Fix: drop redundant change perms checks on cards open button (#14601)

The init-system-checks step already runs them once and all three services
depend on init-complete. Measured idle anon PSS saved on
Python 3.14: consumer 14 MiB, beat 26 MiB, worker 27 MiB.

docker/rootfs/etc/s6-overlay/s6-rc.d/svc-consumer/run

@@ -10,9 +10,11 @@ if [[ -n "${PAPERLESS_CONSUMER_DISABLE}" ]]; then
 else
 	cd ${PAPERLESS_SRC_DIR}
 
+	# Django system checks already ran once in init-system-checks, skip repeating them
+
 	if [[ -n "${USER_IS_NON_ROOT}" ]]; then
-		exec python3 manage.py document_consumer
+		exec python3 manage.py document_consumer --skip-checks
 	else
-		exec s6-setuidgid paperless python3 manage.py document_consumer
+		exec s6-setuidgid paperless python3 manage.py document_consumer --skip-checks
 	fi
 fi

docker/rootfs/etc/s6-overlay/s6-rc.d/svc-scheduler/run

@@ -3,8 +3,9 @@
 
 cd ${PAPERLESS_SRC_DIR}
 
+# Django system checks already ran once in init-system-checks, skip repeating them
 if [[ -n "${USER_IS_NON_ROOT}" ]]; then
-	exec /usr/local/bin/celery --app paperless beat --loglevel INFO
+	exec /usr/local/bin/celery --skip-checks --app paperless beat --loglevel INFO
 else
-	exec s6-setuidgid paperless /usr/local/bin/celery --app paperless beat --loglevel INFO
+	exec s6-setuidgid paperless /usr/local/bin/celery --skip-checks --app paperless beat --loglevel INFO
 fi

docker/rootfs/etc/s6-overlay/s6-rc.d/svc-worker/run

@@ -3,8 +3,9 @@
 
 cd ${PAPERLESS_SRC_DIR}
 
+# Django system checks already ran once in init-system-checks, skip repeating them
 if [[ -n "${USER_IS_NON_ROOT}" ]]; then
-	exec /usr/local/bin/celery --app paperless worker --loglevel INFO --without-mingle --without-gossip
+	exec /usr/local/bin/celery --skip-checks --app paperless worker --loglevel INFO --without-mingle --without-gossip
 else
-	exec s6-setuidgid paperless /usr/local/bin/celery --app paperless worker --loglevel INFO --without-mingle --without-gossip
+	exec s6-setuidgid paperless /usr/local/bin/celery --skip-checks --app paperless worker --loglevel INFO --without-mingle --without-gossip
 fi

src-ui/e2e/backend.py

@@ -43,6 +43,7 @@ def configure_environment(instance_root: Path) -> None:
 
 
 def seed_database() -> None:
+    from django.contrib.auth.models import Permission
     from django.contrib.auth.models import User
     from django.core.management import call_command
     from django.utils import timezone
@@ -63,11 +64,18 @@ def seed_database() -> None:
         username="playwright",
         password="playwright",  # NOSONAR
     )
-    User.objects.create_user(
+    viewer = User.objects.create_user(
         username="viewer",
         password="viewer",  # NOSONAR
     )
 
+    viewer.user_permissions.add(
+        Permission.objects.get(
+            content_type__app_label="documents",
+            codename="view_uisettings",
+        ),
+    )
+
     inbox = Tag.objects.create(name="Inbox", is_inbox_tag=True, owner=admin)
     quick_filter = Tag.objects.create(name="Another Sample Tag", owner=admin)
     Tag.objects.create(name="TagWithPartial", owner=admin)

src-ui/e2e/document-detail/document-detail.spec.ts

@@ -80,12 +80,21 @@ test('should show a list of notes', async ({ page }) => {
 
 test('should support quick filters', async ({ page }) => {
   await page.goto('/documents/1/details')
+  // filter button is available before tags have loaded
+  const selectedTags = page.locator('pngx-input-tags .ng-value pngx-tag')
+  await expect(selectedTags).toHaveCount(2)
+  await expect(
+    selectedTags.getByText('Another Sample Tag', { exact: true })
+  ).toBeVisible()
+  await expect(selectedTags.getByText('Inbox', { exact: true })).toBeVisible()
   await page
     .getByRole('button', { name: 'Filter documents with these Tags' })
     .click()
-  await expect(page).toHaveURL(
-    /tags__id__all=2,1&sort=created&reverse=1&page=1/
-  )
+  await expect(page).toHaveURL((url) => {
+    const tags = url.searchParams.get('tags__id__all')?.split(',').sort()
+    return url.pathname === '/documents' && tags?.join(',') === '1,2'
+  })
+  await expect(page.locator('pngx-document-list')).toHaveText(/one document/i)
 })
 
 test('should finish reloading the preview after a remote document update', async ({

src-ui/e2e/permissions/global-permissions.spec.ts

@@ -6,8 +6,15 @@ test.use({
   },
 })
 
-test('should not allow user to edit settings', async ({ page }) => {
+test.beforeEach(async ({ page }) => {
+  const settingsLoaded = page.waitForResponse((response) =>
+    response.url().endsWith('/api/ui_settings/')
+  )
   await page.goto('/dashboard')
+  expect((await settingsLoaded).status()).toBe(200)
+})
+
+test('should not allow user to edit settings', async ({ page }) => {
   await expect(page.getByRole('link', { name: 'Settings' })).not.toBeAttached()
   await page.goto('/settings')
   await expect(page.locator('body')).toHaveText(
@@ -16,7 +23,6 @@ test('should not allow user to edit settings', async ({ page }) => {
 })
 
 test('should not allow user to view documents', async ({ page }) => {
-  await page.goto('/dashboard')
   await expect(
     page.locator('nav').getByRole('link', { name: 'Documents' })
   ).not.toBeAttached()
@@ -31,7 +37,6 @@ test('should not allow user to view documents', async ({ page }) => {
 })
 
 test('should not allow user to view correspondents', async ({ page }) => {
-  await page.goto('/dashboard')
   await expect(
     page.getByRole('link', { name: 'Attributes' })
   ).not.toBeAttached()
@@ -42,7 +47,6 @@ test('should not allow user to view correspondents', async ({ page }) => {
 })
 
 test('should not allow user to view tags', async ({ page }) => {
-  await page.goto('/dashboard')
   await expect(
     page.getByRole('link', { name: 'Attributes' })
   ).not.toBeAttached()
@@ -53,7 +57,6 @@ test('should not allow user to view tags', async ({ page }) => {
 })
 
 test('should not allow user to view document types', async ({ page }) => {
-  await page.goto('/dashboard')
   await expect(
     page.getByRole('link', { name: 'Attributes' })
   ).not.toBeAttached()
@@ -64,7 +67,6 @@ test('should not allow user to view document types', async ({ page }) => {
 })
 
 test('should not allow user to view storage paths', async ({ page }) => {
-  await page.goto('/dashboard')
   await expect(
     page.getByRole('link', { name: 'Attributes' })
   ).not.toBeAttached()
@@ -75,7 +77,6 @@ test('should not allow user to view storage paths', async ({ page }) => {
 })
 
 test('should not allow user to view logs', async ({ page }) => {
-  await page.goto('/dashboard')
   await expect(page.getByRole('link', { name: 'Logs' })).not.toBeAttached()
   await page.goto('/logs')
   await expect(page.locator('body')).toHaveText(
@@ -84,7 +85,6 @@ test('should not allow user to view logs', async ({ page }) => {
 })
 
 test('should not allow user to view tasks', async ({ page }) => {
-  await page.goto('/dashboard')
   await expect(page.getByRole('link', { name: 'Tasks' })).not.toBeAttached()
   await page.goto('/tasks')
   await expect(page.locator('body')).toHaveText(

src-ui/messages.xlf

@@ -10334,18 +10334,18 @@
           <context context-type="linenumber">294,295</context>
         </context-group>
       </trans-unit>
-      <trans-unit id="2179847500064178686" datatype="html">
-        <source>Edit document</source>
+      <trans-unit id="3420321797707163677" datatype="html">
+        <source>Preview document</source>
         <context-group purpose="location">
           <context context-type="sourcefile">src/app/components/document-list/document-list.component.html</context>
           <context context-type="linenumber">328</context>
         </context-group>
       </trans-unit>
-      <trans-unit id="3420321797707163677" datatype="html">
-        <source>Preview document</source>
+      <trans-unit id="2179847500064178686" datatype="html">
+        <source>Edit document</source>
         <context-group purpose="location">
           <context context-type="sourcefile">src/app/components/document-list/document-list.component.html</context>
-          <context context-type="linenumber">329</context>
+          <context context-type="linenumber">331</context>
         </context-group>
       </trans-unit>
       <trans-unit id="4512084577073831437" datatype="html">

src-ui/src/app/components/common/custom-fields-dropdown/custom-fields-dropdown.component.scss

@@ -1,10 +1,3 @@
 .custom-fields-dropdown {
     min-width: 300px;
-
-    // correct position on mobile
-    @media (max-width: 575.98px) {
-        &.show {
-            margin-left: -245px !important;
-        }
-    }
 }

src-ui/src/app/components/document-detail/document-detail.component.html

@@ -308,7 +308,7 @@
                   </tr>
                   <tr class="align-middle">
                     <td i18n>Original SHA256 checksum</td>
-                    <td><span class="badge bg-light font-monospace text-wrap text-break text-start">{{metadata()?.original_checksum}}</span></td>
+                    <td><span class="badge bg-light text-dark font-monospace text-wrap text-break text-start">{{metadata()?.original_checksum}}</span></td>
                   </tr>
                   <tr>
                     <td i18n>Original file size</td>
@@ -465,7 +465,7 @@
     @switch (archiveContentRenderType) {
       @case (ContentRenderType.PDF) {
         @if (!useNativePdfViewer) {
-          <div class="preview-sticky pdf-viewer-container">
+          <div class="preview-sticky pdf-viewer-container" tabindex="-1">
             <pngx-pdf-viewer
               [src]="pdfSource()"
               [sourceRevision]="previewRevision()"
@@ -477,7 +477,7 @@
               (loadError)="onError($event)"
               (afterLoadComplete)="pdfPreviewLoaded($event)">
             </pngx-pdf-viewer>
-            <div class="position-absolute bottom-0 start-50 translate-middle-x z-1 mb-3 d-flex gap-2 p-1 bg-body border rounded shadow">
+            <div class="preview-controls position-absolute bottom-0 start-50 translate-middle-x z-1 mb-3 d-flex gap-2 p-1 bg-body border rounded shadow">
               @if (previewNumPages()) {
                 <div class="input-group input-group-sm w-auto flex-nowrap">
                   <input class="form-control flex-grow-0 w-auto" type="number" min="1" [max]="previewNumPages()" [ngModelOptions]="{standalone: true}" [ngModel]="previewCurrentPage()" (ngModelChange)="previewCurrentPage.set($event)" aria-label="Page" i18n-aria-label />

src-ui/src/app/components/document-detail/document-detail.component.scss

@@ -12,6 +12,22 @@
     width: 100%;
     height: 100%;
   }
+
+  &:focus {
+    outline: none;
+  }
+
+  .preview-controls {
+    opacity: 0;
+    pointer-events: none;
+    transition: opacity 0.15s ease-in-out;
+  }
+
+  &:hover .preview-controls,
+  &:focus-within .preview-controls {
+    opacity: 1;
+    pointer-events: auto;
+  }
 }
 
 .btn-group .dropdown-toggle-split {

src-ui/src/app/components/document-list/document-card-large/document-card-large.component.html

@@ -72,7 +72,7 @@
               <a class="btn btn-sm btn-outline-secondary" (click)="clickMoreLike.emit()">
                 <i-bs name="diagram-3" class="me-1"></i-bs><span class="d-none d-md-inline" i18n>More like this</span>
               </a>
-              <a routerLink="/documents/{{document().id}}" class="btn btn-sm btn-outline-secondary" *pngxIfPermissions="{ action: PermissionAction.Change, type: PermissionType.Document }">
+              <a routerLink="/documents/{{document().id}}" class="btn btn-sm btn-outline-secondary">
                 <i-bs name="file-earmark-richtext" class="me-1"></i-bs><span class="d-none d-md-inline" i18n>Open</span>
               </a>
               <pngx-preview-popup [document]="document()" #popupPreview>

src-ui/src/app/components/document-list/document-card-large/document-card-large.component.spec.ts

@@ -10,7 +10,6 @@ import {
   NgbTooltipModule,
 } from '@ng-bootstrap/ng-bootstrap'
 import { NgxBootstrapIconsModule, allIcons } from 'ngx-bootstrap-icons'
-import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { CustomDatePipe } from 'src/app/pipes/custom-date.pipe'
 import { DocumentTitlePipe } from 'src/app/pipes/document-title.pipe'
 import { IsNumberPipe } from 'src/app/pipes/is-number.pipe'
@@ -52,7 +51,6 @@ describe('DocumentCardLargeComponent', () => {
         DocumentCardLargeComponent,
         DocumentTitlePipe,
         CustomDatePipe,
-        IfPermissionsDirective,
         SafeUrlPipe,
         IsNumberPipe,
         PreviewPopupComponent,

src-ui/src/app/components/document-list/document-card-large/document-card-large.component.ts

@@ -20,7 +20,6 @@ import {
   Document,
 } from 'src/app/data/document'
 import { SETTINGS_KEYS } from 'src/app/data/ui-settings'
-import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { CorrespondentNamePipe } from 'src/app/pipes/correspondent-name.pipe'
 import { CustomDatePipe } from 'src/app/pipes/custom-date.pipe'
 import { DocumentTitlePipe } from 'src/app/pipes/document-title.pipe'
@@ -51,7 +50,6 @@ import { LoadingComponentWithPermissions } from '../../loading-component/loading
     CorrespondentNamePipe,
     DocumentTypeNamePipe,
     StoragePathNamePipe,
-    IfPermissionsDirective,
     CustomDatePipe,
     RouterModule,
     NgbTooltipModule,

src-ui/src/app/components/document-list/document-card-small/document-card-small.component.html

@@ -144,7 +144,7 @@
       <div class="d-flex justify-content-between align-items-center">
         <div class="btn-group w-100">
            @if (document()) {
-            <a routerLink="/documents/{{document().id}}" class="btn btn-sm btn-outline-secondary" title="Open" i18n-title *pngxIfPermissions="{ action: PermissionAction.Change, type: PermissionType.Document }" i18n-title>
+            <a routerLink="/documents/{{document().id}}" class="btn btn-sm btn-outline-secondary" title="Open" i18n-title>
               <i-bs name="file-earmark-richtext"></i-bs>
             </a>
             <pngx-preview-popup [document]="document()" #popupPreview>

src-ui/src/app/components/document-list/document-card-small/document-card-small.component.ts

@@ -20,7 +20,6 @@ import {
   Document,
 } from 'src/app/data/document'
 import { SETTINGS_KEYS } from 'src/app/data/ui-settings'
-import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { CorrespondentNamePipe } from 'src/app/pipes/correspondent-name.pipe'
 import { CustomDatePipe } from 'src/app/pipes/custom-date.pipe'
 import { DocumentTitlePipe } from 'src/app/pipes/document-title.pipe'
@@ -51,7 +50,6 @@ import { LoadingComponentWithPermissions } from '../../loading-component/loading
     CorrespondentNamePipe,
     DocumentTypeNamePipe,
     StoragePathNamePipe,
-    IfPermissionsDirective,
     CustomDatePipe,
     RouterModule,
     NgbTooltipModule,

src-ui/src/app/components/document-list/document-list.component.html

@@ -208,7 +208,7 @@
                   i18n>Correspondent</th>
               }
               @if (activeDisplayFields.includes(DisplayField.TITLE)) {
-                <th class="cursor-pointer"
+                <th class="cursor-pointer ps-5"
                   pngxSortable="title"
                   title="Sort by title" i18n-title
                   [currentSortField]="list.sortField"
@@ -322,13 +322,13 @@
                   </td>
                 }
                 @if (activeDisplayFields.includes(DisplayField.TITLE) || activeDisplayFields.includes(DisplayField.TAGS)) {
-                  <td width="30%">
+                  <td width="30%" class="position-relative" [class.ps-5]="activeDisplayFields.includes(DisplayField.TITLE)">
                     @if (activeDisplayFields.includes(DisplayField.TITLE)) {
                       <div class="d-inline-block" (mouseleave)="popupPreview.close()">
-                        <a routerLink="/documents/{{d.id}}" title="Edit document" i18n-title style="overflow-wrap: anywhere;">{{d.title | documentTitle}}</a>
-                        <pngx-preview-popup [document]="d" linkClasses="btn btn-sm btn-link text-secondary" linkTitle="Preview document" (click)="$event.stopPropagation()" i18n-linkTitle #popupPreview>
+                        <pngx-preview-popup class="position-absolute start-0" [document]="d" linkClasses="btn btn-sm btn-link text-secondary" linkTitle="Preview document" (click)="$event.stopPropagation()" i18n-linkTitle #popupPreview>
                           <i-bs name="eye"></i-bs>
                         </pngx-preview-popup>
+                        <a routerLink="/documents/{{d.id}}" title="Edit document" i18n-title style="overflow-wrap: anywhere;">{{d.title | documentTitle}}</a>
                       </div>
                     }
                     @if (activeDisplayFields.includes(DisplayField.TAGS)) {

src-ui/src/app/components/document-list/document-list.component.scss

@@ -87,6 +87,10 @@ $paperless-card-breakpoints: (
   }
 }
 
+.table .ps-5 {
+  padding-left: 2.2rem !important;
+}
+
 a {
   cursor: pointer;
 }

src/documents/barcodes.py

@@ -10,9 +10,6 @@ from typing import TYPE_CHECKING
 import regex as regex_mod
 from django.conf import settings
 from pdf2image import convert_from_path
-from pikepdf import Page
-from pikepdf import PasswordError
-from pikepdf import Pdf
 
 from documents.converters import convert_from_tiff_to_pdf
 from documents.data_models import ConsumableDocument
@@ -34,6 +31,7 @@ from documents.utils import maybe_override_pixel_limit
 from paperless.config import BarcodeConfig
 
 if TYPE_CHECKING:
+    from pikepdf import Page
     from PIL import Image
 
 logger = logging.getLogger("paperless.barcodes")
@@ -282,6 +280,8 @@ class BarcodePlugin(ConsumeTaskPlugin):
         # No op if not a TIFF
         self.convert_from_tiff_to_pdf()
 
+        from pikepdf import PasswordError
+
         try:
             self.barcodes = scan_pdf(
                 self.pdf_file,
@@ -433,6 +433,8 @@ class BarcodePlugin(ConsumeTaskPlugin):
 
         document_paths = []
         fname: str = self.input_doc.original_file.stem
+        from pikepdf import Pdf
+
         with Pdf.open(self.pdf_file) as input_pdf:
             # Start with an empty document
             current_document: list[Page] = []
@@ -516,6 +518,8 @@ def scan_pdf(pdf_path: Path, settings: BarcodeConfig, work_dir: Path) -> list[Ba
     """
     barcodes: list[Barcode] = []
 
+    from pikepdf import Pdf
+
     with Pdf.open(pdf_path) as pdf:
         num_of_pages = len(pdf.pages)
     logger.debug(f"PDF has {num_of_pages} pages")

src/documents/converters.py

@@ -1,6 +1,5 @@
 from pathlib import Path
 
-import img2pdf
 from PIL import Image
 
 from documents.utils import copy_basic_file_stats
@@ -17,6 +16,8 @@ def convert_from_tiff_to_pdf(tiff_path: Path, target_directory: Path) -> Path:
 
     Returns the path of the PDF created.
     """
+    import img2pdf
+
     # override pixel setting if needed
     maybe_override_pixel_limit()
 

src/documents/double_sided.py

@@ -6,7 +6,6 @@ from pathlib import Path
 from typing import Final
 
 from django.conf import settings
-from pikepdf import Pdf
 
 from documents.consumer import ConsumerError
 from documents.converters import convert_from_tiff_to_pdf
@@ -85,6 +84,8 @@ class CollatePlugin(NoCleanupPluginMixin, NoSetupPluginMixin, ConsumeTaskPlugin)
                 valid_staging_exists = True
 
         if valid_staging_exists:
+            from pikepdf import Pdf
+
             try:
                 # Collate pages from second PDF in reverse order
                 with Pdf.open(staging) as pdf1, Pdf.open(pdf_file) as pdf2:

src/documents/filters.py

@@ -11,7 +11,6 @@ from decimal import InvalidOperation
 from typing import TYPE_CHECKING
 from typing import Any
 
-import dateparser
 from django.conf import settings
 from django.contrib.contenttypes.models import ContentType
 from django.db.models import Case
@@ -350,6 +349,8 @@ class DateQueryField(serializers.DateField):
             if not isinstance(value, str):
                 raise
 
+        import dateparser
+
         try:
             parsed = dateparser.parse(
                 value,

src/documents/pdf_ops.py

@@ -14,8 +14,6 @@ from __future__ import annotations
 from typing import TYPE_CHECKING
 from typing import NamedTuple
 
-import pikepdf
-
 if TYPE_CHECKING:
     from collections.abc import Callable
     from collections.abc import Iterable
@@ -43,6 +41,8 @@ def rotate_pdf(src: Path, dst: Path, degrees: int) -> None:
     Rotate every page relatively on the opened document, not a rebuild, so Info,
     XMP and outlines are kept. ``src`` is not modified.
     """
+    import pikepdf
+
     with pikepdf.open(src) as pdf:
         for page in pdf.pages:
             page.rotate(degrees, relative=True)
@@ -58,6 +58,8 @@ def remove_pages(src: Path, dst: Path, pages: Iterable[int]) -> None:
     Duplicates are ignored. Pages are removed highest first so earlier removals
     never shift the index of later ones.
     """
+    import pikepdf
+
     unique = sorted(set(pages))
     _require_positive(unique)
     with pikepdf.open(src) as pdf:
@@ -81,6 +83,8 @@ def build_pdfs(
     (Info, XMP, outlines) is not carried over. Returns the written paths in output
     order.
     """
+    import pikepdf
+
     for specs, _ in outputs:
         _require_positive(spec.page for spec in specs)
 
@@ -140,6 +144,8 @@ def needs_decrypt(src: Path) -> bool:
     True if ``src`` is encrypted. A PDF that needs a password to open at all
     counts as encrypted.
     """
+    import pikepdf
+
     try:
         with pikepdf.open(src) as pdf:
             return bool(pdf.is_encrypted)
@@ -154,6 +160,8 @@ def decrypt_pdf(src: Path, make_dst: Callable[[], Path], password: str) -> Path:
     ``make_dst`` is only called once the password has been accepted, so a wrong
     password never leaves a destination behind.
     """
+    import pikepdf
+
     with pikepdf.open(src, password=password) as pdf:
         pdf.remove_unreferenced_resources()
         dst = make_dst()
@@ -170,6 +178,8 @@ class PdfMerger:
     """
 
     def __init__(self) -> None:
+        import pikepdf
+
         self._pdf = pikepdf.new()
         self._version: str = self._pdf.pdf_version
 
@@ -185,6 +195,8 @@ class PdfMerger:
         self._pdf.close()
 
     def add(self, path: Path) -> None:
+        import pikepdf
+
         with pikepdf.open(str(path)) as pdf:
             self._version = max(self._version, pdf.pdf_version)
             self._pdf.pages.extend(pdf.pages)

src/documents/permissions.py

@@ -576,9 +576,12 @@ def annotate_document_count_by_ids(
                        or CustomFieldInstance)
         related_object_field: field on the relation pointing back to queryset pk
         document_ids: the document ids to count against -- a concrete list/set,
-                       or a simple (already resolved) queryset of ids. Keep
-                       the list small: it is embedded in the statement as
-                       literal parameters.
+                       or a simple (already resolved) queryset of ids. Callers
+                       that need this filtered by a complex condition (e.g. a
+                       permission check) should resolve it to a concrete list
+                       first if the same ids will be reused across multiple
+                       calls, rather than passing the complex queryset itself
+                       into each -- see ``_get_selection_data_for_queryset``.
         target_field: field on the relation pointing to Document id
     """
 

src/documents/plugins/date_parsing/base.py

@@ -7,8 +7,6 @@ from dataclasses import dataclass
 from types import TracebackType
 from typing import Self
 
-import dateparser
-
 logger = logging.getLogger(__name__)
 
 
@@ -80,6 +78,8 @@ class DateParserPluginBase(ABC):
 
         Uses configuration from `self.config`.
         """
+        import dateparser
+
         try:
             return dateparser.parse(
                 date_string,

src/documents/tasks.py

@@ -76,9 +76,6 @@ from paperless.logging import consume_task_id
 from paperless.parsers import ParserContext
 from paperless.parsers.registry import get_parser_registry
 from paperless_ai.exceptions import LLMTimeoutError
-from paperless_ai.indexing import llm_index_add_or_update_document
-from paperless_ai.indexing import llm_index_remove_document
-from paperless_ai.indexing import update_llm_index
 
 if settings.AUDIT_LOG_ENABLED:
     from auditlog.models import LogEntry
@@ -339,6 +336,8 @@ def bulk_update_documents(document_ids) -> None:
 
     ai_config = AIConfig()
     if ai_config.llm_index_enabled:
+        from paperless_ai.indexing import update_llm_index
+
         update_llm_index(
             rebuild=False,
             document_ids=document_ids,
@@ -499,6 +498,8 @@ def update_document_content_maybe_archive_file(
 
             ai_config = AIConfig()
             if ai_config.llm_index_enabled:
+                from paperless_ai.indexing import llm_index_add_or_update_document
+
                 llm_index_add_or_update_document(document)
 
             clear_document_caches(document.pk)
@@ -800,11 +801,15 @@ def apply_ai_suggestions(self, action_id: int, document_id: int) -> None:
 
 @shared_task
 def update_document_in_llm_index(document) -> None:
+    from paperless_ai.indexing import llm_index_add_or_update_document
+
     llm_index_add_or_update_document(document)
 
 
 @shared_task
 def remove_document_from_llm_index(document) -> None:
+    from paperless_ai.indexing import llm_index_remove_document
+
     llm_index_remove_document(document)
 
 

src/documents/tests/date_parsing/test_date_parsing.py

@@ -71,7 +71,7 @@ class TestParseString:
         ):
             # We still need to mock dateparser.parse to force the exception
             mocker.patch(
-                "documents.plugins.date_parsing.base.dateparser.parse",
+                "dateparser.parse",
                 side_effect=ValueError(
                     "Parsing error: 01/01/2024",
                 ),
@@ -260,7 +260,7 @@ class TestRegexDateParser:
         parser = RegexDateParserPlugin(base_config)
 
         # Patch the dateparser.parse
-        target = "documents.plugins.date_parsing.base.dateparser.parse"
+        target = "dateparser.parse"
 
         def fake_parse(
             date_string: str,
@@ -335,7 +335,7 @@ class TestRegexDateParser:
         )
         parser = RegexDateParserPlugin(cfg)
 
-        target = "documents.plugins.date_parsing.base.dateparser.parse"
+        target = "dateparser.parse"
 
         def fake_parse(
             date_string: str,
@@ -400,7 +400,7 @@ class TestRegexDateParser:
         parser = RegexDateParserPlugin(cfg)
 
         # Patch the module's dateparser.parse so we can inspect calls
-        target = "documents.plugins.date_parsing.base.dateparser.parse"
+        target = "dateparser.parse"
 
         def fake_parse(
             date_string: str,

src/documents/tests/test_api_app_config.py

@@ -1070,7 +1070,7 @@ class TestApiAppConfig(DirectoriesMixin, APITestCase):
 
         with (
             patch("documents.tasks.llmindex_index.apply_async") as mock_update,
-            patch("paperless.views.llm_index_exists") as mock_exists,
+            patch("paperless_ai.indexing.llm_index_exists") as mock_exists,
         ):
             mock_exists.return_value = False
             self.client.patch(
@@ -1095,7 +1095,7 @@ class TestApiAppConfig(DirectoriesMixin, APITestCase):
 
         with (
             patch("documents.tasks.llmindex_index.apply_async") as mock_update,
-            patch("paperless.views.llm_index_exists", return_value=False),
+            patch("paperless_ai.indexing.llm_index_exists", return_value=False),
         ):
             self.client.patch(
                 f"{self.ENDPOINT}1/",
@@ -1115,7 +1115,7 @@ class TestApiAppConfig(DirectoriesMixin, APITestCase):
 
         with (
             patch("documents.tasks.llmindex_index.apply_async") as mock_update,
-            patch("paperless.views.llm_index_exists") as mock_exists,
+            patch("paperless_ai.indexing.llm_index_exists") as mock_exists,
         ):
             mock_exists.return_value = True
             self.client.patch(
@@ -1136,7 +1136,7 @@ class TestApiAppConfig(DirectoriesMixin, APITestCase):
 
         with (
             patch("documents.tasks.llmindex_index.apply_async") as mock_update,
-            patch("paperless.views.llm_index_exists") as mock_exists,
+            patch("paperless_ai.indexing.llm_index_exists") as mock_exists,
         ):
             mock_exists.return_value = True
             self.client.patch(
@@ -1174,7 +1174,7 @@ class TestApiAppConfig(DirectoriesMixin, APITestCase):
 
         with (
             patch("documents.tasks.llmindex_index.apply_async") as mock_update,
-            patch("paperless.views.llm_index_exists") as mock_exists,
+            patch("paperless_ai.indexing.llm_index_exists") as mock_exists,
         ):
             mock_exists.return_value = True
             self.client.patch(

src/documents/tests/test_api_bulk_edit.py

@@ -1141,8 +1141,8 @@ class TestBulkEditAPI(DirectoriesMixin, APITestCase):
         self.assertCountEqual(
             response.data["selected_document_types"],
             [
-                {"id": self.dt1.id, "document_count": 1},
-                {"id": self.dt2.id, "document_count": 0},
+                {"id": self.c1.id, "document_count": 1},
+                {"id": self.c2.id, "document_count": 0},
             ],
         )
 

src/documents/tests/test_api_chat.py

@@ -61,7 +61,7 @@ class TestChatStreamingViewInputValidation(APITestCase):
                 return_value=self._mock_ai_enabled(),
             ),
             mock.patch(
-                "documents.views.stream_chat_with_documents",
+                "paperless_ai.chat.stream_chat_with_documents",
                 return_value=iter(chunks),
             ),
         ):
@@ -104,7 +104,7 @@ class TestChatStreamingViewUnrestrictedFlag:
         """
         mocker.patch("documents.views.AIConfig").return_value.ai_enabled = True
         return mocker.patch(
-            "documents.views.stream_chat_with_documents",
+            "paperless_ai.chat.stream_chat_with_documents",
             return_value=iter(()),
         )
 

src/documents/tests/test_api_document_versions.py

@@ -1110,7 +1110,7 @@ class TestVersionActionPermissions(DirectoriesMixin, APITestCase):
             self.assertEqual(response.status_code, 403)
             with (
                 mock.patch("documents.views.AIConfig") as ai_config,
-                mock.patch("documents.views.stream_chat_with_documents") as chat,
+                mock.patch("paperless_ai.chat.stream_chat_with_documents") as chat,
             ):
                 ai_config.return_value.ai_enabled = True
                 response = self.client.post(

src/documents/tests/test_api_documents.py

@@ -49,7 +49,6 @@ from documents.models import WorkflowTrigger
 from documents.search import get_backend
 from documents.signals.handlers import run_workflows
 from documents.tests.utils import ConsumeTaskMixin
-from documents.views import DocumentViewSet
 from paperless_testing.dirs import DirectoriesMixin
 from paperless_testing.factories import DocumentFactory
 from paperless_testing.factories import TagFactory
@@ -1756,106 +1755,6 @@ class TestDocumentApi(DirectoriesMixin, ConsumeTaskMixin, APITestCase):
         }
         self.assertEqual(counts_by_field[custom_field.id], 1)
 
-    def test_selection_data_counts_with_shared_by_filter(self) -> None:
-        """
-        GIVEN:
-            - Several documents owned by one user with the same correspondent
-              and document type, only one of which is shared
-        WHEN:
-            - Requesting the document list filtered by shared_by__id with
-              include_selection_data=true
-        THEN:
-            - Counts reflect only the shared document, not every document the
-              owner has with that correspondent or document type, however the
-              link-table lookups are batched
-        """
-        owner = User.objects.create_user(username="owner")
-        other = User.objects.create_user(username="other")
-        correspondent = Correspondent.objects.create(name="c1")
-        doc_type = DocumentType.objects.create(name="dt1")
-        tag = Tag.objects.create(name="tag")
-        docs = [
-            Document.objects.create(
-                checksum=str(i),
-                owner=owner,
-                correspondent=correspondent,
-                document_type=doc_type,
-            )
-            for i in range(3)
-        ]
-        for doc in docs:
-            doc.tags.add(tag)
-        grant_object(other, docs[0], "view_document")
-
-        for batch_size in (10_000, 1):
-            with (
-                self.subTest(link_batch_size=batch_size),
-                mock.patch.object(
-                    DocumentViewSet,
-                    "selection_data_link_batch_size",
-                    batch_size,
-                ),
-            ):
-                response = self.client.get(
-                    f"/api/documents/?shared_by__id={owner.id}"
-                    "&include_selection_data=true",
-                )
-
-                self.assertEqual(response.status_code, status.HTTP_200_OK)
-                self.assertEqual(response.data["count"], 1)
-                selection_data = response.data["selection_data"]
-                for key, obj in (
-                    ("selected_correspondents", correspondent),
-                    ("selected_document_types", doc_type),
-                    ("selected_tags", tag),
-                ):
-                    counts = {
-                        item["id"]: item["document_count"]
-                        for item in selection_data[key]
-                    }
-                    self.assertEqual(counts[obj.id], 1, key)
-
-    def test_selection_data_counts_documents_once_when_filter_joins(self) -> None:
-        """
-        GIVEN:
-            - A document matching a multi-valued filter through two tags
-        WHEN:
-            - Requesting the document list with include_selection_data=true
-        THEN:
-            - The document is counted once per correspondent, not once per
-              matching tag, however the link-table lookups are batched
-        """
-        correspondent = Correspondent.objects.create(name="c1")
-        tag_a = Tag.objects.create(name="a")
-        tag_b = Tag.objects.create(name="b")
-        both = Document.objects.create(checksum="1", correspondent=correspondent)
-        both.tags.add(tag_a, tag_b)
-        only_a = Document.objects.create(checksum="2", correspondent=correspondent)
-        only_a.tags.add(tag_a)
-
-        for batch_size in (10_000, 1):
-            with (
-                self.subTest(link_batch_size=batch_size),
-                mock.patch.object(
-                    DocumentViewSet,
-                    "selection_data_link_batch_size",
-                    batch_size,
-                ),
-            ):
-                response = self.client.get(
-                    f"/api/documents/?tags__id__in={tag_a.id},{tag_b.id}"
-                    "&include_selection_data=true",
-                )
-
-                self.assertEqual(response.status_code, status.HTTP_200_OK)
-                counts = {
-                    item["id"]: item["document_count"]
-                    for item in response.data["selection_data"][
-                        "selected_correspondents"
-                    ]
-                }
-                self.assertEqual(counts[correspondent.id], 2)
-
     def test_statistics(self) -> None:
         doc1 = Document.objects.create(
             title="none1",
@@ -2787,7 +2686,7 @@ class TestDocumentApi(DirectoriesMixin, ConsumeTaskMixin, APITestCase):
         response = self.client.get("/api/documents/34676/suggestions/")
         self.assertEqual(response.status_code, status.HTTP_404_NOT_FOUND)
 
-    @mock.patch("documents.views.get_ai_document_classification")
+    @mock.patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(AI_ENABLED=True)
     def test_suggestions_still_uses_classifier_when_ai_enabled(
         self,

src/documents/tests/test_permission_filtering_security.py

@@ -572,7 +572,7 @@ class TestAiChatAllDocumentsPermissionBoundary:
     ENDPOINT = "/api/documents/chat/"
 
     @override_settings(AI_ENABLED=True)
-    @patch("documents.views.stream_chat_with_documents")
+    @patch("paperless_ai.chat.stream_chat_with_documents")
     def test_chat_all_documents_excludes_unshared_document(self, mock_stream_chat):
         mock_stream_chat.return_value = iter([b"data"])
 

src/documents/tests/test_tasks.py

@@ -322,7 +322,7 @@ class TestUpdateContent(DirectoriesMixin, TestCase):
         )
 
     @override_settings(AI_ENABLED=True, LLM_EMBEDDING_BACKEND="huggingface")
-    @mock.patch("documents.tasks.llm_index_add_or_update_document")
+    @mock.patch("paperless_ai.indexing.llm_index_add_or_update_document")
     @mock.patch("documents.search.get_backend")
     def test_update_content_version_updates_llm_index(
         self,
@@ -450,7 +450,7 @@ class TestAIIndex(DirectoriesMixin, TestCase):
             checksum="wow",
         )
         with mock.patch(
-            "documents.tasks.llm_index_add_or_update_document",
+            "paperless_ai.indexing.llm_index_add_or_update_document",
         ) as llm_index_add_or_update_document:
             tasks.update_document_in_llm_index(doc)
             llm_index_add_or_update_document.assert_called_once_with(doc)
@@ -470,7 +470,7 @@ class TestAIIndex(DirectoriesMixin, TestCase):
             checksum="wow",
         )
         with mock.patch(
-            "documents.tasks.llm_index_remove_document",
+            "paperless_ai.indexing.llm_index_remove_document",
         ) as llm_index_remove_document:
             tasks.remove_document_from_llm_index(doc)
             llm_index_remove_document.assert_called_once_with(doc)
@@ -495,7 +495,7 @@ class TestAIIndex(DirectoriesMixin, TestCase):
                 "documents.tasks.update_document_in_llm_index",
             ) as update_document_in_llm_index,
             mock.patch(
-                "documents.tasks.update_llm_index",
+                "paperless_ai.indexing.update_llm_index",
             ) as update_llm_index,
         ):
             doc_ids = [doc.pk for doc in docs]

src/documents/tests/test_views.py

@@ -474,7 +474,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
         self.assertEqual(response.json()["tags"], [])
         self.assertEqual(response.json()["suggested_tags"], [])
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="mock_backend",
@@ -524,7 +524,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
             None,
         )
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="mock_backend",
@@ -562,7 +562,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
             "KI Title",
         )
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="mock_backend",
@@ -601,7 +601,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
             "Titre IA",
         )
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="mock_backend",
@@ -637,7 +637,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
             ),
         )
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="mock_backend",
@@ -709,7 +709,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
             ),
         )
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="openai-like",
@@ -738,7 +738,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
             get_llm_suggestion_cache(self.document.pk, backend="openai-like"),
         )
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="openai-like",
@@ -765,7 +765,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
             get_llm_suggestion_cache(self.document.pk, backend="openai-like"),
         )
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="openai-like",
@@ -797,7 +797,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
             get_llm_suggestion_cache(self.document.pk, backend="openai-like"),
         )
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="mock_backend",
@@ -835,7 +835,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
         self.assertEqual(response.json()["tags"], [self.tag1.pk])
         self.assertEqual(response.json()["suggested_tags"], ["Follow-up"])
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="mock_backend",
@@ -874,7 +874,7 @@ class TestAISuggestions(DirectoriesMixin, TestCase):
         self.assertEqual(response.json()["tags"], [self.tag1.pk])
         self.assertEqual(response.json()["suggested_tags"], [])
 
-    @patch("documents.views.get_ai_document_classification")
+    @patch("paperless_ai.ai_classifier.get_ai_document_classification")
     @override_settings(
         AI_ENABLED=True,
         LLM_BACKEND="mock_backend",
@@ -996,7 +996,7 @@ class TestAIChatStreamingView(DirectoriesMixin, TestCase):
         self.assertEqual(response.status_code, 400)
         self.assertIn(b"AI is required for this feature", response.content)
 
-    @patch("documents.views.stream_chat_with_documents")
+    @patch("paperless_ai.chat.stream_chat_with_documents")
     @patch("documents.views.permitted_document_ids")
     @override_settings(AI_ENABLED=True)
     def test_post_no_document_id(self, mock_permitted_ids, mock_stream_chat) -> None:
@@ -1016,7 +1016,7 @@ class TestAIChatStreamingView(DirectoriesMixin, TestCase):
         self.assertEqual(list(call_kwargs["documents"]), [self.document])
         self.assertIsNone(call_kwargs["output_language"])
 
-    @patch("documents.views.stream_chat_with_documents")
+    @patch("paperless_ai.chat.stream_chat_with_documents")
     @patch("documents.views.permitted_document_ids")
     @override_settings(AI_ENABLED=True)
     def test_post_uses_user_display_language(
@@ -1042,7 +1042,7 @@ class TestAIChatStreamingView(DirectoriesMixin, TestCase):
         self.assertEqual(list(call_kwargs["documents"]), [self.document])
         self.assertEqual(call_kwargs["output_language"], "de-de")
 
-    @patch("documents.views.stream_chat_with_documents")
+    @patch("paperless_ai.chat.stream_chat_with_documents")
     @override_settings(AI_ENABLED=True)
     def test_post_with_document_id(self, mock_stream_chat) -> None:
         self.grant_view_document_permission()
@@ -1079,7 +1079,7 @@ class TestAIChatStreamingView(DirectoriesMixin, TestCase):
         self.assertEqual(response.status_code, 403)
         self.assertIn(b"Insufficient permissions", response.content)
 
-    @patch("documents.views.stream_chat_with_documents")
+    @patch("paperless_ai.chat.stream_chat_with_documents")
     @override_settings(AI_ENABLED=True)
     def test_post_no_document_id_requires_view_document_permission(
         self,
@@ -1093,7 +1093,7 @@ class TestAIChatStreamingView(DirectoriesMixin, TestCase):
         self.assertEqual(response.status_code, 403)
         mock_stream_chat.assert_not_called()
 
-    @patch("documents.views.stream_chat_with_documents")
+    @patch("paperless_ai.chat.stream_chat_with_documents")
     @override_settings(AI_ENABLED=True)
     def test_post_with_document_id_requires_view_document_permission(
         self,

src/documents/views.py

@@ -5,7 +5,6 @@ import platform
 import re
 import tempfile
 import zipfile
-from collections import Counter
 from collections import deque
 from collections.abc import Callable
 from datetime import datetime
@@ -169,6 +168,7 @@ from documents.permissions import PaperlessNotePermissions
 from documents.permissions import PaperlessObjectPermissions
 from documents.permissions import TrashPermissions
 from documents.permissions import ViewDocumentsPermissions
+from documents.permissions import annotate_document_count_by_ids
 from documents.permissions import annotate_document_count_for_related_queryset
 from documents.permissions import get_document_count_filter_for_user
 from documents.permissions import get_permission_grantees
@@ -250,20 +250,8 @@ from paperless.parsers.remote import RemoteEngineConfig
 from paperless.serialisers import GroupSerializer
 from paperless.serialisers import UserSerializer
 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
-from paperless_ai.matching import match_document_types_by_name
-from paperless_ai.matching import match_storage_paths_by_name
-from paperless_ai.matching import match_tags_by_name
-from paperless_ai.matching import resolve_correspondent_ids
-from paperless_ai.matching import resolve_document_type_ids
-from paperless_ai.matching import resolve_storage_path_ids
-from paperless_ai.matching import resolve_tag_ids
 from paperless_mail.models import MailAccount
 from paperless_mail.models import MailRule
 from paperless_mail.oauth import PaperlessMailOAuth2Manager
@@ -1022,83 +1010,71 @@ class DocumentViewSet(
         "custom_field_",
     )
 
-    # Matching document ids are looked up in the link tables in batches of this
-    # size: one literal IN (...) over every match is a multi-MB statement per
-    # relation and, on SQLite, exceeds the bound-parameter limit.
-    selection_data_link_batch_size = 10_000
-
     def _get_selection_data_for_queryset(self, queryset):
-        # The filtered queryset is evaluated once, as the page's own filter is,
-        # and never wrapped in a subquery: re-deriving a join or a HAVING
-        # (shared_by__id) per statement is very slow on some databases, and
-        # regrouping the queryset by another column would miscount.
-        # order_by() drops the default/user ordering, which only costs a sort.
-        correspondents: Counter = Counter()
-        document_types: Counter = Counter()
-        storage_paths: Counter = Counter()
-        document_ids: list[int] = []
-        seen: set[int] = set()
-        rows = queryset.order_by().values_list(
-            "pk",
-            "correspondent_id",
-            "document_type_id",
-            "storage_path_id",
+        # Resolve once instead of once per model below. `queryset` can carry an
+        # arbitrarily expensive WHERE clause (user filters plus the permission
+        # filter); re-embedding it as a subquery inside 5 separate Count(...)
+        # calls forces the database to re-evaluate that whole thing 5 times, and
+        # -- for FK relations especially -- can defeat semi-join planning
+        # entirely at scale. A concrete id list is cheap to reuse.
+        # order_by() drops the default/user ordering -- irrelevant for a plain
+        # id list, but left in place it forces a sort over the full filtered
+        # set before the ids can even be collected.
+        document_ids = list(queryset.order_by().values_list("pk", flat=True))
+
+        correspondents = Correspondent.objects.annotate(
+            document_count=Count(
+                "documents",
+                filter=Q(documents__id__in=document_ids),
+                distinct=True,
+            ),
+        )
+        document_types = DocumentType.objects.annotate(
+            document_count=Count(
+                "documents",
+                filter=Q(documents__id__in=document_ids),
+                distinct=True,
+            ),
+        )
+        storage_paths = StoragePath.objects.annotate(
+            document_count=Count(
+                "documents",
+                filter=Q(documents__id__in=document_ids),
+                distinct=True,
+            ),
+        )
+        # Tag and CustomField reach Document through an M2M/through-model table;
+        # a plain Count(filter=...) there is a much more expensive plan than the
+        # FK relations above once the bridge table is large -- see
+        # annotate_document_count_by_ids() for why.
+        tags = annotate_document_count_by_ids(
+            Tag.objects.all(),
+            through_model=Document.tags.through,
+            related_object_field="tag_id",
+            document_ids=document_ids,
+        )
+        custom_fields = annotate_document_count_by_ids(
+            CustomField.objects.all(),
+            through_model=CustomFieldInstance,
+            related_object_field="field_id",
+            document_ids=document_ids,
         )
-        for pk, correspondent, document_type, storage_path in rows:
-            # A filter can join, repeating a document row.
-            if pk in seen:
-                continue
-            seen.add(pk)
-            document_ids.append(pk)
-            # Documents without a value count under None, which selection()
-            # never looks up.
-            correspondents[correspondent] += 1
-            document_types[document_type] += 1
-            storage_paths[storage_path] += 1
-
-        def link_table_counts(link_model, field_name: str) -> Counter:
-            result: Counter = Counter()
-            batch_size = self.selection_data_link_batch_size
-            for start in range(0, len(document_ids), batch_size):
-                result.update(
-                    dict(
-                        link_model.objects.filter(
-                            document_id__in=document_ids[start : start + batch_size],
-                        )
-                        .values_list(field_name)
-                        .annotate(document_count=Count("document_id")),
-                    ),
-                )
-            return result
-
-        def selection(model, counts: Counter) -> list[dict[str, int]]:
-            # Every object is listed, with 0 for those with no matching document.
-            return [
-                {"id": pk, "document_count": counts.get(pk, 0)}
-                for pk in model.objects.values_list("pk", flat=True)
-            ]
-
         return {
-            "selected_correspondents": selection(
-                Correspondent,
-                correspondents,
-            ),
-            "selected_tags": selection(
-                Tag,
-                link_table_counts(Document.tags.through, "tag_id"),
-            ),
-            "selected_document_types": selection(
-                DocumentType,
-                document_types,
-            ),
-            "selected_storage_paths": selection(
-                StoragePath,
-                storage_paths,
-            ),
-            "selected_custom_fields": selection(
-                CustomField,
-                link_table_counts(CustomFieldInstance, "field_id"),
-            ),
+            "selected_correspondents": [
+                {"id": t.id, "document_count": t.document_count} for t in correspondents
+            ],
+            "selected_tags": [
+                {"id": t.id, "document_count": t.document_count} for t in tags
+            ],
+            "selected_document_types": [
+                {"id": t.id, "document_count": t.document_count} for t in document_types
+            ],
+            "selected_storage_paths": [
+                {"id": t.id, "document_count": t.document_count} for t in storage_paths
+            ],
+            "selected_custom_fields": [
+                {"id": t.id, "document_count": t.document_count} for t in custom_fields
+            ],
         }
 
     @classmethod
@@ -1625,6 +1601,18 @@ class DocumentViewSet(
         if not ai_config.ai_enabled:
             return HttpResponseBadRequest("AI is required for this feature")
 
+        from paperless_ai.ai_classifier import get_ai_document_classification
+        from paperless_ai.ai_classifier import get_llm_output_language
+        from paperless_ai.matching import extract_unmatched_names
+        from paperless_ai.matching import match_correspondents_by_name
+        from paperless_ai.matching import match_document_types_by_name
+        from paperless_ai.matching import match_storage_paths_by_name
+        from paperless_ai.matching import match_tags_by_name
+        from paperless_ai.matching import resolve_correspondent_ids
+        from paperless_ai.matching import resolve_document_type_ids
+        from paperless_ai.matching import resolve_storage_path_ids
+        from paperless_ai.matching import resolve_tag_ids
+
         output_language = get_llm_output_language(
             ai_config=ai_config,
             user=request.user,
@@ -2458,6 +2446,9 @@ class ChatStreamingView(GenericAPIView[Any]):
             )
             unrestricted = user_is_unrestricted(request.user)
 
+        from paperless_ai.ai_classifier import get_llm_output_language
+        from paperless_ai.chat import stream_chat_with_documents
+
         output_language = get_llm_output_language(
             ai_config=ai_config,
             user=request.user,

src/locale/en_US/LC_MESSAGES/django.po

@@ -2,7 +2,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: paperless-ngx\n"
 "Report-Msgid-Bugs-To: \n"
-"POT-Creation-Date: 2026-10-29 17:25+0000\n"
+"POT-Creation-Date: 2026-10-30 22:26+0000\n"
 "PO-Revision-Date: 2022-02-17 04:17\n"
 "Last-Translator: \n"
 "Language-Team: English\n"
@@ -21,43 +21,43 @@ msgstr ""
 msgid "Documents"
 msgstr ""
 
-#: documents/filters.py:337
+#: documents/filters.py:336
 msgid "Expected a date or a relative date such as \"today\" or \"in 30 days\"."
 msgstr ""
 
-#: documents/filters.py:513
+#: documents/filters.py:514
 msgid "Value must be valid JSON."
 msgstr ""
 
-#: documents/filters.py:534
+#: documents/filters.py:535
 msgid "Invalid custom field query expression"
 msgstr ""
 
-#: documents/filters.py:544
+#: documents/filters.py:545
 msgid "Invalid expression list. Must be nonempty."
 msgstr ""
 
-#: documents/filters.py:565
+#: documents/filters.py:566
 msgid "Invalid logical operator {op!r}"
 msgstr ""
 
-#: documents/filters.py:579
+#: documents/filters.py:580
 msgid "Maximum number of query conditions exceeded."
 msgstr ""
 
-#: documents/filters.py:643
+#: documents/filters.py:644
 msgid "{name!r} is not a valid custom field."
 msgstr ""
 
-#: documents/filters.py:680
+#: documents/filters.py:681
 msgid "{data_type} does not support query expr {expr!r}."
 msgstr ""
 
-#: documents/filters.py:801 documents/models.py:137
+#: documents/filters.py:802 documents/models.py:137
 msgid "Maximum nesting depth exceeded."
 msgstr ""
 
-#: documents/filters.py:1178
+#: documents/filters.py:1179
 msgid "Custom field not found"
 msgstr ""
 
@@ -1657,7 +1657,7 @@ msgid "workflow runs"
 msgstr ""
 
 #: documents/serialisers.py:483 documents/serialisers.py:840
-#: documents/serialisers.py:2872 documents/views.py:360 documents/views.py:2730
+#: documents/serialisers.py:2872 documents/views.py:348 documents/views.py:2733
 #: paperless_mail/serialisers.py:156
 msgid "Insufficient permissions."
 msgstr ""
@@ -1698,7 +1698,7 @@ msgstr ""
 msgid "Duplicate document identifiers are not allowed."
 msgstr ""
 
-#: documents/serialisers.py:2958 documents/views.py:4780
+#: documents/serialisers.py:2958 documents/views.py:4783
 #, python-format
 msgid "Documents not found: %(ids)s"
 msgstr ""
@@ -1949,11 +1949,11 @@ msgstr ""
 msgid ", "
 msgstr ""
 
-#: documents/views.py:315
+#: documents/views.py:303
 msgid "Specify only one of text, title_search, query, or more_like_id."
 msgstr ""
 
-#: documents/views.py:353 documents/views.py:2727
+#: documents/views.py:341 documents/views.py:2730
 msgid "Invalid more_like_id"
 msgstr ""
 
@@ -1969,20 +1969,20 @@ msgstr ""
 msgid "AI backend rejected the request. Check logs for details."
 msgstr ""
 
-#: documents/views.py:4796
+#: documents/views.py:4799
 #, python-format
 msgid "Insufficient permissions to share document %(id)s."
 msgstr ""
 
-#: documents/views.py:4843
+#: documents/views.py:4846
 msgid "Bundle is already being processed."
 msgstr ""
 
-#: documents/views.py:4909
+#: documents/views.py:4912
 msgid "The share link bundle is still being prepared. Please try again later."
 msgstr ""
 
-#: documents/views.py:4923
+#: documents/views.py:4926
 msgid "The share link bundle is unavailable."
 msgstr ""
 

src/paperless/asgi.py

@@ -17,7 +17,7 @@ from paperless.urls import websocket_urlpatterns  # noqa: E402
 
 application = ProtocolTypeRouter(
     {
-        "http": get_asgi_application(),
+        "http": django_asgi_app,
         "websocket": AuthMiddlewareStack(URLRouter(websocket_urlpatterns)),
     },
 )

src/paperless/parsers/mail.py

@@ -36,9 +36,6 @@ from humanize import naturalsize
 from imap_tools import MailAttachment
 from imap_tools import MailMessage
 from tika_client import TikaClient
-from turbohtml.clean import Linkify
-from turbohtml.clean import linkify
-from turbohtml.migration.bleach import clean
 
 from documents.parsers import ParseError
 from documents.parsers import make_thumbnail_from_pdf
@@ -622,6 +619,10 @@ class MailDocumentParser:
 
         def clean_html(text: str) -> str:
             """Attempt to clean, escape, and linkify the given HTML string."""
+            from turbohtml.clean import Linkify
+            from turbohtml.clean import linkify
+            from turbohtml.migration.bleach import clean
+
             if isinstance(text, list):
                 text = "\n".join([str(e) for e in text])
             if not isinstance(text, str):

src/paperless/parsers/tesseract.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import functools
 import importlib.resources
 import logging
 import os
@@ -39,9 +40,12 @@ if TYPE_CHECKING:
 
 logger = logging.getLogger("paperless.parsing.tesseract")
 
-_SRGB_ICC_DATA: Final[bytes] = (
-    importlib.resources.files("ocrmypdf.data").joinpath("sRGB.icc").read_bytes()
-)
+
+@functools.cache
+def _srgb_icc_data() -> bytes:
+    # Read on first use, importing ocrmypdf.data pulls in the whole ocrmypdf package
+    return importlib.resources.files("ocrmypdf.data").joinpath("sRGB.icc").read_bytes()
+
 
 _SUPPORTED_MIME_TYPES: Final[dict[str, str]] = {
     "application/pdf": ".pdf",
@@ -403,7 +407,7 @@ class RasterisedDocumentParser:
         pdfa_path = Path(self.tempdir) / "archive.pdf"
         try:
             with pikepdf.open(plain_pdf_path) as pdf:
-                cs = pdf.make_stream(_SRGB_ICC_DATA)
+                cs = pdf.make_stream(_srgb_icc_data())
                 cs["/N"] = 3
                 output_intent = pikepdf.Dictionary(
                     Type=pikepdf.Name("/OutputIntent"),

src/paperless/settings/custom.py

@@ -6,7 +6,6 @@ from pathlib import Path
 from typing import Any
 
 from celery.schedules import crontab
-from dateparser.languages.loader import LocaleDataLoader
 
 from paperless.settings.parsers import get_choice_from_env
 from paperless.settings.parsers import get_int_from_env
@@ -352,6 +351,8 @@ def parse_dateparser_languages(languages: str | None) -> list[str]:
             )
             language_list.append("zh")
 
+    from dateparser.languages.loader import LocaleDataLoader
+
     return list(LocaleDataLoader().get_locale_map(locales=language_list))
 
 

src/paperless/tests/test_utils.py

@@ -59,7 +59,10 @@ def test_ocr_to_dateparser_languages_exception(
             raise RuntimeError("Simulated error")
 
     with caplog.at_level(logging.WARNING):
-        monkeypatch.setattr(utils, "LocaleDataLoader", lambda: DummyLoader())
+        monkeypatch.setattr(
+            "dateparser.languages.loader.LocaleDataLoader",
+            lambda: DummyLoader(),
+        )
         result = utils.ocr_to_dateparser_languages("eng+fra")
         assert result == []
         assert (

src/paperless/utils.py

@@ -1,7 +1,5 @@
 import logging
 
-from dateparser.languages.loader import LocaleDataLoader
-
 logger = logging.getLogger("paperless.utils")
 
 OCR_TO_DATEPARSER_LANGUAGES = {
@@ -125,6 +123,8 @@ def ocr_to_dateparser_languages(ocr_languages: str) -> list[str]:
     - If a language cannot be mapped or validated, it is skipped with a warning.
     - Returns a list of valid locales, or an empty list if none could be converted.
     """
+    from dateparser.languages.loader import LocaleDataLoader
+
     loader = LocaleDataLoader()
     result = []
     try:

src/paperless/validators.py

@@ -3,7 +3,6 @@ from io import BytesIO
 from django.conf import settings
 from django.core.exceptions import ValidationError
 from django.core.files.uploadedfile import UploadedFile
-from lxml import etree
 from PIL import Image
 
 ALLOWED_SVG_TAGS: set[str] = {
@@ -186,6 +185,7 @@ def reject_dangerous_svg(file: UploadedFile) -> None:
     Raises ValidationError if unsafe content is found.
     See GHSA-6p53-hqqw-8j62
     """
+    from lxml import etree
 
     try:
         parser = etree.XMLParser(resolve_entities=False)

src/paperless/views.py

@@ -49,7 +49,6 @@ from paperless.serialisers import GroupSerializer
 from paperless.serialisers import PaperlessAuthTokenSerializer
 from paperless.serialisers import ProfileSerializer
 from paperless.serialisers import UserSerializer
-from paperless_ai.indexing import llm_index_exists
 
 
 class PaperlessObtainAuthTokenView(ObtainAuthToken):
@@ -471,6 +470,8 @@ class ApplicationConfigurationViewSet(ModelViewSet[ApplicationConfiguration]):
             or old_llm_embedding_model != new_llm_embedding_model
             or old_llm_context_size != new_llm_context_size
         )
+        from paperless_ai.indexing import llm_index_exists
+
         rebuild_needed = new_ai_index_enabled and (
             not llm_index_exists() or embedding_config_changed
         )