this.settingsForm.controls.sidebarHiddenItems.setValue(hiddenItems)
-      )
    this.settings.settingsSaved.subscribe(() => {
      if (!this.savePending) this.initialize()
      this.savedViewsService.maybeRefreshDocumentCounts()
@@ -302,21 +279,14 @@ export class SettingsComponent

    this.activatedRoute.paramMap.subscribe((paramMap) => {
      const section = paramMap.get('section')
-      let navID = SettingsNavIDs.General
      if (section) {
        const navIDKey: string = Object.keys(SettingsNavIDs).find(
          (navID) => navID.toLowerCase() == section
        )
        if (navIDKey) {
-          navID = SettingsNavIDs[navIDKey]
+          this.activeNavID.set(SettingsNavIDs[navIDKey])
        }
      }
-      this.activeNavID.set(navID)
-      this.settings.sidebarHiddenItemsEditing.set(
-        navID === SettingsNavIDs.General
-          ? [...this.settingsForm.controls.sidebarHiddenItems.value]
-          : null
-      )
    })
  }

@@ -340,7 +310,6 @@ export class SettingsComponent
        SETTINGS_KEYS.DOCUMENT_LIST_SIZE
      ),
      slimSidebarEnabled: this.settings.get(SETTINGS_KEYS.SLIM_SIDEBAR),
-      sidebarHiddenItems: this.settings.get(SETTINGS_KEYS.SIDEBAR_HIDDEN_ITEMS),
      darkModeUseSystem: this.settings.get(SETTINGS_KEYS.DARK_MODE_USE_SYSTEM),
      darkModeEnabled: this.settings.get(SETTINGS_KEYS.DARK_MODE_ENABLED),
      darkModeInvertThumbs: this.settings.get(
@@ -467,12 +436,6 @@ export class SettingsComponent
      this.settingsForm.patchValue(currentFormValue)
    }

-    if (this.settings.organizingSidebarItems()) {
-      this.settings.sidebarHiddenItemsEditing.set([
-        ...this.settingsForm.controls.sidebarHiddenItems.value,
-      ])
-    }
-
    if (this.canViewSystemStatus) {
      this.systemStatusService.get().subscribe((status) => {
        this.systemStatus.set(status)
@@ -481,18 +444,8 @@ export class SettingsComponent
  }

  ngOnDestroy() {
-    this.settings.sidebarHiddenItemsEditing.set(null)
    if (this.isDirty) this.settings.updateAppearanceSettings() // in case user changed appearance but didn't save
    this.storeSub && this.storeSub.unsubscribe()
-    this.sidebarItemsSub.unsubscribe()
-  }
-
-  isSidebarItemShown(item: HideableSidebarItemID): boolean {
-    return !(this.settingsForm.value.sidebarHiddenItems || []).includes(item)
-  }
-
-  toggleSidebarItem(item: HideableSidebarItemID, checked: boolean): void {
-    this.settings.updateSidebarItemVisibility(item, checked)
  }

  public saveSettings() {
@@ -520,10 +473,6 @@ export class SettingsComponent
      SETTINGS_KEYS.SLIM_SIDEBAR,
      this.settingsForm.value.slimSidebarEnabled
    )
-    this.settings.set(
-      SETTINGS_KEYS.SIDEBAR_HIDDEN_ITEMS,
-      this.settingsForm.value.sidebarHiddenItems
-    )
    this.settings.set(
      SETTINGS_KEYS.DARK_MODE_USE_SYSTEM,
      this.settingsForm.value.darkModeUseSystem
@@ -683,11 +632,6 @@ export class SettingsComponent

  reset() {
    this.settingsForm.patchValue(this.store.getValue())
-    if (this.settings.organizingSidebarItems()) {
-      this.settings.sidebarHiddenItemsEditing.set([
-        ...this.settingsForm.controls.sidebarHiddenItems.value,
-      ])
-    }
  }

  clearThemeColor() {
@@ -86,15 +86,12 @@
      }
      <div class="sidebar-sticky pt-3 pb-1 d-flex flex-column justify-space-around">
        <ul class="nav flex-column">
-          <li class="nav-item app-link position-relative" [class.d-none]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.Dashboard) && !settingsService.organizingSidebarItems()">
-            <a class="nav-link" [class.opacity-50]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.Dashboard)" [class.pe-5]="settingsService.organizingSidebarItems() && !slimSidebarEnabled && !slimSidebarAnimating()" routerLink="dashboard" routerLinkActive="active" (click)="closeMenu()"
+          <li class="nav-item app-link">
+            <a class="nav-link" routerLink="dashboard" routerLinkActive="active" (click)="closeMenu()"
              ngbPopover="Dashboard" i18n-ngbPopover [disablePopover]="!slimSidebarPopoversEnabled" placement="end"
              container="body" triggers="mouseenter:mouseleave" popoverClass="popover-slim">
              <i-bs class="me-2" name="house"></i-bs><span class="nav-link-label"><ng-container i18n>Dashboard</ng-container></span>
            </a>
-            @if (settingsService.organizingSidebarItems()) {
-              <pngx-input-switch class="position-absolute top-50 end-0 translate-middle-y me-1" [class.d-none]="slimSidebarEnabled || slimSidebarAnimating()" [compact]="true" title="Dashboard" i18n-title [ngModel]="!settingsService.sidebarItemIsHidden(HideableSidebarItemID.Dashboard)" (ngModelChange)="toggleSidebarItem(HideableSidebarItemID.Dashboard, $event)"></pngx-input-switch>
-            }
          </li>
          <li class="nav-item app-link" *pngxIfPermissions="{ action: PermissionAction.View, type: PermissionType.Document }">
            <a class="nav-link" routerLink="documents" routerLinkActive="active"
@@ -240,38 +237,29 @@
                </div>
              </li>
            }
-            <li class="nav-item app-link position-relative" [class.d-none]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.SavedViews) && !settingsService.organizingSidebarItems()" *pngxIfPermissions="{ action: PermissionAction.View, type: PermissionType.SavedView }">
-              <a class="nav-link" [class.opacity-50]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.SavedViews)" [class.pe-5]="settingsService.organizingSidebarItems() && !slimSidebarEnabled && !slimSidebarAnimating()" routerLink="savedviews" routerLinkActive="active" (click)="closeMenu()"
+            <li class="nav-item app-link" *pngxIfPermissions="{ action: PermissionAction.View, type: PermissionType.SavedView }">
+              <a class="nav-link" routerLink="savedviews" routerLinkActive="active" (click)="closeMenu()"
                ngbPopover="Saved Views" i18n-ngbPopover [disablePopover]="!slimSidebarPopoversEnabled" placement="end"
                container="body" triggers="mouseenter:mouseleave" popoverClass="popover-slim">
                <i-bs class="me-2" name="window-stack"></i-bs><span class="nav-link-label"><ng-container i18n>Saved Views</ng-container></span>
              </a>
-              @if (settingsService.organizingSidebarItems()) {
-                <pngx-input-switch class="position-absolute top-50 end-0 translate-middle-y me-1" [class.d-none]="slimSidebarEnabled || slimSidebarAnimating()" [compact]="true" title="Saved Views" i18n-title [ngModel]="!settingsService.sidebarItemIsHidden(HideableSidebarItemID.SavedViews)" (ngModelChange)="toggleSidebarItem(HideableSidebarItemID.SavedViews, $event)"></pngx-input-switch>
-              }
            </li>
-            <li class="nav-item app-link position-relative" [class.d-none]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.Workflows) && !settingsService.organizingSidebarItems()"
+            <li class="nav-item app-link"
              *pngxIfPermissions="{ action: PermissionAction.View, type: PermissionType.Workflow }"
              tourAnchor="tour.workflows">
-              <a class="nav-link" [class.opacity-50]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.Workflows)" [class.pe-5]="settingsService.organizingSidebarItems() && !slimSidebarEnabled && !slimSidebarAnimating()" routerLink="workflows" routerLinkActive="active" (click)="closeMenu()"
+              <a class="nav-link" routerLink="workflows" routerLinkActive="active" (click)="closeMenu()"
                ngbPopover="Workflows" i18n-ngbPopover [disablePopover]="!slimSidebarPopoversEnabled" placement="end"
                container="body" triggers="mouseenter:mouseleave" popoverClass="popover-slim">
                <i-bs class="me-2" name="boxes"></i-bs><span class="nav-link-label"><ng-container i18n>Workflows</ng-container></span>
              </a>
-              @if (settingsService.organizingSidebarItems()) {
-                <pngx-input-switch class="position-absolute top-50 end-0 translate-middle-y me-1" [class.d-none]="slimSidebarEnabled || slimSidebarAnimating()" [compact]="true" title="Workflows" i18n-title [ngModel]="!settingsService.sidebarItemIsHidden(HideableSidebarItemID.Workflows)" (ngModelChange)="toggleSidebarItem(HideableSidebarItemID.Workflows, $event)"></pngx-input-switch>
-              }
            </li>
-            <li class="nav-item app-link position-relative" [class.d-none]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.Mail) && !settingsService.organizingSidebarItems()" *pngxIfPermissions="{ action: PermissionAction.View, type: PermissionType.MailAccount }"
+            <li class="nav-item app-link" *pngxIfPermissions="{ action: PermissionAction.View, type: PermissionType.MailAccount }"
              tourAnchor="tour.mail">
-              <a class="nav-link" [class.opacity-50]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.Mail)" [class.pe-5]="settingsService.organizingSidebarItems() && !slimSidebarEnabled && !slimSidebarAnimating()" routerLink="mail" routerLinkActive="active" (click)="closeMenu()" ngbPopover="Mail"
+              <a class="nav-link" routerLink="mail" routerLinkActive="active" (click)="closeMenu()" ngbPopover="Mail"
                i18n-ngbPopover [disablePopover]="!slimSidebarPopoversEnabled" placement="end" container="body"
                triggers="mouseenter:mouseleave" popoverClass="popover-slim">
                <i-bs class="me-2" name="envelope"></i-bs><span class="nav-link-label"><ng-container i18n>Mail</ng-container></span>
              </a>
-              @if (settingsService.organizingSidebarItems()) {
-                <pngx-input-switch class="position-absolute top-50 end-0 translate-middle-y me-1" [class.d-none]="slimSidebarEnabled || slimSidebarAnimating()" [compact]="true" title="Mail" i18n-title [ngModel]="!settingsService.sidebarItemIsHidden(HideableSidebarItemID.Mail)" (ngModelChange)="toggleSidebarItem(HideableSidebarItemID.Mail, $event)"></pngx-input-switch>
-              }
            </li>
            <li class="nav-item app-link" *pngxIfPermissions="{ action: PermissionAction.Delete, type: PermissionType.Document }">
              <a class="nav-link" routerLink="trash" routerLinkActive="active" (click)="closeMenu()" ngbPopover="Trash"
@@ -334,16 +322,13 @@
                </a>
              </li>
            }
-            <li class="nav-item mt-2 position-relative" [class.d-none]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.Documentation) && !settingsService.organizingSidebarItems()" tourAnchor="tour.outro">
-              <a class="text-muted small d-flex align-items-center flex-wrap text-decoration-none nav-anchor" [class.opacity-50]="settingsService.sidebarItemIsHidden(HideableSidebarItemID.Documentation)" [class.pe-5]="settingsService.organizingSidebarItems() && !slimSidebarEnabled && !slimSidebarAnimating()"
+            <li class="nav-item mt-2" tourAnchor="tour.outro">
+              <a class="text-muted small d-flex align-items-center flex-wrap text-decoration-none nav-anchor"
                target="_blank" rel="noopener noreferrer" href="https://docs.paperless-ngx.com" ngbPopover="Documentation"
                i18n-ngbPopover [disablePopover]="!slimSidebarPopoversEnabled" placement="end" container="body"
                triggers="mouseenter:mouseleave" popoverClass="popover-slim">
                <i-bs class="d-flex me-2" name="question-circle"></i-bs><span><ng-container i18n>Documentation</ng-container></span>
              </a>
-              @if (settingsService.organizingSidebarItems()) {
-                <pngx-input-switch class="position-absolute top-50 end-0 translate-middle-y me-1" [class.d-none]="slimSidebarEnabled || slimSidebarAnimating()" [compact]="true" title="Documentation" i18n-title [ngModel]="!settingsService.sidebarItemIsHidden(HideableSidebarItemID.Documentation)" (ngModelChange)="toggleSidebarItem(HideableSidebarItemID.Documentation, $event)"></pngx-input-switch>
-              }
            </li>
            <li class="nav-item" [class.visually-hidden]="slimSidebarEnabled">
              <div class="text-muted small d-flex align-items-center flex-wrap nav-label">
@@ -15,7 +15,7 @@ import { provideUiTour } from 'ngx-ui-tour-ng-bootstrap'
 import { of, throwError } from 'rxjs'
 import { routes } from 'src/app/app-routing.module'
 import { SavedView } from 'src/app/data/saved-view'
-import { HideableSidebarItemID, SETTINGS_KEYS } from 'src/app/data/ui-settings'
+import { SETTINGS_KEYS } from 'src/app/data/ui-settings'
 import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { PermissionsGuard } from 'src/app/guards/permissions.guard'
 import {
@@ -287,82 +287,6 @@ describe('AppFrameComponent', () => {
    jest.useRealTimers()
  })

-  it('should hide configured sidebar items', () => {
-    settingsService.set(SETTINGS_KEYS.SIDEBAR_HIDDEN_ITEMS, [
-      HideableSidebarItemID.Dashboard,
-      HideableSidebarItemID.Workflows,
-    ])
-    fixture.detectChanges()
-
-    expect(
-      fixture.nativeElement.querySelector('[routerLink="dashboard"]')
-        .parentElement.classList
-    ).toContain('d-none')
-    expect(
-      fixture.nativeElement.querySelector('[routerLink="workflows"]')
-        .parentElement.classList
-    ).toContain('d-none')
-    expect(
-      fixture.nativeElement.querySelector('[routerLink="mail"]').parentElement
-        .classList
-    ).not.toContain('d-none')
-  })
-
-  it('should show hidden items and visibility switches while customizing', () => {
-    settingsService.set(SETTINGS_KEYS.SIDEBAR_HIDDEN_ITEMS, [
-      HideableSidebarItemID.Dashboard,
-    ])
-    settingsService.sidebarHiddenItemsEditing.set([
-      HideableSidebarItemID.Dashboard,
-    ])
-    fixture.detectChanges()
-
-    expect(
-      fixture.nativeElement.querySelectorAll('pngx-input-switch').length
-    ).toBe(5)
-    expect(
-      fixture.nativeElement.querySelector('[routerLink="dashboard"]')
-        .parentElement.classList
-    ).not.toContain('d-none')
-    expect(
-      fixture.nativeElement.querySelector('[routerLink="dashboard"]').classList
-    ).toContain('opacity-50')
-
-    settingsService.set(SETTINGS_KEYS.SLIM_SIDEBAR, true)
-    fixture.detectChanges()
-
-    expect(
-      Array.from(
-        fixture.nativeElement.querySelectorAll('pngx-input-switch')
-      ).every((toggle: HTMLElement) => toggle.classList.contains('d-none'))
-    ).toBe(true)
-    expect(
-      fixture.nativeElement.querySelector('[routerLink="dashboard"]').classList
-    ).not.toContain('pe-5')
-
-    settingsService.set(SETTINGS_KEYS.SLIM_SIDEBAR, false)
-    component.slimSidebarAnimating.set(true)
-    fixture.detectChanges()
-
-    expect(
-      Array.from(
-        fixture.nativeElement.querySelectorAll('pngx-input-switch')
-      ).every((toggle: HTMLElement) => toggle.classList.contains('d-none'))
-    ).toBe(true)
-
-    component.slimSidebarAnimating.set(false)
-    fixture.detectChanges()
-
-    expect(
-      Array.from(
-        fixture.nativeElement.querySelectorAll('pngx-input-switch')
-      ).every((toggle: HTMLElement) => !toggle.classList.contains('d-none'))
-    ).toBe(true)
-    expect(
-      fixture.nativeElement.querySelector('[routerLink="dashboard"]').classList
-    ).toContain('pe-5')
-  })
-
  it('should show error on toggle slim sidebar if store settings fails', () => {
    jest.spyOn(console, 'warn').mockImplementation(() => {})
    const toastSpy = jest.spyOn(toastService, 'showError')
@@ -7,7 +7,6 @@ import {
 } from '@angular/cdk/drag-drop'
 import { NgClass } from '@angular/common'
 import { Component, HostListener, inject, OnInit, signal } from '@angular/core'
-import { FormsModule } from '@angular/forms'
 import { ActivatedRoute, Router, RouterModule } from '@angular/router'
 import {
  NgbCollapseModule,
@@ -22,11 +21,7 @@ import { Observable } from 'rxjs'
 import { first } from 'rxjs/operators'
 import { Document } from 'src/app/data/document'
 import { SavedView } from 'src/app/data/saved-view'
-import {
-  CollapsibleSection,
-  HideableSidebarItemID,
-  SETTINGS_KEYS,
-} from 'src/app/data/ui-settings'
+import { CollapsibleSection, SETTINGS_KEYS } from 'src/app/data/ui-settings'
 import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { ComponentCanDeactivate } from 'src/app/guards/dirty-doc.guard'
 import { DocumentTitlePipe } from 'src/app/pipes/document-title.pipe'
@@ -53,7 +48,6 @@ import { ChatComponent } from '../chat/chat/chat.component'
 import { BrandMarkComponent } from '../common/logo/brand-mark/brand-mark.component'
 import { LogoComponent } from '../common/logo/logo.component'
 import { ProfileEditDialogComponent } from '../common/profile-edit-dialog/profile-edit-dialog.component'
-import { SwitchComponent } from '../common/input/switch/switch.component'
 import { DocumentDetailComponent } from '../document-detail/document-detail.component'
 import { ComponentWithPermissions } from '../with-permissions/with-permissions.component'
 import { GlobalSearchComponent } from './global-search/global-search.component'
@@ -82,8 +76,6 @@ const SCROLL_THRESHOLD = 16
    NgxBootstrapIconsModule,
    DragDropModule,
    TourNgBootstrap,
-    FormsModule,
-    SwitchComponent,
  ],
 })
 export class AppFrameComponent
@@ -106,7 +98,6 @@ export class AppFrameComponent
  readonly isMenuCollapsed = signal(true)
  readonly slimSidebarAnimating = signal(false)
  readonly mobileSearchHidden = signal(false)
-  readonly HideableSidebarItemID = HideableSidebarItemID
  private readonly versionSetting = this.settingsService.getSignal<string>(
    SETTINGS_KEYS.VERSION
  )
@@ -204,10 +195,6 @@ export class AppFrameComponent
    }, 200) // slightly longer than css animation for slim sidebar
  }

-  toggleSidebarItem(item: HideableSidebarItemID, visible: boolean): void {
-    this.settingsService.updateSidebarItemVisibility(item, visible)
-  }
-
  toggleAttributesSections(event?: Event): void {
    event?.preventDefault()
    event?.stopPropagation()
@@ -1,6 +1,6 @@
-<div [class.mb-3]="!compact">
-  <div [class.row]="!compact">
-    @if (!horizontal && !compact) {
+<div class="mb-3">
+  <div class="row">
+    @if (!horizontal) {
      <div class="d-flex align-items-center position-relative hidden-button-container col-md-3">
        <label class="form-label" [for]="inputId" [ngbTooltip]="showUnsetNote && isUnset ? tipContent: null" placement="end">
          {{title}}
@@ -17,8 +17,8 @@
    }
    <div [ngClass]="{'align-items-center': horizontal, 'd-flex': horizontal}">
      <div class="form-check form-switch">
-        <input #inputField type="checkbox" class="form-check-input" [id]="inputId" [(ngModel)]="value" [ngModelOptions]="{standalone: true}" (change)="onChange(value)" (blur)="onTouched()" [disabled]="disabled" [attr.aria-label]="compact ? title : null">
-        @if (horizontal && !compact) {
+        <input #inputField type="checkbox" class="form-check-input" [id]="inputId" [(ngModel)]="value" [ngModelOptions]="{standalone: true}" (change)="onChange(value)" (blur)="onTouched()" [disabled]="disabled">
+        @if (horizontal) {
          <label class="form-check-label" [class.text-muted]="showUnsetNote && isUnset" [for]="inputId" [ngbTooltip]="showUnsetNote && isUnset ? tipContent: null" placement="end">
            {{title}}
            @if (showUnsetNote && isUnset) {
@@ -48,14 +48,4 @@ describe('SwitchComponent', () => {
    component.value = undefined
    expect(component.isUnset).toBeTruthy()
  })
-
-  it('should support a compact layout', () => {
-    component.compact = true
-    component.title = 'Test switch'
-    fixture.detectChanges()
-
-    expect(fixture.nativeElement.querySelector('.mb-3')).toBeNull()
-    expect(fixture.nativeElement.querySelector('.row')).toBeNull()
-    expect(input.getAttribute('aria-label')).toEqual('Test switch')
-  })
 })
@@ -25,9 +25,6 @@ export class SwitchComponent extends AbstractInputComponent<boolean> {
  @Input()
  showUnsetNote: boolean = false

-  @Input()
-  compact: boolean = false
-
  constructor() {
    super()
  }
@@ -24,16 +24,6 @@ export enum CollapsibleSection {
  ATTRIBUTES = 'attributes',
 }

-export enum HideableSidebarItemID {
-  Dashboard = 'dashboard',
-  SavedViews = 'saved_views',
-  Workflows = 'workflows',
-  Mail = 'mail',
-  Documentation = 'documentation',
-}
-
-export const HIDEABLE_SIDEBAR_ITEM_IDS = Object.values(HideableSidebarItemID)
-
 export const PAPERLESS_GREEN_HEX = '#17541f'

 export const SETTINGS_KEYS = {
@@ -66,7 +56,6 @@ export const SETTINGS_KEYS = {
  NOTES_ENABLED: 'general-settings:notes-enabled',
  AUDITLOG_ENABLED: 'general-settings:auditlog-enabled',
  SLIM_SIDEBAR: 'general-settings:slim-sidebar',
-  SIDEBAR_HIDDEN_ITEMS: 'general-settings:sidebar:hidden-items',
  ATTRIBUTES_SECTIONS_COLLAPSED:
    'general-settings:attributes-sections-collapsed',
  UPDATE_CHECKING_ENABLED: 'general-settings:update-checking:enabled',
@@ -138,11 +127,6 @@ export const SETTINGS: UiSetting[] = [
    type: 'boolean',
    default: false,
  },
-  {
-    key: SETTINGS_KEYS.SIDEBAR_HIDDEN_ITEMS,
-    type: 'array',
-    default: [],
-  },
  {
    key: SETTINGS_KEYS.ATTRIBUTES_SECTIONS_COLLAPSED,
    type: 'array',
@@ -14,11 +14,7 @@ import { CustomFieldDataType } from '../data/custom-field'
 import { DEFAULT_DISPLAY_FIELDS, DisplayField } from '../data/document'
 import { SavedView } from '../data/saved-view'
 import { RemoteOCRModeConfig } from '../data/paperless-config'
-import {
-  HideableSidebarItemID,
-  SETTINGS_KEYS,
-  UiSettings,
-} from '../data/ui-settings'
+import { SETTINGS_KEYS, UiSettings } from '../data/ui-settings'
 import { PermissionsService } from './permissions.service'
 import { CustomFieldsService } from './rest/custom-fields.service'
 import { SettingsService } from './settings.service'
@@ -234,35 +230,6 @@ describe('SettingsService', () => {
    expect(notesEnabled()).toBeFalsy()
  })

-  it('updates sidebar item visibility', () => {
-    httpTestingController
-      .expectOne(`${environment.apiBaseUrl}ui_settings/`)
-      .flush(ui_settings)
-
-    expect(
-      settingsService.sidebarItemIsHidden(HideableSidebarItemID.Workflows)
-    ).toBe(false)
-
-    settingsService.updateSidebarItemVisibility(
-      HideableSidebarItemID.Workflows,
-      false
-    )
-
-    expect(
-      settingsService.sidebarItemIsHidden(HideableSidebarItemID.Workflows)
-    ).toBe(true)
-    expect(settingsService.get(SETTINGS_KEYS.SIDEBAR_HIDDEN_ITEMS)).toEqual([])
-
-    settingsService.updateSidebarItemVisibility(
-      HideableSidebarItemID.Workflows,
-      true
-    )
-
-    expect(
-      settingsService.sidebarItemIsHidden(HideableSidebarItemID.Workflows)
-    ).toBe(false)
-  })
-
  it('updates setting signals when settings are reinitialized', () => {
    let req = httpTestingController.expectOne(
      `${environment.apiBaseUrl}ui_settings/`
@@ -24,7 +24,6 @@ import { DEFAULT_DISPLAY_FIELDS, DisplayField } from '../data/document'
 import { RemoteOCRModeConfig } from '../data/paperless-config'
 import { SavedView } from '../data/saved-view'
 import {
-  HideableSidebarItemID,
  PAPERLESS_GREEN_HEX,
  SETTINGS,
  SETTINGS_KEYS,
@@ -314,18 +313,6 @@ export class SettingsService {
  readonly globalDropzoneEnabled = signal(true)
  readonly globalDropzoneActive = signal(false)
  readonly organizingSidebarSavedViews = signal(false)
-  readonly sidebarHiddenItemsEditing = signal<HideableSidebarItemID[] | null>(
-    null
-  )
-  readonly organizingSidebarItems = computed(
-    () => this.sidebarHiddenItemsEditing() !== null
-  )
-  readonly sidebarHiddenItemsEditingChanged = new EventEmitter<
-    HideableSidebarItemID[]
-  >()
-  readonly hiddenSidebarItems = this.getSignal<HideableSidebarItemID[]>(
-    SETTINGS_KEYS.SIDEBAR_HIDDEN_ITEMS
-  )

  readonly allDisplayFields = signal<Array<{ id: DisplayField; name: string }>>(
    DEFAULT_DISPLAY_FIELDS
@@ -762,29 +749,6 @@ export class SettingsService {
    return this.storeSettings()
  }

-  sidebarItemIsHidden(item: HideableSidebarItemID): boolean {
-    return (
-      this.sidebarHiddenItemsEditing() ?? this.hiddenSidebarItems()
-    ).includes(item)
-  }
-
-  updateSidebarItemVisibility(
-    item: HideableSidebarItemID,
-    visible: boolean
-  ): void {
-    const hiddenItems = new Set(
-      this.sidebarHiddenItemsEditing() ?? this.hiddenSidebarItems()
-    )
-    if (visible) {
-      hiddenItems.delete(item)
-    } else {
-      hiddenItems.add(item)
-    }
-    const updatedHiddenItems = [...hiddenItems]
-    this.sidebarHiddenItemsEditing.set(updatedHiddenItems)
-    this.sidebarHiddenItemsEditingChanged.emit(updatedHiddenItems)
-  }
-
  updateSavedViewsVisibility(
    dashboardVisibleViewIds: number[],
    sidebarVisibleViewIds: number[]
@@ -2,6 +2,7 @@ from __future__ import annotations

 import logging
 import tempfile
+import uuid
 from pathlib import Path
 from typing import TYPE_CHECKING
 from typing import Literal
@@ -379,7 +380,7 @@ def delete(doc_ids: list[int]) -> Literal["OK"]:
        )
        delete_ids = list({*doc_ids, *version_ids})

-        Document.objects.filter(id__in=delete_ids).delete()
+        Document.objects.filter(id__in=delete_ids).delete(transaction_id=uuid.uuid4())

        from documents.search import get_backend

@@ -156,6 +156,15 @@ class FileStabilityTracker:
                    logger.debug(f"File disappeared during stability check: {path}")
                continue

+            # Stable, but empty: some scanners create a zero byte placeholder
+            # and only write the page some time later. Consuming it now can
+            # only fail  so drop it and let the writer's next event
+            # (or the periodic rescan) bring it back once it has content
+            if not tracked.last_size:
+                to_remove.append(path)
+                logger.debug("Ignoring stable but empty file: %s", path)
+                continue
+
            # File is stable, we can return it
            to_yield.append(path)
            logger.info(f"File is stable: {path}")
@@ -1,4 +1,5 @@
 import datetime
+import uuid
 from pathlib import Path
 from typing import Final

@@ -514,13 +515,20 @@ class Document(SoftDeleteModel, ModelWithOwner):  # type: ignore[django-manager-
    def delete(
        self,
        *args,
+        transaction_id=None,
        **kwargs,
    ):
-        # If deleting a root document, move all its versions to trash as well.
+        # Versions must share the root's transaction ID so they are restored
+        # together by django-softdelete.
+        if transaction_id is None:
+            transaction_id = uuid.uuid4()
        if self.root_document_id is None:
-            Document.objects.filter(root_document=self).delete()
+            Document.objects.filter(root_document=self).delete(
+                transaction_id=transaction_id,
+            )
        return super().delete(
            *args,
+            transaction_id=transaction_id,
            **kwargs,
        )

@@ -207,3 +207,65 @@ class TestTrashAPI(DirectoriesMixin, APITestCase):
        )
        self.assertEqual(resp.status_code, status.HTTP_400_BAD_REQUEST)
        self.assertIn("have not yet been deleted", resp.data["documents"][0])
+
+    def _make_versioned_document(self) -> tuple[Document, list[Document]]:
+        root = Document.objects.create(
+            title="root",
+            content="root-content",
+            checksum="root",
+            mime_type="application/pdf",
+        )
+        versions = [
+            Document.objects.create(
+                title=f"v{index}",
+                content=f"v{index}-content",
+                checksum=f"v{index}",
+                mime_type="application/pdf",
+                root_document=root,
+                version_index=index,
+            )
+            for index in range(1, 3)
+        ]
+        return root, versions
+
+    def test_api_trash_restore_document_restores_its_versions(self) -> None:
+        """
+        GIVEN:
+            - Existing document with two versions
+        WHEN:
+            - API request to delete the document
+            - API request to restore it from the trash
+        THEN:
+            - Only the document itself is listed in the trash
+            - A version cannot be restored without its root
+            - The document is restored together with all of its versions
+        """
+        root, versions = self._make_versioned_document()
+
+        self.client.force_login(user=self.user)
+        self.client.delete(f"/api/documents/{root.pk}/")
+        self.assertEqual(Document.deleted_objects.count(), 3)
+
+        resp = self.client.get("/api/trash/")
+        self.assertEqual(resp.status_code, status.HTTP_200_OK)
+        self.assertEqual(resp.data["count"], 1)
+        self.assertEqual(resp.data["results"][0]["id"], root.pk)
+
+        # A version cannot be restored while its root remains in the trash.
+        resp = self.client.post(
+            "/api/trash/",
+            {"action": "restore", "documents": [versions[0].pk]},
+        )
+        self.assertEqual(resp.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertIn("Restore the root document", resp.data["documents"][0])
+
+        resp = self.client.post(
+            "/api/trash/",
+            {"action": "restore", "documents": [root.pk]},
+        )
+        self.assertEqual(resp.status_code, status.HTTP_200_OK)
+        self.assertEqual(Document.deleted_objects.count(), 0)
+        self.assertCountEqual(
+            Document.objects.filter(root_document=root).values_list("id", flat=True),
+            [version.pk for version in versions],
+        )
@@ -392,6 +392,11 @@ class TestBulkEdit(DirectoriesMixin, TestCase):
        self.assertFalse(Document.objects.filter(id=self.doc1.id).exists())
        self.assertFalse(Document.objects.filter(id=version.id).exists())

+        Document.deleted_objects.get(id=self.doc1.id).restore(strict=False)
+
+        self.assertTrue(Document.objects.filter(id=self.doc1.id).exists())
+        self.assertTrue(Document.objects.filter(id=version.id).exists())
+
    def test_delete_version_document_keeps_root(self) -> None:
        version = Document.objects.create(
            checksum="A-v1",
@@ -110,7 +110,7 @@ class TestDocument(TestCase):
            checksum="checksum",
            mime_type="application/pdf",
        )
-        Document.objects.create(
+        version = Document.objects.create(
            root_document=root,
            correspondent=root.correspondent,
            title="Version",
@@ -124,6 +124,10 @@ class TestDocument(TestCase):
        self.assertEqual(Document.objects.count(), 0)
        self.assertEqual(Document.deleted_objects.count(), 2)

+        root.restore(strict=False)
+
+        self.assertTrue(Document.objects.filter(pk=version.pk).exists())
+
    def test_file_name(self) -> None:
        doc = Document(
            mime_type="application/pdf",
@@ -136,6 +136,23 @@ def wait_for_mock_call(
    return False


+def sleep_past_stability(
+    owner: FileStabilityTracker | ConsumerThread,
+    *,
+    windows: float = 1.5,
+) -> None:
+    """
+    Block until a tracked file's stability window has certainly elapsed.
+
+    Args:
+        owner: The tracker, or the consumer thread running one, whose
+               configured stability delay sets the wait.
+        windows: How many stability windows to wait, giving slop for a slow
+                 or loaded test runner.
+    """
+    sleep(owner.stability_delay * windows)
+
+
 class TestTrackedFile:
    """Tests for the TrackedFile dataclass."""

@@ -261,6 +278,56 @@ class TestFileStabilityTracker:
        assert len(stable) == 0
        assert stability_tracker.pending_count == 1

+    def test_get_stable_files_skips_empty_file(
+        self,
+        stability_tracker: FileStabilityTracker,
+        tmp_path: Path,
+    ) -> None:
+        """
+        GIVEN:
+            - A zero byte file, tracked and past its stability delay
+        WHEN:
+            - Stable files are collected
+        THEN:
+            - The file is not yielded for consumption
+            - The file is dropped from tracking rather than held, so an
+              abandoned placeholder does not keep the watch loop awake
+        """
+        empty = tmp_path / "scan.pdf"
+        empty.write_bytes(b"")
+        stability_tracker.track(empty, Change.added)
+        sleep_past_stability(stability_tracker)
+
+        stable = list(stability_tracker.get_stable_files())
+
+        assert stable == []
+        assert stability_tracker.pending_count == 0
+
+    def test_empty_file_is_yielded_once_content_arrives(
+        self,
+        stability_tracker: FileStabilityTracker,
+        tmp_path: Path,
+    ) -> None:
+        """
+        GIVEN:
+            - A zero byte file which was dropped from tracking while empty
+        WHEN:
+            - The writer fills the file and a new event re-tracks it
+        THEN:
+            - The file is yielded for consumption once it is stable
+        """
+        target = tmp_path / "scan.pdf"
+        target.write_bytes(b"")
+        stability_tracker.track(target, Change.added)
+        sleep_past_stability(stability_tracker)
+        assert list(stability_tracker.get_stable_files()) == []
+
+        target.write_bytes(b"%PDF-1.4 content")
+        stability_tracker.track(target, Change.modified)
+        sleep_past_stability(stability_tracker)
+
+        assert list(stability_tracker.get_stable_files()) == [target]
+
    def test_get_stable_files_deleted_during_check(self, temp_file: Path) -> None:
        """Test deleted file is not returned during stability check."""
        tracker = FileStabilityTracker(stability_delay=0.1)
@@ -879,6 +946,51 @@ class TestCommandWatch:

        mock_consume_file_delay.apply_async.assert_called()

+    def test_scanner_placeholder_is_not_consumed_while_empty(
+        self,
+        consumption_dir: Path,
+        sample_pdf: Path,
+        mock_consume_file_delay: MagicMock,
+        start_consumer: Callable[..., ConsumerThread],
+    ) -> None:
+        """
+        GIVEN:
+            - A scanner which creates a zero byte placeholder and only writes
+              the page some time later (GH discussion #13969)
+        WHEN:
+            - The placeholder sits untouched well past the stability delay
+            - The scanner then writes the real content
+        THEN:
+            - The empty placeholder is never queued, as it could only fail
+              with "Unsupported mime type inode/x-empty"
+            - The file is queued exactly once, when the content lands
+        """
+        thread = start_consumer(stability_delay=0.2)
+
+        target = consumption_dir / "scan.pdf"
+        target.write_bytes(b"")  # the scanner's placeholder
+
+        # Well past the stability delay: the old behaviour queued it here.
+        sleep_past_stability(thread, windows=5)
+        if thread.exception:
+            raise thread.exception
+        assert mock_consume_file_delay.apply_async.call_count == 0
+
+        shutil.copy(sample_pdf, target)  # the scanner finishes the page
+
+        assert wait_for_mock_call(
+            mock_consume_file_delay.apply_async,
+            timeout_s=5.0,
+        )
+        if thread.exception:
+            raise thread.exception
+
+        assert mock_consume_file_delay.apply_async.call_count == 1
+        queued_doc = mock_consume_file_delay.apply_async.call_args.kwargs["kwargs"][
+            "input_doc"
+        ]
+        assert queued_doc.original_file.name == "scan.pdf"
+
    def test_ignores_macos_files(
        self,
        consumption_dir: Path,
@@ -32,6 +32,7 @@ from documents.signals.handlers import update_llm_suggestions_cache
 from documents.tests.utils import DirectoriesMixin
 from documents.tests.utils import read_streaming_response
 from paperless.models import ApplicationConfiguration
+from paperless_ai.exceptions import LLMProviderError
 from paperless_ai.exceptions import LLMTimeoutError


@@ -737,6 +738,38 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
            get_llm_suggestion_cache(self.document.pk, backend="openai-like"),
        )

+    @patch("documents.views.get_ai_document_classification")
+    @override_settings(
+        AI_ENABLED=True,
+        LLM_BACKEND="openai-like",
+    )
+    def test_ai_suggestions_with_llm_provider_error(
+        self,
+        mock_get_ai_classification,
+    ) -> None:
+        mock_get_ai_classification.side_effect = LLMProviderError(
+            "confidential provider response",
+        )
+
+        self.client.force_login(user=self.user)
+        response = self.client.get(
+            f"/api/documents/{self.document.pk}/ai_suggestions/",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_502_BAD_GATEWAY)
+        self.assertEqual(
+            response.json(),
+            {
+                "ai": [
+                    "AI backend rejected the request. Check logs for details.",
+                ],
+            },
+        )
+        self.assertNotIn("confidential provider response", response.content.decode())
+        self.assertIsNone(
+            get_llm_suggestion_cache(self.document.pk, backend="openai-like"),
+        )
+
    @patch("documents.views.get_ai_document_classification")
    @override_settings(
        AI_ENABLED=True,
@@ -252,6 +252,7 @@ from paperless.views import StandardPagination
 from paperless_ai.ai_classifier import get_ai_document_classification
 from paperless_ai.ai_classifier import get_llm_output_language
 from paperless_ai.chat import stream_chat_with_documents
+from paperless_ai.exceptions import LLMProviderError
 from paperless_ai.exceptions import LLMTimeoutError
 from paperless_ai.matching import extract_unmatched_names
 from paperless_ai.matching import match_correspondents_by_name
@@ -1603,6 +1604,22 @@ class DocumentViewSet(
                    {"ai": [_("AI backend request timed out.")]},
                    status=status.HTTP_503_SERVICE_UNAVAILABLE,
                )
+            except LLMProviderError:
+                logger.exception(
+                    "AI backend rejected the request for document %s",
+                    doc.pk,
+                )
+                return Response(
+                    {
+                        "ai": [
+                            _(
+                                "AI backend rejected the request. "
+                                "Check logs for details.",
+                            ),
+                        ],
+                    },
+                    status=status.HTTP_502_BAD_GATEWAY,
+                )
            set_llm_suggestions_cache(
                doc.pk,
                llm_suggestions,
@@ -5437,7 +5454,10 @@ class TrashView(ListModelMixin, PassUserMixin):

    model = Document

-    queryset = Document.deleted_objects.all()
+    # A version is listed separately only when its root is not in the trash.
+    queryset = Document.deleted_objects.exclude(
+        root_document_id__in=Document.deleted_objects.values("id"),
+    )

    def get(self, request: Request, format: str | None = None) -> Response:
        self.serializer_class = DocumentSerializer
@@ -5468,7 +5488,15 @@ class TrashView(ListModelMixin, PassUserMixin):
            return HttpResponseForbidden("Insufficient permissions")
        action = serializer.validated_data.get("action")
        if action == "restore":
-            restored = list(Document.deleted_objects.filter(id__in=doc_ids))
+            restored = list(self.get_queryset().filter(id__in=doc_ids))
+            if len(restored) != len(doc_ids):
+                raise ValidationError(
+                    {
+                        "documents": [
+                            "Restore the root document instead of one of its versions.",
+                        ],
+                    },
+                )
            for doc in restored:
                doc.restore(strict=False)
            if restored:
@@ -4,21 +4,24 @@ from django.conf import settings
 from django.contrib.auth.models import User

 from documents.models import Document
-from documents.permissions import get_objects_for_user_owner_aware
+from documents.permissions import permitted_object_ids
+from documents.permissions import restrict_queryset_to_visible
+from documents.permissions import user_is_unrestricted
 from paperless.config import AIConfig
 from paperless_ai.base_model import ClassificationSuggestions
 from paperless_ai.base_model import TaxonomyChoiceDict
 from paperless_ai.base_model import classification_suggestions_to_model
 from paperless_ai.client import AIClient
 from paperless_ai.db import db_connection_released
-from paperless_ai.indexing import _node_document_ids
 from paperless_ai.indexing import retrieve_similar_nodes
 from paperless_ai.indexing import truncate_content
 from paperless_ai.prompts.context import ClassificationPromptContext
 from paperless_ai.prompts.context import LocalizationPromptContext
 from paperless_ai.prompts.context import RagContextPromptContext
 from paperless_ai.prompts.render import render_prompt
+from paperless_ai.taxonomy import SimilarDocument
 from paperless_ai.taxonomy import TaxonomyCandidates
+from paperless_ai.taxonomy import _node_document_weights
 from paperless_ai.taxonomy import build_taxonomy_candidates
 from paperless_ai.taxonomy import empty_taxonomy_candidates
 from paperless_ai.taxonomy import format_taxonomy_for_prompt
@@ -37,6 +40,48 @@ logger = logging.getLogger("paperless_ai.rag_classifier")
 TAXONOMY_CANDIDATE_TOP_K = 15


+def _fulltext_similar_documents(
+    document: Document,
+    user: User | None,
+    top_k: int,
+) -> list[SimilarDocument]:
+    """Rank-based fallback when no embedding backend is configured. Uses
+    Tantivy's "More Like This" (term-overlap similarity) instead of vector
+    similarity - cruder, but far better than no candidates at all.
+    more_like_this_ids returns only a ranked ID list, no scores, so weight is
+    synthesized from rank (descending from top_k) rather than claiming a
+    similarity magnitude that doesn't exist. An unrestricted user (none, or an
+    active superuser - see user_is_unrestricted) is normalized to ``None``
+    before calling, since the backend's permission filter has no superuser
+    short-circuit of its own. Results are re-checked with
+    restrict_queryset_to_visible() since Tantivy's indexed permission fields
+    lag the DB via async reindexing.
+    """
+    from documents.search import get_backend
+
+    unrestricted = user_is_unrestricted(user)
+    search_user = None if unrestricted else user
+    backend = get_backend()
+    similar_ids = backend.more_like_this_ids(
+        document.pk,
+        user=search_user,
+        limit=top_k,
+    )
+    if not unrestricted:
+        allowed_ids = set(
+            restrict_queryset_to_visible(
+                Document.objects.filter(pk__in=similar_ids),
+                user,
+                "view_document",
+            ).values_list("pk", flat=True),
+        )
+        similar_ids = [doc_id for doc_id in similar_ids if doc_id in allowed_ids]
+    return [
+        SimilarDocument(document_id=doc_id, weight=float(top_k - rank))
+        for rank, doc_id in enumerate(similar_ids)
+    ]
+
+
 def get_language_name(language_code: str) -> str:
    normalized_language_code = language_code.lower()
    for code, name in settings.LANGUAGES:
@@ -136,43 +181,52 @@ def get_taxonomy_context(
    user: User | None = None,
    max_docs: int = 5,
 ) -> tuple[TaxonomyCandidates, str]:
-    """One retrieval feeds both taxonomy candidates and RAG text context.
-    On any retrieval failure, degrades to empty candidates/context rather than
-    propagating the exception - a vector-store outage should not block
-    classification, only its RAG-assisted enrichment.
+    """One retrieval feeds both taxonomy candidates and RAG text context. Uses
+    vector similarity when an embedding backend is configured, otherwise
+    falls back to Tantivy full-text "More Like This" similarity - see
+    _fulltext_similar_documents. On any retrieval failure, degrades to empty
+    candidates/context rather than propagating the exception - neither a
+    vector-store outage nor a search-index issue should block classification,
+    only its context-assisted enrichment.
    """
+    ai_config = AIConfig()
    try:
-        # None means "no restriction" to retrieve_similar_nodes. A superuser
-        # (like no user at all) can see every document, so skip materializing
-        # every visible pk into a Python list and passing it through as an IN
-        # filter: for a large library that is a wasted quadratic scan in the
-        # vector store at best, and past ~32,763 documents a hard
-        # sqlite3.OperationalError (SQLite's bound-parameter limit) at worst.
-        # get_objects_for_user_owner_aware() would return every Document for a
-        # superuser anyway (guardian's own with_superuser shortcut), so this
-        # changes nothing about which documents are considered -- only how we
-        # get there.
-        visible_document_ids = (
-            None
-            if user is None or user.is_superuser
-            else list(
-                get_objects_for_user_owner_aware(
-                    user,
-                    "view_document",
-                    Document,
-                ).values_list("pk", flat=True),
+        if ai_config.llm_embedding_backend:
+            # None means "no restriction" to retrieve_similar_nodes. An
+            # unrestricted user (no user at all, or an active superuser -- see
+            # user_is_unrestricted) can see every document, so skip
+            # materializing every visible pk into a Python list and passing it
+            # through as an IN filter: for a large library that is a wasted
+            # quadratic scan in the vector store at best, and past ~32,763
+            # documents a hard sqlite3.OperationalError (SQLite's
+            # bound-parameter limit) at worst.
+            # permitted_object_ids() has its own superuser shortcut that would
+            # return every Document's id anyway, so this changes nothing about
+            # which documents are considered -- only how we get there.
+            visible_document_ids = (
+                None
+                if user_is_unrestricted(user)
+                else list(permitted_object_ids(user, Document, "view_document"))
+            )
+            nodes = retrieve_similar_nodes(
+                document,
+                top_k=TAXONOMY_CANDIDATE_TOP_K,
+                document_ids=visible_document_ids,
+            )
+            similar_documents = _node_document_weights(nodes)
+        else:
+            # See _fulltext_similar_documents: it applies its own permission
+            # filter via `user`, so no visible-document-id list is needed here.
+            similar_documents = _fulltext_similar_documents(
+                document,
+                user,
+                top_k=TAXONOMY_CANDIDATE_TOP_K,
            )
-        )
-        nodes = retrieve_similar_nodes(
-            document,
-            top_k=TAXONOMY_CANDIDATE_TOP_K,
-            document_ids=visible_document_ids,
-        )

-        candidates = build_taxonomy_candidates(nodes, user)
+        candidates = build_taxonomy_candidates(similar_documents, user)

-        # ``nodes`` are already ordered by descending vector similarity; don't lose it.
-        similar_document_ids = list(dict.fromkeys(_node_document_ids(nodes)))
+        # similar_documents is already ordered by descending weight; don't lose it.
+        similar_document_ids = [s["document_id"] for s in similar_documents]
        similar_documents_by_id = Document.objects.in_bulk(similar_document_ids)
        similar_docs = [
            similar_documents_by_id[document_id]
@@ -186,8 +240,8 @@ def get_taxonomy_context(
            context_blocks.append(f"TITLE: {title}\n{text}")
    except Exception:
        logger.exception(
-            "Failed to retrieve RAG neighbours for document %s; continuing "
-            "without taxonomy candidates or similar-document context.",
+            "Failed to retrieve similar-document context for document %s; "
+            "continuing without taxonomy candidates or similar-document context.",
            document.pk,
        )
        return empty_taxonomy_candidates(), ""
@@ -241,17 +295,13 @@ def get_ai_document_classification(
 ) -> ClassificationSuggestions:
    ai_config = AIConfig()

-    if ai_config.llm_embedding_backend:
-        candidates, context = get_taxonomy_context(document, user)
-        prompt = build_prompt_with_rag(
-            document,
-            ai_config,
-            candidates=candidates,
-            context=context,
-        )
-    else:
-        candidates = empty_taxonomy_candidates()
-        prompt = build_prompt_without_rag(document, ai_config, candidates=candidates)
+    candidates, context = get_taxonomy_context(document, user)
+    prompt = build_prompt_with_rag(
+        document,
+        ai_config,
+        candidates=candidates,
+        context=context,
+    )

    client = AIClient()
    # Hand the pooled DB connection back while the (slow) LLM query runs so it
@@ -22,6 +22,7 @@ from paperless.network import validate_outbound_http_url
 from paperless_ai.base_model import ClassificationSuggestions
 from paperless_ai.base_model import DocumentClassifierSchema
 from paperless_ai.base_model import model_to_classification_suggestions
+from paperless_ai.exceptions import LLMProviderError
 from paperless_ai.exceptions import LLMTimeoutError

 logger = logging.getLogger("paperless_ai.client")
@@ -132,7 +133,7 @@ class AIClient:
        from llama_index.core.llms import ChatMessage

        if self.settings.llm_backend == LLMBackend.OLLAMA:
-            with self._normalize_timeouts():
+            with self._normalize_errors():
                result = self.llm.chat(
                    [ChatMessage(role="user", content=prompt)],
                    format=DocumentClassifierSchema.model_json_schema(),
@@ -153,7 +154,7 @@ class AIClient:
            content=f"{prompt}\n\n"
            f"Answer by calling the {tool.metadata.name} tool. Do not write the answer as text.",
        )
-        with self._normalize_timeouts():
+        with self._normalize_errors():
            result = self.llm.chat_with_tools(
                tools=[tool],
                user_msg=user_msg,
@@ -173,7 +174,7 @@ class AIClient:
        )

    @contextmanager
-    def _normalize_timeouts(self) -> Iterator[None]:
+    def _normalize_errors(self) -> Iterator[None]:
        try:
            yield
        except httpx.TimeoutException as exc:
@@ -181,8 +182,23 @@ class AIClient:
        except Exception as exc:
            if self._is_openai_timeout(exc):
                raise LLMTimeoutError from exc
+            if self._is_provider_error(exc):
+                raise LLMProviderError from exc
            raise

+    def _is_provider_error(self, exc: Exception) -> bool:
+        if self.settings.llm_backend == LLMBackend.OLLAMA:
+            from ollama import ResponseError
+
+            return isinstance(exc, ResponseError)
+
+        if self.settings.llm_backend == LLMBackend.OPENAI_LIKE:
+            from openai import APIStatusError
+
+            return isinstance(exc, APIStatusError)
+
+        return False
+
    def _is_openai_timeout(self, exc: Exception) -> bool:
        if self.settings.llm_backend != LLMBackend.OPENAI_LIKE:
            return False
@@ -1,2 +1,6 @@
 class LLMTimeoutError(Exception):
    pass
+
+
+class LLMProviderError(Exception):
+    """The LLM backend rejected the request."""
@@ -721,20 +721,3 @@ def retrieve_similar_nodes(
            continue
        filtered.append(node)
    return filtered
-
-
-def _node_document_ids(nodes: list["NodeWithScore"]) -> list[int]:
-    document_ids: list[int] = []
-    for node in nodes:
-        document_id = node.metadata.get("document_id")
-        if document_id is None:  # pragma: no cover
-            # See the matching guard in retrieve_similar_nodes() above.
-            continue
-        try:
-            document_ids.append(int(document_id))
-        except ValueError:  # pragma: no cover
-            logger.warning(
-                "Skipping LLM index result with invalid document_id %r.",
-                document_id,
-            )
-    return document_ids
@@ -31,6 +31,11 @@ class TaxonomyCandidate(TypedDict):
    weight: float


+class SimilarDocument(TypedDict):
+    document_id: int
+    weight: float
+
+
 class TaxonomyCandidates(TypedDict):
    tags: list[TaxonomyCandidate]
    document_types: list[TaxonomyCandidate]
@@ -49,10 +54,10 @@ def empty_taxonomy_candidates() -> TaxonomyCandidates:
    )


-def _node_document_weights(nodes: list["NodeWithScore"]) -> dict[int, float]:
-    """document_id -> that node's similarity score, summed if a document_id
-    appears more than once across the retrieved nodes (e.g. multiple chunks
-    of the same source document)."""
+def _node_document_weights(nodes: list["NodeWithScore"]) -> list[SimilarDocument]:
+    """Sum each node's similarity score into its document_id (a document can
+    appear via multiple chunks/nodes) and return one SimilarDocument per
+    distinct document_id."""
    weights: dict[int, float] = defaultdict(float)
    for node in nodes:
        document_id = node.metadata.get("document_id")
@@ -65,7 +70,14 @@ def _node_document_weights(nodes: list["NodeWithScore"]) -> dict[int, float]:
            weights[int(document_id)] += float(node.score or 0.0)
        except (TypeError, ValueError):  # pragma: no cover
            continue
-    return weights
+    return sorted(
+        (
+            SimilarDocument(document_id=document_id, weight=weight)
+            for document_id, weight in weights.items()
+        ),
+        key=lambda similar: similar["weight"],
+        reverse=True,
+    )


 def _visible_ranked_candidates(
@@ -101,21 +113,26 @@ def _visible_ranked_candidates(


 def build_taxonomy_candidates(
-    nodes: list["NodeWithScore"],
+    similar_documents: list[SimilarDocument],
    user: User | None,
 ) -> TaxonomyCandidates:
-    """Resolve each neighbour node's document_id to a live Document, read its
-    *current* tags/type/correspondent/storage_path via the ORM (never the
-    possibly-stale names cached in vector-index node metadata), weight each
-    distinct taxonomy object by aggregate neighbour similarity, permission-filter
+    """Resolve each similar document's id to a live Document, read its
+    *current* tags/type/correspondent/storage_path via the ORM (never any
+    possibly-stale names an adapter's source might have cached), weight each
+    distinct taxonomy object by aggregate similarity weight, permission-filter
    against what ``user`` can see, and return each category ranked by weight
-    and capped.
+    and capped. ``similar_documents`` may come from either the vector-RAG
+    adapter or the full-text fallback adapter - both produce this same shape.
    """
-
-    document_weights = _node_document_weights(nodes)
-    if not document_weights:
+    if not similar_documents:
        return empty_taxonomy_candidates()

+    # Both adapters guarantee at most one SimilarDocument per document_id, so
+    # this never silently drops a duplicate's weight.
+    document_weights: dict[int, float] = {
+        s["document_id"]: s["weight"] for s in similar_documents
+    }
+
    # Only .tags.all() needs prefetching (a reverse M2M, one extra query for
    # the whole batch). document_type/correspondent/storage_path are read
    # below via their *_id columns (neighbour.document_type_id, etc.), which
@@ -1,4 +1,5 @@
 import datetime
+from collections.abc import Generator
 from types import SimpleNamespace
 from unittest.mock import MagicMock
 from unittest.mock import patch
@@ -6,18 +7,24 @@ from unittest.mock import patch
 import pytest
 import pytest_mock
 from django.test import override_settings
+from guardian.shortcuts import assign_perm
+from guardian.shortcuts import remove_perm

 from documents.models import Document
+from documents.search import TantivyBackend
 from documents.tests.factories import DocumentFactory
 from documents.tests.factories import TagFactory
 from documents.tests.factories import UserFactory
 from paperless.config import AIConfig
+from paperless_ai.ai_classifier import TAXONOMY_CANDIDATE_TOP_K
+from paperless_ai.ai_classifier import _fulltext_similar_documents
 from paperless_ai.ai_classifier import build_localization_prompt
 from paperless_ai.ai_classifier import build_prompt_with_rag
 from paperless_ai.ai_classifier import build_prompt_without_rag
 from paperless_ai.ai_classifier import get_ai_document_classification
 from paperless_ai.ai_classifier import get_language_name
 from paperless_ai.ai_classifier import get_taxonomy_context
+from paperless_ai.taxonomy import SimilarDocument
 from paperless_ai.taxonomy import TaxonomyCandidate
 from paperless_ai.taxonomy import TaxonomyCandidates

@@ -220,12 +227,10 @@ def test_use_rag_if_configured(

@pytest.mark.django_db
@patch("paperless_ai.client.AIClient.run_llm_query")
-@patch("paperless_ai.ai_classifier.build_prompt_without_rag")
-@patch("paperless_ai.ai_classifier.AIConfig")
+@patch("paperless_ai.ai_classifier.build_prompt_with_rag")
@override_settings(LLM_BACKEND="ollama", LLM_MODEL="some_model")
-def test_use_without_rag_if_not_configured(
-    mock_ai_config,
-    mock_build_prompt_without_rag,
+def test_use_rag_prompt_even_without_embedding_backend(
+    mock_build_prompt_with_rag,
    mock_run_llm_query,
    mock_document,
 ):
@@ -235,13 +240,13 @@ def test_use_without_rag_if_not_configured(
    WHEN:
        - get_ai_document_classification() is called
    THEN:
-        - The non-RAG prompt builder is used
+        - The RAG-context prompt builder is still used (fed by the full-text
+          fallback's context/candidates instead of the vector store's)
    """
-    mock_ai_config.return_value.llm_embedding_backend = None
-    mock_build_prompt_without_rag.return_value = "Prompt without RAG"
+    mock_build_prompt_with_rag.return_value = "Prompt with RAG"
    mock_run_llm_query.return_value = NESTED_SUGGESTIONS
    get_ai_document_classification(mock_document)
-    mock_build_prompt_without_rag.assert_called_once()
+    mock_build_prompt_with_rag.assert_called_once()


@pytest.mark.django_db
@@ -320,6 +325,7 @@ def test_build_localization_prompt_preserves_unicode_characters():


@pytest.mark.django_db
+@override_settings(LLM_EMBEDDING_BACKEND="huggingface")
 def test_get_taxonomy_context_assembles_rag_text_and_candidates():
    """
    GIVEN:
@@ -354,6 +360,7 @@ def test_get_taxonomy_context_assembles_rag_text_and_candidates():


@pytest.mark.django_db
+@override_settings(LLM_EMBEDDING_BACKEND="huggingface")
 def test_get_taxonomy_context_preserves_similarity_order_and_distinct_documents():
    """
    GIVEN:
@@ -424,6 +431,7 @@ def test_get_taxonomy_context_preserves_similarity_order_and_distinct_documents(


@pytest.mark.django_db
+@override_settings(LLM_EMBEDDING_BACKEND="huggingface")
 def test_get_taxonomy_context_no_similar_docs():
    """
    GIVEN:
@@ -447,6 +455,67 @@ def test_get_taxonomy_context_no_similar_docs():
    }


+@pytest.mark.django_db
+def test_get_taxonomy_context_uses_fulltext_fallback_when_no_embedding_backend(
+    mocker: pytest_mock.MockerFixture,
+) -> None:
+    """
+    GIVEN:
+        - No LLM embedding backend is configured (the default test settings)
+    WHEN:
+        - get_taxonomy_context() is called
+    THEN:
+        - _fulltext_similar_documents() is called with the document, the user
+          and TAXONOMY_CANDIDATE_TOP_K
+        - retrieve_similar_nodes() (the vector path) is never called
+    """
+    document = DocumentFactory.create(content="Some content")
+    mock_fulltext = mocker.patch(
+        "paperless_ai.ai_classifier._fulltext_similar_documents",
+        return_value=[],
+    )
+    mock_retrieve = mocker.patch("paperless_ai.ai_classifier.retrieve_similar_nodes")
+
+    get_taxonomy_context(document, user=None)
+
+    mock_fulltext.assert_called_once_with(
+        document,
+        None,
+        top_k=TAXONOMY_CANDIDATE_TOP_K,
+    )
+    mock_retrieve.assert_not_called()
+
+
+@pytest.mark.django_db
+@override_settings(LLM_EMBEDDING_BACKEND="huggingface")
+def test_get_taxonomy_context_uses_vector_path_when_embedding_backend_configured(
+    mocker: pytest_mock.MockerFixture,
+) -> None:
+    """
+    GIVEN:
+        - An LLM embedding backend is configured
+    WHEN:
+        - get_taxonomy_context() is called
+    THEN:
+        - retrieve_similar_nodes() (the vector path) is called
+        - _fulltext_similar_documents() (the no-embedding-backend fallback)
+          is never called
+    """
+    document = DocumentFactory.create(content="Some content")
+    mock_retrieve = mocker.patch(
+        "paperless_ai.ai_classifier.retrieve_similar_nodes",
+        return_value=[],
+    )
+    mock_fulltext = mocker.patch(
+        "paperless_ai.ai_classifier._fulltext_similar_documents",
+    )
+
+    get_taxonomy_context(document, user=None)
+
+    mock_retrieve.assert_called_once()
+    mock_fulltext.assert_not_called()
+
+
 class TestGetTaxonomyContextVisibility:
    """get_taxonomy_context must not materialize every visible document id
    for a user who can already see the whole library: a superuser (like no
@@ -459,6 +528,7 @@ class TestGetTaxonomyContextVisibility:
    """

    @pytest.mark.django_db
+    @override_settings(LLM_EMBEDDING_BACKEND="huggingface")
    def test_skips_permission_lookup_for_superuser(
        self,
        mocker: pytest_mock.MockerFixture,
@@ -477,17 +547,18 @@ class TestGetTaxonomyContextVisibility:
            "paperless_ai.ai_classifier.retrieve_similar_nodes",
            return_value=[],
        )
-        mock_get_objects = mocker.patch(
-            "paperless_ai.ai_classifier.get_objects_for_user_owner_aware",
+        mock_permitted = mocker.patch(
+            "paperless_ai.ai_classifier.permitted_object_ids",
        )
        user = UserFactory.create(is_superuser=True)

        get_taxonomy_context(document, user)

-        mock_get_objects.assert_not_called()
+        mock_permitted.assert_not_called()
        assert mock_retrieve.call_args.kwargs["document_ids"] is None

    @pytest.mark.django_db
+    @override_settings(LLM_EMBEDDING_BACKEND="huggingface")
    def test_skips_permission_lookup_when_no_user(
        self,
        mocker: pytest_mock.MockerFixture,
@@ -506,16 +577,17 @@ class TestGetTaxonomyContextVisibility:
            "paperless_ai.ai_classifier.retrieve_similar_nodes",
            return_value=[],
        )
-        mock_get_objects = mocker.patch(
-            "paperless_ai.ai_classifier.get_objects_for_user_owner_aware",
+        mock_permitted = mocker.patch(
+            "paperless_ai.ai_classifier.permitted_object_ids",
        )

        get_taxonomy_context(document, None)

-        mock_get_objects.assert_not_called()
+        mock_permitted.assert_not_called()
        assert mock_retrieve.call_args.kwargs["document_ids"] is None

    @pytest.mark.django_db
+    @override_settings(LLM_EMBEDDING_BACKEND="huggingface")
    def test_restricts_to_visible_documents_for_non_superuser(
        self,
        mocker: pytest_mock.MockerFixture,
@@ -526,7 +598,7 @@ class TestGetTaxonomyContextVisibility:
        WHEN:
            - get_taxonomy_context() is called
        THEN:
-            - The user's visible document ids are looked up and passed to
+            - The user's permitted document ids are looked up and passed to
              retrieve_similar_nodes() as a restriction
        """
        document = DocumentFactory.create(content="Some content")
@@ -534,21 +606,232 @@ class TestGetTaxonomyContextVisibility:
            "paperless_ai.ai_classifier.retrieve_similar_nodes",
            return_value=[],
        )
-        mock_queryset = mocker.MagicMock()
-        mock_queryset.values_list.return_value = [1, 2, 3]
-        mock_get_objects = mocker.patch(
-            "paperless_ai.ai_classifier.get_objects_for_user_owner_aware",
-            return_value=mock_queryset,
+        mock_permitted = mocker.patch(
+            "paperless_ai.ai_classifier.permitted_object_ids",
+            return_value=[1, 2, 3],
        )
        user = UserFactory.create(is_superuser=False)

        get_taxonomy_context(document, user)

-        mock_get_objects.assert_called_once_with(user, "view_document", Document)
+        mock_permitted.assert_called_once_with(user, Document, "view_document")
        assert mock_retrieve.call_args.kwargs["document_ids"] == [1, 2, 3]


@pytest.mark.django_db
+class TestFulltextSimilarDocuments:
+    """_fulltext_similar_documents is the no-embedding-backend fallback: it
+    asks the Tantivy full-text index for "More Like This" neighbours instead
+    of the vector store, and synthesizes a rank-based weight since Tantivy's
+    more_like_this_ids returns only an ordered id list, no scores.
+    """
+
+    @pytest.fixture
+    def fulltext_backend(
+        self,
+        mocker: pytest_mock.MockerFixture,
+    ) -> Generator[TantivyBackend, None, None]:
+        """An in-memory Tantivy backend, wired up as the module-level
+        singleton _fulltext_similar_documents resolves via get_backend()."""
+        backend = TantivyBackend(path=None)
+        backend.open()
+        mocker.patch("documents.search.get_backend", return_value=backend)
+        try:
+            yield backend
+        finally:
+            backend.close()
+
+    def test_ranks_by_rank_based_weight_descending(
+        self,
+        fulltext_backend: TantivyBackend,
+    ) -> None:
+        """
+        GIVEN:
+            - A source document and two similar documents indexed in Tantivy
+        WHEN:
+            - _fulltext_similar_documents() is called
+        THEN:
+            - Each result's weight reflects its rank (first result weighted
+              higher than the second), not a raw similarity score
+        """
+        source = DocumentFactory.create(content="quarterly financial report details")
+        first = DocumentFactory.create(content="quarterly financial report details")
+        second = DocumentFactory.create(content="financial report")
+        for doc in (source, first, second):
+            fulltext_backend.add_or_update(doc)
+
+        result = _fulltext_similar_documents(source, user=None, top_k=5)
+
+        assert len(result) == 2
+        weight_by_id = {s["document_id"]: s["weight"] for s in result}
+        assert weight_by_id[first.pk] > weight_by_id[second.pk]
+
+    def test_excludes_source_document(
+        self,
+        fulltext_backend: TantivyBackend,
+    ) -> None:
+        """
+        GIVEN:
+            - A source document indexed in Tantivy with no other documents
+        WHEN:
+            - _fulltext_similar_documents() is called
+        THEN:
+            - An empty list is returned - the source document is never its
+              own similar document
+        """
+        source = DocumentFactory.create(content="unique unrelated content")
+        fulltext_backend.add_or_update(source)
+
+        result = _fulltext_similar_documents(source, user=None, top_k=5)
+
+        assert result == []
+
+    def test_empty_index_returns_empty_list(
+        self,
+        fulltext_backend: TantivyBackend,
+    ) -> None:
+        """
+        GIVEN:
+            - A document that has never been indexed (fresh/empty Tantivy index)
+        WHEN:
+            - _fulltext_similar_documents() is called
+        THEN:
+            - An empty list is returned rather than raising
+        """
+        source = DocumentFactory.create(content="never indexed")
+
+        result = _fulltext_similar_documents(source, user=None, top_k=5)
+
+        assert result == []
+
+    def test_respects_top_k_limit(
+        self,
+        fulltext_backend: TantivyBackend,
+    ) -> None:
+        """
+        GIVEN:
+            - A source document and four similar documents indexed
+        WHEN:
+            - _fulltext_similar_documents() is called with top_k=2
+        THEN:
+            - At most 2 results are returned
+        """
+        source = DocumentFactory.create(content="shared overlapping keyword text")
+        fulltext_backend.add_or_update(source)
+        for _ in range(4):
+            fulltext_backend.add_or_update(
+                DocumentFactory.create(content="shared overlapping keyword text"),
+            )
+
+        result = _fulltext_similar_documents(source, user=None, top_k=2)
+
+        assert len(result) == 2
+
+    def test_result_shape_is_similar_document(
+        self,
+        fulltext_backend: TantivyBackend,
+    ) -> None:
+        """
+        GIVEN:
+            - A source document and one similar document indexed
+        WHEN:
+            - _fulltext_similar_documents() is called
+        THEN:
+            - Each result is a SimilarDocument (document_id + weight only)
+        """
+        source = DocumentFactory.create(content="shared content phrase")
+        other = DocumentFactory.create(content="shared content phrase")
+        fulltext_backend.add_or_update(source)
+        fulltext_backend.add_or_update(other)
+
+        result = _fulltext_similar_documents(source, user=None, top_k=5)
+
+        # rank 0 (the only/best result) with top_k=5 -> weight = top_k - rank = 5.0,
+        # per the "first result gets top_k, the last gets 1" formula.
+        assert result == [SimilarDocument(document_id=other.pk, weight=5.0)]
+
+    def test_superuser_sees_other_users_documents(
+        self,
+        fulltext_backend: TantivyBackend,
+    ) -> None:
+        """
+        GIVEN:
+            - A source document owned by one user and a similar document
+              owned by a different user, with no sharing between them
+        WHEN:
+            - _fulltext_similar_documents() is called with a superuser
+        THEN:
+            - The other user's document is still returned as a similar
+              document - a superuser must not be narrowed by the backend's
+              owner-based permission filter
+        """
+        owner = UserFactory.create()
+        other_owner = UserFactory.create()
+        superuser = UserFactory.create(is_superuser=True)
+        source = DocumentFactory.create(
+            content="shared content phrase",
+            owner=owner,
+        )
+        other = DocumentFactory.create(
+            content="shared content phrase",
+            owner=other_owner,
+        )
+        fulltext_backend.add_or_update(source)
+        fulltext_backend.add_or_update(other)
+
+        result = _fulltext_similar_documents(source, user=superuser, top_k=5)
+
+        assert [s["document_id"] for s in result] == [other.pk]
+
+    def test_excludes_stale_permitted_document_for_regular_user(
+        self,
+        fulltext_backend: TantivyBackend,
+    ) -> None:
+        """
+        GIVEN:
+            - A regular (non-superuser) user
+            - A similar document the user is permitted to view, and another
+              similar document indexed while the user still had view
+              permission but which has since had that permission revoked in
+              the database, i.e. the Tantivy index has stale permission data
+        WHEN:
+            - _fulltext_similar_documents() is called with that user
+        THEN:
+            - Only the still-permitted document is returned - the DB
+              re-check via restrict_queryset_to_visible() must catch the
+              document Tantivy's stale index still thinks is visible
+        """
+        owner = UserFactory.create()
+        viewer = UserFactory.create(is_superuser=False)
+        source = DocumentFactory.create(
+            content="shared content phrase",
+            owner=owner,
+        )
+        permitted = DocumentFactory.create(
+            content="shared content phrase",
+            owner=owner,
+        )
+        now_private = DocumentFactory.create(
+            content="shared content phrase",
+            owner=owner,
+        )
+        assign_perm("view_document", viewer, permitted)
+        assign_perm("view_document", viewer, now_private)
+        fulltext_backend.add_or_update(source)
+        fulltext_backend.add_or_update(permitted)
+        fulltext_backend.add_or_update(now_private)
+
+        # Revoke access after indexing, without reindexing: the index still
+        # carries viewer as a permitted viewer for `now_private`.
+        remove_perm("view_document", viewer, now_private)
+
+        result = _fulltext_similar_documents(source, user=viewer, top_k=5)
+
+        assert [s["document_id"] for s in result] == [permitted.pk]
+
+
+@pytest.mark.django_db
+@override_settings(LLM_EMBEDDING_BACKEND="huggingface")
@patch("paperless_ai.ai_classifier.retrieve_similar_nodes")
 def test_get_taxonomy_context_retrieval_failure_degrades_to_no_hints(mock_retrieve):
    """
@@ -575,6 +858,7 @@ def test_get_taxonomy_context_retrieval_failure_degrades_to_no_hints(mock_retrie


@pytest.mark.django_db
+@override_settings(LLM_EMBEDDING_BACKEND="huggingface")
@patch("paperless_ai.ai_classifier.build_taxonomy_candidates")
@patch("paperless_ai.ai_classifier.retrieve_similar_nodes")
 def test_get_taxonomy_context_candidate_building_failure_degrades_to_no_hints(
@@ -1188,9 +1188,7 @@ class TestRetrieveSimilarNodesAgainstRealIndex:

        nodes = indexing.retrieve_similar_nodes(a, document_ids=[b.id])

-        assert all(
-            document_id == b.id for document_id in indexing._node_document_ids(nodes)
-        )
+        assert all(int(node.metadata["document_id"]) == b.id for node in nodes)

    def test_excludes_self(
        self,
@@ -1212,7 +1210,7 @@ class TestRetrieveSimilarNodesAgainstRealIndex:

        nodes = indexing.retrieve_similar_nodes(a, top_k=5)

-        assert set(indexing._node_document_ids(nodes)) == {b.id}
+        assert {int(node.metadata["document_id"]) for node in nodes} == {b.id}

    def test_excludes_self_with_multiple_chunks(
        self,
@@ -1235,4 +1233,4 @@ class TestRetrieveSimilarNodesAgainstRealIndex:

        nodes = indexing.retrieve_similar_nodes(a, top_k=3)

-        assert set(indexing._node_document_ids(nodes)) == {b.id}
+        assert {int(node.metadata["document_id"]) for node in nodes} == {b.id}
@@ -4,6 +4,7 @@ from unittest.mock import MagicMock
 from unittest.mock import patch

 import httpx
+import ollama
 import openai
 import pytest
 from llama_index.core.llms.llm import ToolSelection
@@ -11,6 +12,7 @@ from llama_index.core.llms.llm import ToolSelection
 from paperless_ai.client import LLM_SYSTEM_PROMPT
 from paperless_ai.client import PLACEHOLDER_API_KEY
 from paperless_ai.client import AIClient
+from paperless_ai.exceptions import LLMProviderError
 from paperless_ai.exceptions import LLMTimeoutError


@@ -214,6 +216,52 @@ def test_run_llm_query_openai_timeout_raises_local_error(
        client.run_llm_query("test_prompt")


+def test_run_llm_query_openai_status_error_raises_provider_error(
+    mock_ai_config,
+    mock_openai_llm,
+):
+    mock_ai_config.llm_backend = "openai-like"
+    mock_ai_config.llm_model = "test_model"
+    mock_ai_config.llm_endpoint = "http://test-url"
+
+    request = httpx.Request("POST", "http://test-url/v1/chat/completions")
+    body = {"error": {"message": "Thinking mode does not support this tool_choice"}}
+    mock_openai_llm.return_value.chat_with_tools.side_effect = openai.BadRequestError(
+        "Error code: 400",
+        response=httpx.Response(400, request=request, json=body),
+        body=body,
+    )
+
+    client = AIClient()
+
+    with pytest.raises(LLMProviderError) as exc_info:
+        client.run_llm_query("test_prompt")
+    assert str(exc_info.value) == ""
+    assert isinstance(exc_info.value.__cause__, openai.BadRequestError)
+
+
+def test_run_llm_query_ollama_response_error_raises_provider_error(
+    mock_ai_config,
+    mock_ollama_llm,
+):
+    mock_ai_config.llm_backend = "ollama"
+    mock_ai_config.llm_model = "test_model"
+    mock_ai_config.llm_endpoint = "http://test-url"
+
+    response_error = ollama.ResponseError(
+        "confidential provider response",
+        status_code=400,
+    )
+    mock_ollama_llm.return_value.chat.side_effect = response_error
+
+    client = AIClient()
+
+    with pytest.raises(LLMProviderError) as exc_info:
+        client.run_llm_query("test_prompt")
+    assert str(exc_info.value) == ""
+    assert exc_info.value.__cause__ is response_error
+
+
 def test_run_llm_query_httpx_timeout_raises_local_error(
    mock_ai_config,
    mock_ollama_llm,
@@ -1,5 +1,4 @@
 import json
-from types import SimpleNamespace

 import pytest
 import pytest_mock
@@ -10,14 +9,14 @@ from documents.tests.factories import DocumentTypeFactory
 from documents.tests.factories import StoragePathFactory
 from documents.tests.factories import TagFactory
 from documents.tests.factories import UserFactory
+from paperless_ai.taxonomy import SimilarDocument
 from paperless_ai.taxonomy import TaxonomyCandidates
 from paperless_ai.taxonomy import build_taxonomy_candidates
 from paperless_ai.taxonomy import format_taxonomy_for_prompt


-def make_node(document_id: int, score: float) -> SimpleNamespace:
-    """A stand-in for NodeWithScore: only ``.metadata``/``.score`` are read."""
-    return SimpleNamespace(metadata={"document_id": str(document_id)}, score=score)
+def make_similar(document_id: int, weight: float) -> SimilarDocument:
+    return SimilarDocument(document_id=document_id, weight=weight)


@pytest.mark.django_db
@@ -53,9 +52,9 @@ class TestBuildTaxonomyCandidates:
        doc_a.tags.add(tag)
        doc_b = DocumentFactory.create()
        doc_b.tags.add(tag)
-        nodes = [make_node(doc_a.pk, 0.9), make_node(doc_b.pk, 0.4)]
+        similar_documents = [make_similar(doc_a.pk, 0.9), make_similar(doc_b.pk, 0.4)]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert len(result["tags"]) == 1
        assert result["tags"][0]["id"] == tag.pk
@@ -80,9 +79,9 @@ class TestBuildTaxonomyCandidates:
        document.tags.add(tag)
        tag.name = "New Name"
        tag.save()
-        nodes = [make_node(document.pk, 0.5)]
+        similar_documents = [make_similar(document.pk, 0.5)]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert result["tags"][0]["name"] == "New Name"

@@ -102,9 +101,9 @@ class TestBuildTaxonomyCandidates:
        document = DocumentFactory.create()
        document.tags.add(tag)
        tag.delete()
-        nodes = [make_node(document.pk, 0.5)]
+        similar_documents = [make_similar(document.pk, 0.5)]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert result["tags"] == []

@@ -123,9 +122,12 @@ class TestBuildTaxonomyCandidates:
        strong_doc.tags.add(strong_tag)
        weak_doc = DocumentFactory.create()
        weak_doc.tags.add(weak_tag)
-        nodes = [make_node(strong_doc.pk, 0.9), make_node(weak_doc.pk, 0.1)]
+        similar_documents = [
+            make_similar(strong_doc.pk, 0.9),
+            make_similar(weak_doc.pk, 0.1),
+        ]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert [c["name"] for c in result["tags"]] == ["Strong", "Weak"]

@@ -141,9 +143,9 @@ class TestBuildTaxonomyCandidates:
        document = DocumentFactory.create()
        for i in range(15):
            document.tags.add(TagFactory.create(name=f"Tag{i}"))
-        nodes = [make_node(document.pk, 0.5)]
+        similar_documents = [make_similar(document.pk, 0.5)]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert len(result["tags"]) == 10

@@ -157,12 +159,12 @@ class TestBuildTaxonomyCandidates:
            - Only 5 correspondents are returned
        """
        correspondents = CorrespondentFactory.create_batch(7)
-        nodes = [
-            make_node(DocumentFactory.create(correspondent=c).pk, 0.5)
+        similar_documents = [
+            make_similar(DocumentFactory.create(correspondent=c).pk, 0.5)
            for c in correspondents
        ]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert len(result["correspondents"]) == 5

@@ -177,9 +179,9 @@ class TestBuildTaxonomyCandidates:
        """
        document_type = DocumentTypeFactory.create(name="Invoice")
        document = DocumentFactory.create(document_type=document_type)
-        nodes = [make_node(document.pk, 0.5)]
+        similar_documents = [make_similar(document.pk, 0.5)]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert len(result["document_types"]) == 1
        assert result["document_types"][0]["id"] == document_type.pk
@@ -195,12 +197,12 @@ class TestBuildTaxonomyCandidates:
            - Only 5 document_types are returned
        """
        document_types = DocumentTypeFactory.create_batch(7)
-        nodes = [
-            make_node(DocumentFactory.create(document_type=dt).pk, 0.5)
+        similar_documents = [
+            make_similar(DocumentFactory.create(document_type=dt).pk, 0.5)
            for dt in document_types
        ]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert len(result["document_types"]) == 5

@@ -215,9 +217,9 @@ class TestBuildTaxonomyCandidates:
        """
        storage_path = StoragePathFactory.create(name="Invoices")
        document = DocumentFactory.create(storage_path=storage_path)
-        nodes = [make_node(document.pk, 0.5)]
+        similar_documents = [make_similar(document.pk, 0.5)]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert len(result["storage_paths"]) == 1
        assert result["storage_paths"][0]["id"] == storage_path.pk
@@ -233,12 +235,12 @@ class TestBuildTaxonomyCandidates:
            - Only 5 storage_paths are returned
        """
        storage_paths = StoragePathFactory.create_batch(7)
-        nodes = [
-            make_node(DocumentFactory.create(storage_path=sp).pk, 0.5)
+        similar_documents = [
+            make_similar(DocumentFactory.create(storage_path=sp).pk, 0.5)
            for sp in storage_paths
        ]

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert len(result["storage_paths"]) == 5

@@ -258,14 +260,14 @@ class TestBuildTaxonomyCandidates:
        tag = TagFactory.create(name="Restricted")
        document = DocumentFactory.create()
        document.tags.add(tag)
-        nodes = [make_node(document.pk, 0.5)]
+        similar_documents = [make_similar(document.pk, 0.5)]
        user = UserFactory.create()
        mocker.patch(
            "documents.permissions.permitted_object_ids",
            return_value=[],  # user cannot see this tag
        )

-        result = build_taxonomy_candidates(nodes, user=user)
+        result = build_taxonomy_candidates(similar_documents, user=user)

        assert result["tags"] == []

@@ -295,10 +297,10 @@ class TestBuildTaxonomyCandidates:
        tag.save()
        document = DocumentFactory.create()
        document.tags.add(tag)
-        nodes = [make_node(document.pk, 0.5)]
+        similar_documents = [make_similar(document.pk, 0.5)]
        spy = mocker.patch("documents.permissions.permitted_object_ids")

-        result = build_taxonomy_candidates(nodes, user=None)
+        result = build_taxonomy_candidates(similar_documents, user=None)

        assert result["tags"][0]["name"] == "Owned"
        spy.assert_not_called()
@@ -1,7 +1,9 @@
 import logging

+from celery import Task
 from celery import shared_task

+from documents.models import PaperlessTask
 from paperless_mail.mail import MailAccountHandler
 from paperless_mail.mail import MailError
 from paperless_mail.models import MailAccount
@@ -10,8 +12,26 @@ from paperless_mail.models import MailRule
 logger = logging.getLogger("paperless.mail.tasks")


-@shared_task
-def process_mail_accounts(account_ids: list[int] | None = None) -> str:
+@shared_task(bind=True)
+def process_mail_accounts(self: Task, account_ids: list[int] | None = None) -> str:
+    # A scheduled check can still be running (or queued) when the next one
+    # ProcessedMail dedup only records a message once its
+    # handling has finished, so an overlapping run can still pick up the same
+    # not-yet-recorded message. Skip outright rather than race it.
+    other_mail_fetch_running = (
+        PaperlessTask.objects.filter(
+            task_type=PaperlessTask.TaskType.MAIL_FETCH,
+            status__in=[PaperlessTask.Status.PENDING, PaperlessTask.Status.STARTED],
+        )
+        .exclude(task_id=self.request.id)
+        .exists()
+    )
+    if other_mail_fetch_running:
+        logger.info(
+            "Mail account processing is already running; skipping this run.",
+        )
+        return "Skipped: mail account processing already in progress."
+
    total_new_documents = 0
    accounts = (
        MailAccount.objects.filter(pk__in=account_ids)
@@ -0,0 +1,134 @@
+from typing import Final
+
+import pytest
+import pytest_mock
+
+from documents.models import PaperlessTask
+from documents.tests.factories import PaperlessTaskFactory
+from paperless_mail import tasks
+from paperless_mail.tests.factories import MailAccountFactory
+from paperless_mail.tests.factories import MailRuleFactory
+
+NO_DOCUMENTS_ADDED: Final = "No new documents were added."
+SKIPPED: Final = "Skipped: mail account processing already in progress."
+
+
+@pytest.mark.django_db
+@pytest.mark.usefixtures("account_with_rule")
+class TestProcessMailAccountsOverlap:
+    @pytest.fixture
+    def account_with_rule(self) -> None:
+        """An enabled mail account with a single enabled rule."""
+        account = MailAccountFactory.create()
+        MailRuleFactory.create(account=account, enabled=True)
+
+    @pytest.mark.parametrize(
+        ("status", "expected_result", "expected_call_count"),
+        [
+            pytest.param(
+                PaperlessTask.Status.PENDING,
+                SKIPPED,
+                0,
+                id="pending-task-blocks",
+            ),
+            pytest.param(
+                PaperlessTask.Status.STARTED,
+                SKIPPED,
+                0,
+                id="started-task-blocks",
+            ),
+            pytest.param(
+                PaperlessTask.Status.SUCCESS,
+                NO_DOCUMENTS_ADDED,
+                1,
+                id="finished-task-does-not-block",
+            ),
+        ],
+    )
+    def test_skips_only_while_another_mail_fetch_task_runs(
+        self,
+        mocker: pytest_mock.MockerFixture,
+        status: PaperlessTask.Status,
+        expected_result: str,
+        expected_call_count: int,
+    ) -> None:
+        """
+        GIVEN:
+            - An enabled mail account with a rule
+            - Another mail fetch task row in the given status
+        WHEN:
+            - Mail accounts are processed
+        THEN:
+            - Processing is skipped only if that other task is pending or running
+        """
+        PaperlessTaskFactory.create(
+            task_type=PaperlessTask.TaskType.MAIL_FETCH,
+            trigger_source=PaperlessTask.TriggerSource.SCHEDULED,
+            status=status,
+        )
+
+        mocked_handle = mocker.patch.object(
+            tasks.MailAccountHandler,
+            "handle_mail_account",
+            return_value=0,
+        )
+
+        result = tasks.process_mail_accounts()
+
+        assert mocked_handle.call_count == expected_call_count
+        assert result == expected_result
+
+    def test_runs_when_no_other_mail_fetch_task_exists(
+        self,
+        mocker: pytest_mock.MockerFixture,
+    ) -> None:
+        """
+        GIVEN:
+            - An enabled mail account with a rule
+            - No other mail fetch task rows
+        WHEN:
+            - Mail accounts are processed
+        THEN:
+            - The account is handled
+        """
+        mocked_handle = mocker.patch.object(
+            tasks.MailAccountHandler,
+            "handle_mail_account",
+            return_value=0,
+        )
+
+        result = tasks.process_mail_accounts()
+
+        mocked_handle.assert_called_once()
+        assert result == NO_DOCUMENTS_ADDED
+
+    def test_does_not_skip_due_to_its_own_task_row(
+        self,
+        mocker: pytest_mock.MockerFixture,
+    ) -> None:
+        """
+        GIVEN:
+            - An enabled mail account with a rule
+            - A running mail fetch task row belonging to this very task
+        WHEN:
+            - Mail accounts are processed under that task id
+        THEN:
+            - The task does not skip itself and handles the account
+        """
+        PaperlessTaskFactory.create(
+            task_id="self-task-id",
+            task_type=PaperlessTask.TaskType.MAIL_FETCH,
+            trigger_source=PaperlessTask.TriggerSource.SCHEDULED,
+            status=PaperlessTask.Status.STARTED,
+        )
+
+        mocked_handle = mocker.patch.object(
+            tasks.MailAccountHandler,
+            "handle_mail_account",
+            return_value=0,
+        )
+
+        result = tasks.process_mail_accounts.apply(task_id="self-task-id").result
+
+        mocked_handle.assert_called_once()
+        assert result == NO_DOCUMENTS_ADDED
Author	SHA1	Message	Date
Trenton H	310628699d	Fix: If a file remains stable but has zero size after the stability window has passed, drop it from tracking (#14047 ) A slow writer might create a zero byte file, then take longer than the window to finish the write. We would then queue a zero byte file for consumption and race the writer to most likely fail due to still being empty. Instead, drop zero size files at yeild time. The slow writer may or may not finish, but if it does, there will be a Change.modified event fired again	2026-09-09 16:24:17 +00:00
GitHub Actions	aff0f9cf41	Auto translate strings	2026-09-09 16:04:18 +00:00
shamoon	bf716ebfd1	Fixhancement: better LLM errors (#14031 )	2026-09-09 16:02:43 +00:00
GitHub Actions	7d67a10a35	Auto translate strings	2026-09-09 15:50:33 +00:00
shamoon	8d1bc5dd24	Fix: prevent orphaned versions from bulk delete (#14030 )	2026-09-09 08:48:59 -07:00
Trenton H	43a8d7d412	Enhancement: add Tantivy full-text fallback adapter for taxonomy candidates (#13820 ) This brings users without an embedding backend configured to closer parity with those who do. Reuse the search backend to locate similar documents and use them to provide the LLM with the better suggestion pool to draw from	2026-09-09 15:15:15 +00:00
Trenton H	c40922440b	Fix: prevent overlapping mail-account processing runs (#14046 ) process_mail_accounts had no guard against a scheduled run still being in progress when the next one fires. Skip a run outright if another MAIL_FETCH task is already PENDING/STARTED.	2026-09-09 07:49:49 -07:00