Fix tests

[ci skip]

CONTRIBUTING.md

@@ -2,11 +2,9 @@
 
 If you feel like contributing to the project, please do! Bug fixes and improvements are always welcome.
 
-⚠️ Please note: Pull requests that implement a new feature or enhancement _should almost always target an existing feature request_ with evidence of community interest and discussion. This is in order to balance the work of implementing and maintaining new features / enhancements. Pull requests that are opened without meeting this requirement may not be merged.
-
 If you want to implement something big:
 
-- As above, please start with a discussion! Maybe something similar is already in development and we can make it happen together.
+- Please start a discussion about that in the issues! Maybe something similar is already in development and we can make it happen together.
 - When making additions to the project, consider if the majority of users will benefit from your change. If not, you're probably better of forking the project.
 - Also consider if your change will get in the way of other users. A good change is a change that enhances the experience of some users who want that change and does not affect users who do not care about the change.
 - Please see the [paperless-ngx merge process](#merging-prs) below.

docs/configuration.md

@@ -1759,11 +1759,6 @@ started by the container.
 
 : Path to an image file in the /media/logo directory, must include 'logo', e.g. `/logo/Atari_logo.svg`
 
-!!! note
-
-    The logo file will be viewable by anyone with access to the Paperless instance login page,
-    so consider your choice of logo carefully and removing exif data from images before uploading.
-
 #### [`PAPERLESS_ENABLE_UPDATE_CHECK=<bool>`](#PAPERLESS_ENABLE_UPDATE_CHECK) {#PAPERLESS_ENABLE_UPDATE_CHECK}
 
 !!! note
@@ -1805,23 +1800,3 @@ password. All of these options come from their similarly-named [Django settings]
 #### [`PAPERLESS_EMAIL_USE_SSL=<bool>`](#PAPERLESS_EMAIL_USE_SSL) {#PAPERLESS_EMAIL_USE_SSL}
 
 : Defaults to false.
-
-## Remote OCR
-
-#### [`PAPERLESS_REMOTE_OCR_ENGINE=<str>`](#PAPERLESS_REMOTE_OCR_ENGINE) {#PAPERLESS_REMOTE_OCR_ENGINE}
-
-: The remote OCR engine to use. Currently only Azure AI is supported as "azureai".
-
-    Defaults to None, which disables remote OCR.
-
-#### [`PAPERLESS_REMOTE_OCR_API_KEY=<str>`](#PAPERLESS_REMOTE_OCR_API_KEY) {#PAPERLESS_REMOTE_OCR_API_KEY}
-
-: The API key to use for the remote OCR engine.
-
-    Defaults to None.
-
-#### [`PAPERLESS_REMOTE_OCR_ENDPOINT=<str>`](#PAPERLESS_REMOTE_OCR_ENDPOINT) {#PAPERLESS_REMOTE_OCR_ENDPOINT}
-
-: The endpoint to use for the remote OCR engine. This is required for Azure AI.
-
-    Defaults to None.

docs/index.md

@@ -25,10 +25,9 @@ physical documents into a searchable online archive so you can keep, well, _less
 ## Features
 
 -   **Organize and index** your scanned documents with tags, correspondents, types, and more.
--   _Your_ data is stored locally on _your_ server and is never transmitted or shared in any way, unless you explicitly choose to do so.
+-   _Your_ data is stored locally on _your_ server and is never transmitted or shared in any way.
 -   Performs **OCR** on your documents, adding searchable and selectable text, even to documents scanned with only images.
-    -   Utilizes the open-source Tesseract engine to recognize more than 100 languages.
-    -   _New!_ Supports remote OCR with Azure AI (opt-in).
+-   Utilizes the open-source Tesseract engine to recognize more than 100 languages.
 -   Documents are saved as PDF/A format which is designed for long term storage, alongside the unaltered originals.
 -   Uses machine-learning to automatically add tags, correspondents and document types to your documents.
 -   Supports PDF documents, images, plain text files, Office documents (Word, Excel, PowerPoint, and LibreOffice equivalents)[^1] and more.

docs/usage.md

@@ -878,21 +878,6 @@ how regularly you intend to scan documents and use paperless.
     performed the task associated with the document, move it to the
     inbox.
 
-## Remote OCR
-
-!!! important
-
-    This feature is disabled by default and will always remain strictly "opt-in".
-
-Paperless-ngx supports performing OCR on documents using remote services. At the moment, this is limited to
-[Microsoft's Azure "Document Intelligence" service](https://azure.microsoft.com/en-us/products/ai-services/ai-document-intelligence).
-This is of course a paid service (with a free tier) which requires an Azure account and subscription. Azure AI is not affiliated with
-Paperless-ngx in any way. When enabled, Paperless-ngx will automatically send appropriate documents to Azure for OCR processing, bypassing
-the local OCR engine. See the [configuration](configuration.md#PAPERLESS_REMOTE_OCR_ENGINE) options for more details.
-
-Additionally, when using a commercial service with this feature, consider both potential costs as well as any associated file size
-or page limitations (e.g. with a free tier).
-
 ## Architecture
 
 Paperless-ngx consists of the following components:

pyproject.toml

@@ -15,7 +15,6 @@ classifiers = [
 # This will allow testing to not install a webserver, mysql, etc
 
 dependencies = [
-  "azure-ai-documentintelligence>=1.0.2",
   "babel>=2.17",
   "bleach~=6.2.0",
   "celery[redis]~=5.5.1",
@@ -234,7 +233,6 @@ testpaths = [
   "src/paperless_tesseract/tests/",
   "src/paperless_tika/tests",
   "src/paperless_text/tests/",
-  "src/paperless_remote/tests/",
 ]
 addopts = [
   "--pythonwarnings=all",

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

@@ -1,7 +1,30 @@
 <pngx-page-header [(title)]="title">
+
+  @if (document?.versions?.length > 0) {
+    <div class="btn-group" ngbDropdown role="group">
+      <div class="btn-group" ngbDropdown role="group">
+        <button class="btn btn-sm btn-outline-secondary dropdown-toggle" ngbDropdownToggle [disabled]="!hasVersions">
+          <i-bs name="layers"></i-bs>
+          <span class="d-none d-lg-inline ps-1" i18n>Version</span>
+        </button>
+        <div class="dropdown-menu shadow" ngbDropdownMenu>
+          @for (vid of document.versions; track vid) {
+            <button ngbDropdownItem (click)="selectVersion(vid)">
+              <span i18n>Version</span> {{vid}}
+              @if (selectedVersionId === vid) { <span>&nbsp;✓</span> }
+            </button>
+          }
+        </div>
+      </div>
+      <input #versionFileInput type="file" class="visually-hidden" (change)="onVersionFileSelected($event)" />
+      <button class="btn btn-sm btn-outline-secondary" title="Upload new version" i18n-title (click)="versionFileInput.click()" [disabled]="!userIsOwner || !userCanEdit">
+        <i-bs name="file-earmark-plus"></i-bs><span class="visually-hidden" i18n>Upload new version</span>
+      </button>
+    </div>
+  }
   @if (archiveContentRenderType === ContentRenderType.PDF && !useNativePdfViewer) {
     @if (previewNumPages) {
-      <div class="input-group input-group-sm d-none d-md-flex">
+      <div class="input-group input-group-sm ms-2 d-none d-md-flex">
         <div class="input-group-text" i18n>Page</div>
           <input class="form-control flex-grow-0 w-auto" type="number" min="1" [max]="previewNumPages" [(ngModel)]="previewCurrentPage" />
         <div class="input-group-text" i18n>of {{previewNumPages}}</div>

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

@@ -222,6 +222,8 @@ export class DocumentDetailComponent
   titleSubject: Subject<string> = new Subject()
   previewUrl: string
   thumbUrl: string
+  // Versioning: which document ID to use for file preview/download
+  selectedVersionId: number
   previewText: string
   previewLoaded: boolean = false
   tiffURL: string
@@ -270,6 +272,7 @@ export class DocumentDetailComponent
   public readonly DataType = DataType
 
   @ViewChild('nav') nav: NgbNav
+  @ViewChild('versionFileInput') versionFileInput
   @ViewChild('pdfPreview') set pdfPreview(element) {
     // this gets called when component added or removed from DOM
     if (
@@ -402,7 +405,10 @@ export class DocumentDetailComponent
   }
 
   private loadDocument(documentId: number): void {
-    this.previewUrl = this.documentsService.getPreviewUrl(documentId)
+    this.selectedVersionId = documentId
+    this.previewUrl = this.documentsService.getPreviewUrl(
+      this.selectedVersionId
+    )
     this.http
       .get(this.previewUrl, { responseType: 'text' })
       .pipe(
@@ -417,7 +423,7 @@ export class DocumentDetailComponent
             err.message ?? err.toString()
           }`),
       })
-    this.thumbUrl = this.documentsService.getThumbUrl(documentId)
+    this.thumbUrl = this.documentsService.getThumbUrl(this.selectedVersionId)
     this.documentsService
       .get(documentId)
       .pipe(
@@ -638,6 +644,10 @@ export class DocumentDetailComponent
 
   updateComponent(doc: Document) {
     this.document = doc
+    // Default selected version is the newest version
+    this.selectedVersionId = doc.versions?.length
+      ? Math.max(...doc.versions)
+      : doc.id
     this.requiresPassword = false
     this.updateFormForCustomFields()
     if (this.archiveContentRenderType === ContentRenderType.TIFF) {
@@ -702,6 +712,36 @@ export class DocumentDetailComponent
     this.prepareForm(doc)
   }
 
+  get hasVersions(): boolean {
+    return this.document?.versions?.length > 1
+  }
+
+  // Update file preview and download target to a specific version (by document id)
+  selectVersion(versionId: number) {
+    this.selectedVersionId = versionId
+    this.previewUrl = this.documentsService.getPreviewUrl(
+      this.documentId,
+      false,
+      this.selectedVersionId
+    )
+    this.thumbUrl = this.documentsService.getThumbUrl(this.selectedVersionId)
+    // For text previews, refresh content
+    this.http
+      .get(this.previewUrl, { responseType: 'text' })
+      .pipe(
+        first(),
+        takeUntil(this.unsubscribeNotifier),
+        takeUntil(this.docChangeNotifier)
+      )
+      .subscribe({
+        next: (res) => (this.previewText = res.toString()),
+        error: (err) =>
+          (this.previewText = $localize`An error occurred loading content: ${
+            err.message ?? err.toString()
+          }`),
+      })
+  }
+
   get customFieldFormFields(): FormArray {
     return this.documentForm.get('custom_fields') as FormArray
   }
@@ -1049,10 +1089,36 @@ export class DocumentDetailComponent
     })
   }
 
+  onVersionFileSelected(event: Event) {
+    const input = event.target as HTMLInputElement
+    if (!input?.files || input.files.length === 0) return
+    const file = input.files[0]
+    // Reset input to allow re-selection of the same file later
+    input.value = ''
+    this.documentsService
+      .uploadVersion(this.documentId, file)
+      .pipe(first())
+      .subscribe({
+        next: () => {
+          this.toastService.showInfo(
+            $localize`Uploading new version. Processing will happen in the background.`
+          )
+          // Refresh metadata to reflect that versions changed (when ready)
+          this.openDocumentService.refreshDocument(this.documentId)
+        },
+        error: (error) => {
+          this.toastService.showError(
+            $localize`Error uploading new version`,
+            error
+          )
+        },
+      })
+  }
+
   download(original: boolean = false) {
     this.downloading = true
     const downloadUrl = this.documentsService.getDownloadUrl(
-      this.documentId,
+      this.selectedVersionId || this.documentId,
       original
     )
     this.http

src-ui/src/app/components/manage/tag-list/tag-list.component.spec.ts

@@ -71,20 +71,4 @@ describe('TagListComponent', () => {
       'Do you really want to delete the tag "Tag1"?'
     )
   })
-
-  it('should filter out child tags if name filter is empty, otherwise show all', () => {
-    const tags = [
-      { id: 1, name: 'Tag1', parent: null },
-      { id: 2, name: 'Tag2', parent: 1 },
-      { id: 3, name: 'Tag3', parent: null },
-    ]
-    component['_nameFilter'] = null // Simulate empty name filter
-    const filtered = component.filterData(tags as any)
-    expect(filtered.length).toBe(2)
-    expect(filtered.find((t) => t.id === 2)).toBeUndefined()
-
-    component['_nameFilter'] = 'Tag2' // Simulate non-empty name filter
-    const filteredWithName = component.filterData(tags as any)
-    expect(filteredWithName.length).toBe(3)
-  })
 })

src-ui/src/app/components/manage/tag-list/tag-list.component.ts

@@ -62,8 +62,6 @@ export class TagListComponent extends ManagementListComponent<Tag> {
   }
 
   filterData(data: Tag[]) {
-    return this.nameFilter?.length
-      ? [...data]
-      : data.filter((tag) => !tag.parent)
+    return data.filter((tag) => !tag.parent)
   }
 }

src-ui/src/app/data/document.ts

@@ -159,6 +159,10 @@ export interface Document extends ObjectWithPermissions {
 
   page_count?: number
 
+  // Versioning
+  head_version?: number
+  versions?: number[]
+
   // Frontend only
   __changedFields?: string[]
 }

src-ui/src/app/services/rest/document.service.ts

@@ -163,12 +163,19 @@ export class DocumentService extends AbstractPaperlessService<Document> {
     })
   }
 
-  getPreviewUrl(id: number, original: boolean = false): string {
+  getPreviewUrl(
+    id: number,
+    original: boolean = false,
+    versionID: number = null
+  ): string {
     let url = new URL(this.getResourceUrl(id, 'preview'))
     if (this._searchQuery) url.hash = `#search="${this.searchQuery}"`
     if (original) {
       url.searchParams.append('original', 'true')
     }
+    if (versionID) {
+      url.searchParams.append('version', versionID.toString())
+    }
     return url.toString()
   }
 
@@ -184,6 +191,16 @@ export class DocumentService extends AbstractPaperlessService<Document> {
     return url
   }
 
+  uploadVersion(documentId: number, file: File) {
+    const formData = new FormData()
+    formData.append('document', file, file.name)
+    return this.http.post(
+      this.getResourceUrl(documentId, 'update_version'),
+      formData,
+      { reportProgress: true, observe: 'events' }
+    )
+  }
+
   getNextAsn(): Observable<number> {
     return this.http.get<number>(this.getResourceUrl(null, 'next_asn'))
   }

src-ui/src/main.ts

@@ -78,6 +78,7 @@ import {
   fileEarmarkFill,
   fileEarmarkLock,
   fileEarmarkMinus,
+  fileEarmarkPlus,
   fileEarmarkRichtext,
   fileText,
   files,
@@ -94,6 +95,7 @@ import {
   house,
   infoCircle,
   journals,
+  layers,
   link,
   listNested,
   listTask,
@@ -290,6 +292,7 @@ const icons = {
   fileEarmarkFill,
   fileEarmarkLock,
   fileEarmarkMinus,
+  fileEarmarkPlus,
   fileEarmarkRichtext,
   files,
   fileText,
@@ -306,6 +309,7 @@ const icons = {
   house,
   infoCircle,
   journals,
+  layers,
   link,
   listNested,
   listTask,

src/documents/bulk_edit.py

@@ -1,6 +1,5 @@
 from __future__ import annotations
 
-import hashlib
 import logging
 import tempfile
 from pathlib import Path
@@ -333,10 +332,8 @@ def rotate(doc_ids: list[int], degrees: int) -> Literal["OK"]:
         f"Attempting to rotate {len(doc_ids)} documents by {degrees} degrees.",
     )
     qs = Document.objects.filter(id__in=doc_ids)
-    affected_docs: list[int] = []
     import pikepdf
 
-    rotate_tasks = []
     for doc in qs:
         if doc.mime_type != "application/pdf":
             logger.warning(
@@ -344,28 +341,34 @@ def rotate(doc_ids: list[int], degrees: int) -> Literal["OK"]:
             )
             continue
         try:
-            with pikepdf.open(doc.source_path, allow_overwriting_input=True) as pdf:
+            # Write rotated output to a temp file and create a new version via consume pipeline
+            filepath: Path = (
+                Path(tempfile.mkdtemp(dir=settings.SCRATCH_DIR))
+                / f"{doc.id}_rotated.pdf"
+            )
+            with pikepdf.open(doc.source_path) as pdf:
                 for page in pdf.pages:
                     page.rotate(degrees, relative=True)
-                pdf.save()
-                doc.checksum = hashlib.md5(doc.source_path.read_bytes()).hexdigest()
-                doc.save()
-                rotate_tasks.append(
-                    update_document_content_maybe_archive_file.s(
-                        document_id=doc.id,
-                    ),
-                )
-                logger.info(
-                    f"Rotated document {doc.id} by {degrees} degrees",
-                )
-                affected_docs.append(doc.id)
+                pdf.remove_unreferenced_resources()
+                pdf.save(filepath)
+
+            # Preserve metadata/permissions via overrides; mark as new version
+            overrides = DocumentMetadataOverrides().from_document(doc)
+
+            consume_file.delay(
+                ConsumableDocument(
+                    source=DocumentSource.ConsumeFolder,
+                    original_file=filepath,
+                    head_version_id=doc.id,
+                ),
+                overrides,
+            )
+            logger.info(
+                f"Queued new rotated version for document {doc.id} by {degrees} degrees",
+            )
         except Exception as e:
             logger.exception(f"Error rotating document {doc.id}: {e}")
 
-    if len(affected_docs) > 0:
-        bulk_update_task = bulk_update_documents.si(document_ids=affected_docs)
-        chord(header=rotate_tasks, body=bulk_update_task).delay()
-
     return "OK"
 
 
@@ -528,19 +531,31 @@ def delete_pages(doc_ids: list[int], pages: list[int]) -> Literal["OK"]:
     import pikepdf
 
     try:
-        with pikepdf.open(doc.source_path, allow_overwriting_input=True) as pdf:
+        # Produce edited PDF to a temp file and create a new version
+        filepath: Path = (
+            Path(tempfile.mkdtemp(dir=settings.SCRATCH_DIR))
+            / f"{doc.id}_pages_deleted.pdf"
+        )
+        with pikepdf.open(doc.source_path) as pdf:
             offset = 1  # pages are 1-indexed
             for page_num in pages:
                 pdf.pages.remove(pdf.pages[page_num - offset])
                 offset += 1  # remove() changes the index of the pages
             pdf.remove_unreferenced_resources()
-            pdf.save()
-            doc.checksum = hashlib.md5(doc.source_path.read_bytes()).hexdigest()
-            if doc.page_count is not None:
-                doc.page_count = doc.page_count - len(pages)
-            doc.save()
-            update_document_content_maybe_archive_file.delay(document_id=doc.id)
-            logger.info(f"Deleted pages {pages} from document {doc.id}")
+            pdf.save(filepath)
+
+        overrides = DocumentMetadataOverrides().from_document(doc)
+        consume_file.delay(
+            ConsumableDocument(
+                source=DocumentSource.ConsumeFolder,
+                original_file=filepath,
+                head_version_id=doc.id,
+            ),
+            overrides,
+        )
+        logger.info(
+            f"Queued new version for document {doc.id} after deleting pages {pages}",
+        )
     except Exception as e:
         logger.exception(f"Error deleting pages from document {doc.id}: {e}")
 
@@ -592,17 +607,29 @@ def edit_pdf(
                     dst.pages[-1].rotate(op["rotate"], relative=True)
 
         if update_document:
-            temp_path = doc.source_path.with_suffix(".tmp.pdf")
+            # Create a new version from the edited PDF rather than replacing in-place
             pdf = pdf_docs[0]
             pdf.remove_unreferenced_resources()
-            # save the edited PDF to a temporary file in case of errors
-            pdf.save(temp_path)
-            # replace the original document with the edited one
-            temp_path.replace(doc.source_path)
-            doc.checksum = hashlib.md5(doc.source_path.read_bytes()).hexdigest()
-            doc.page_count = len(pdf.pages)
-            doc.save()
-            update_document_content_maybe_archive_file.delay(document_id=doc.id)
+            filepath: Path = (
+                Path(tempfile.mkdtemp(dir=settings.SCRATCH_DIR))
+                / f"{doc.id}_edited.pdf"
+            )
+            pdf.save(filepath)
+            overrides = (
+                DocumentMetadataOverrides().from_document(doc)
+                if include_metadata
+                else DocumentMetadataOverrides()
+            )
+            if user is not None:
+                overrides.owner_id = user.id
+            consume_file.delay(
+                ConsumableDocument(
+                    source=DocumentSource.ConsumeFolder,
+                    original_file=filepath,
+                    head_version_id=doc.id,
+                ),
+                overrides,
+            )
         else:
             consume_tasks = []
             overrides = (

src/documents/conditionals.py

@@ -14,6 +14,51 @@ from documents.classifier import DocumentClassifier
 from documents.models import Document
 
 
+def _resolve_effective_doc(pk: int, request) -> Document | None:
+    """
+    Resolve which Document row should be considered for caching keys:
+    - If a version is requested, use that version
+    - If pk is a head doc, use its newest child version if present, else the head.
+    - Else, pk is a version, use that version.
+    Returns None if resolution fails (treat as no-cache).
+    """
+    try:
+        request_doc = Document.objects.only("id", "head_version_id").get(pk=pk)
+    except Document.DoesNotExist:
+        return None
+
+    head_doc = (
+        request_doc
+        if request_doc.head_version_id is None
+        else Document.objects.only("id").get(id=request_doc.head_version_id)
+    )
+
+    version_param = (
+        request.query_params.get("version")
+        if hasattr(request, "query_params")
+        else None
+    )
+    if version_param:
+        try:
+            version_id = int(version_param)
+            candidate = Document.objects.only("id", "head_version_id").get(
+                id=version_id,
+            )
+            if candidate.id != head_doc.id and candidate.head_version_id != head_doc.id:
+                return None
+            return candidate
+        except Exception:
+            return None
+
+    # Default behavior: if pk is a head doc, prefer its newest child version
+    if request_doc.head_version_id is None:
+        latest = head_doc.versions.only("id").order_by("id").last()
+        return latest or head_doc
+
+    # pk is already a version
+    return request_doc
+
+
 def suggestions_etag(request, pk: int) -> str | None:
     """
     Returns an optional string for the ETag, allowing browser caching of
@@ -71,11 +116,10 @@ def metadata_etag(request, pk: int) -> str | None:
     Metadata is extracted from the original file, so use its checksum as the
     ETag
     """
-    try:
-        doc = Document.objects.only("checksum").get(pk=pk)
-        return doc.checksum
-    except Document.DoesNotExist:  # pragma: no cover
+    doc = _resolve_effective_doc(pk, request)
+    if doc is None:
         return None
+    return doc.checksum
     return None
 
 
@@ -85,11 +129,10 @@ def metadata_last_modified(request, pk: int) -> datetime | None:
     not the modification of the original file, but of the database object, but might as well
     error on the side of more cautious
     """
-    try:
-        doc = Document.objects.only("modified").get(pk=pk)
-        return doc.modified
-    except Document.DoesNotExist:  # pragma: no cover
+    doc = _resolve_effective_doc(pk, request)
+    if doc is None:
         return None
+    return doc.modified
     return None
 
 
@@ -97,15 +140,15 @@ def preview_etag(request, pk: int) -> str | None:
     """
     ETag for the document preview, using the original or archive checksum, depending on the request
     """
-    try:
-        doc = Document.objects.only("checksum", "archive_checksum").get(pk=pk)
-        use_original = (
-            "original" in request.query_params
-            and request.query_params["original"] == "true"
-        )
-        return doc.checksum if use_original else doc.archive_checksum
-    except Document.DoesNotExist:  # pragma: no cover
+    doc = _resolve_effective_doc(pk, request)
+    if doc is None:
         return None
+    use_original = (
+        hasattr(request, "query_params")
+        and "original" in request.query_params
+        and request.query_params["original"] == "true"
+    )
+    return doc.checksum if use_original else doc.archive_checksum
     return None
 
 
@@ -114,11 +157,10 @@ def preview_last_modified(request, pk: int) -> datetime | None:
     Uses the documents modified time to set the Last-Modified header.  Not strictly
     speaking correct, but close enough and quick
     """
-    try:
-        doc = Document.objects.only("modified").get(pk=pk)
-        return doc.modified
-    except Document.DoesNotExist:  # pragma: no cover
+    doc = _resolve_effective_doc(pk, request)
+    if doc is None:
         return None
+    return doc.modified
     return None
 
 
@@ -128,10 +170,13 @@ def thumbnail_last_modified(request, pk: int) -> datetime | None:
     Cache should be (slightly?) faster than filesystem
     """
     try:
-        doc = Document.objects.only("storage_type").get(pk=pk)
+        doc = _resolve_effective_doc(pk, request)
+        if doc is None:
+            return None
         if not doc.thumbnail_path.exists():
             return None
-        doc_key = get_thumbnail_modified_key(pk)
+        # Use the effective document id for cache key
+        doc_key = get_thumbnail_modified_key(doc.id)
 
         cache_hit = cache.get(doc_key)
         if cache_hit is not None:

src/documents/consumer.py

@@ -113,6 +113,12 @@ class ConsumerPluginMixin:
 
         self.filename = self.metadata.filename or self.input_doc.original_file.name
 
+        if input_doc.head_version_id:
+            self.log.debug(f"Document head version id: {input_doc.head_version_id}")
+            head_version = Document.objects.get(pk=input_doc.head_version_id)
+            version_index = head_version.versions.count()
+            self.filename += f"_v{version_index}"
+
     def _send_progress(
         self,
         current_progress: int,
@@ -470,12 +476,44 @@ class ConsumerPlugin(
         try:
             with transaction.atomic():
                 # store the document.
-                document = self._store(
-                    text=text,
-                    date=date,
-                    page_count=page_count,
-                    mime_type=mime_type,
-                )
+                if self.input_doc.head_version_id:
+                    # If this is a new version of an existing document, we need
+                    # to make sure we're not creating a new document, but updating
+                    # the existing one.
+                    original_document = Document.objects.get(
+                        pk=self.input_doc.head_version_id,
+                    )
+                    self.log.debug("Saving record for updated version to database")
+                    original_document.pk = None
+                    original_document.head_version = Document.objects.get(
+                        pk=self.input_doc.head_version_id,
+                    )
+                    file_for_checksum = (
+                        self.unmodified_original
+                        if self.unmodified_original is not None
+                        else self.working_copy
+                    )
+                    original_document.checksum = hashlib.md5(
+                        file_for_checksum.read_bytes(),
+                    ).hexdigest()
+                    original_document.content = text
+                    original_document.page_count = page_count
+                    original_document.mime_type = mime_type
+                    original_document.original_filename = self.filename
+                    # Clear unique file path fields so they can be generated uniquely later
+                    original_document.filename = None
+                    original_document.archive_filename = None
+                    original_document.archive_checksum = None
+                    original_document.modified = timezone.now()
+                    original_document.save()
+                    document = original_document
+                else:
+                    document = self._store(
+                        text=text,
+                        date=date,
+                        page_count=page_count,
+                        mime_type=mime_type,
+                    )
 
                 # If we get here, it was successful. Proceed with post-consume
                 # hooks. If they fail, nothing will get changed.

src/documents/data_models.py

@@ -156,6 +156,7 @@ class ConsumableDocument:
 
     source: DocumentSource
     original_file: Path
+    head_version_id: int | None = None
     mailrule_id: int | None = None
     mime_type: str = dataclasses.field(init=False, default=None)
 

src/documents/management/commands/document_consumer.py

@@ -82,13 +82,6 @@ def _is_ignored(filepath: Path) -> bool:
 
 
 def _consume(filepath: Path) -> None:
-    # Check permissions early
-    try:
-        filepath.stat()
-    except (PermissionError, OSError):
-        logger.warning(f"Not consuming file {filepath}: Permission denied.")
-        return
-
     if filepath.is_dir() or _is_ignored(filepath):
         return
 
@@ -330,12 +323,7 @@ class Command(BaseCommand):
 
                         # Also make sure the file exists still, some scanners might write a
                         # temporary file first
-                        try:
-                            file_still_exists = filepath.exists() and filepath.is_file()
-                        except (PermissionError, OSError):  # pragma: no cover
-                            # If we can't check, let it fail in the _consume function
-                            file_still_exists = True
-                            continue
+                        file_still_exists = filepath.exists() and filepath.is_file()
 
                         if waited_long_enough and file_still_exists:
                             _consume(filepath)

src/documents/migrations/1072_document_head_version.py

@@ -0,0 +1,26 @@
+# Generated by Django 5.1.6 on 2025-02-26 17:08
+
+import django.db.models.deletion
+from django.db import migrations
+from django.db import models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("documents", "1071_tag_tn_ancestors_count_tag_tn_ancestors_pks_and_more"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="document",
+            name="head_version",
+            field=models.ForeignKey(
+                blank=True,
+                null=True,
+                on_delete=django.db.models.deletion.CASCADE,
+                related_name="versions",
+                to="documents.document",
+                verbose_name="head version of document",
+            ),
+        ),
+    ]

src/documents/models.py

@@ -313,6 +313,15 @@ class Document(SoftDeleteModel, ModelWithOwner):
         ),
     )
 
+    head_version = models.ForeignKey(
+        "self",
+        blank=True,
+        null=True,
+        related_name="versions",
+        on_delete=models.CASCADE,
+        verbose_name=_("head version of document"),
+    )
+
     class Meta:
         ordering = ("-created",)
         verbose_name = _("document")

src/documents/serialisers.py

@@ -974,6 +974,8 @@ class DocumentSerializer(
     page_count = SerializerMethodField()
 
     notes = NotesSerializer(many=True, required=False, read_only=True)
+    head_version = serializers.PrimaryKeyRelatedField(read_only=True)
+    versions = serializers.PrimaryKeyRelatedField(many=True, read_only=True)
 
     custom_fields = CustomFieldInstanceSerializer(
         many=True,
@@ -1016,6 +1018,10 @@ class DocumentSerializer(
             request.version if request else settings.REST_FRAMEWORK["DEFAULT_VERSION"],
         )
 
+        if doc.get("versions") is not None:
+            doc["versions"] = sorted(doc["versions"], reverse=True)
+            doc["versions"].append(doc["id"])
+
         if api_version < 9:
             # provide created as a datetime for backwards compatibility
             from django.utils import timezone
@@ -1184,6 +1190,8 @@ class DocumentSerializer(
             "remove_inbox_tags",
             "page_count",
             "mime_type",
+            "head_version",
+            "versions",
         )
         list_serializer_class = OwnedObjectListSerializer
 
@@ -1867,6 +1875,15 @@ class PostDocumentSerializer(serializers.Serializer):
             return created.date()
 
 
+class DocumentVersionSerializer(serializers.Serializer):
+    document = serializers.FileField(
+        label="Document",
+        write_only=True,
+    )
+
+    validate_document = PostDocumentSerializer().validate_document
+
+
 class BulkDownloadSerializer(DocumentListSerializer):
     content = serializers.ChoiceField(
         choices=["archive", "originals", "both"],

src/documents/tasks.py

@@ -145,13 +145,17 @@ def consume_file(
     if overrides is None:
         overrides = DocumentMetadataOverrides()
 
-    plugins: list[type[ConsumeTaskPlugin]] = [
-        ConsumerPreflightPlugin,
-        CollatePlugin,
-        BarcodePlugin,
-        WorkflowTriggerPlugin,
-        ConsumerPlugin,
-    ]
+    plugins: list[type[ConsumeTaskPlugin]] = (
+        [ConsumerPreflightPlugin, ConsumerPlugin]
+        if input_doc.head_version_id is not None
+        else [
+            ConsumerPreflightPlugin,
+            CollatePlugin,
+            BarcodePlugin,
+            WorkflowTriggerPlugin,
+            ConsumerPlugin,
+        ]
+    )
 
     with (
         ProgressManager(

src/documents/tests/test_bulk_edit.py

@@ -787,10 +787,8 @@ class TestPDFActions(DirectoriesMixin, TestCase):
 
         mock_consume_file.assert_not_called()
 
-    @mock.patch("documents.tasks.bulk_update_documents.si")
-    @mock.patch("documents.tasks.update_document_content_maybe_archive_file.s")
-    @mock.patch("celery.chord.delay")
-    def test_rotate(self, mock_chord, mock_update_document, mock_update_documents):
+    @mock.patch("documents.tasks.consume_file.delay")
+    def test_rotate(self, mock_consume_delay):
         """
         GIVEN:
             - Existing documents
@@ -801,19 +799,22 @@ class TestPDFActions(DirectoriesMixin, TestCase):
         """
         doc_ids = [self.doc1.id, self.doc2.id]
         result = bulk_edit.rotate(doc_ids, 90)
-        self.assertEqual(mock_update_document.call_count, 2)
-        mock_update_documents.assert_called_once()
-        mock_chord.assert_called_once()
+        self.assertEqual(mock_consume_delay.call_count, 2)
+        for call, expected_id in zip(
+            mock_consume_delay.call_args_list,
+            doc_ids,
+        ):
+            consumable, overrides = call.args
+            self.assertEqual(consumable.head_version_id, expected_id)
+            self.assertIsNotNone(overrides)
         self.assertEqual(result, "OK")
 
-    @mock.patch("documents.tasks.bulk_update_documents.si")
-    @mock.patch("documents.tasks.update_document_content_maybe_archive_file.s")
+    @mock.patch("documents.tasks.consume_file.delay")
     @mock.patch("pikepdf.Pdf.save")
     def test_rotate_with_error(
         self,
         mock_pdf_save,
-        mock_update_archive_file,
-        mock_update_documents,
+        mock_consume_delay,
     ):
         """
         GIVEN:
@@ -832,16 +833,12 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             error_str = cm.output[0]
             expected_str = "Error rotating document"
             self.assertIn(expected_str, error_str)
-            mock_update_archive_file.assert_not_called()
+            mock_consume_delay.assert_not_called()
 
-    @mock.patch("documents.tasks.bulk_update_documents.si")
-    @mock.patch("documents.tasks.update_document_content_maybe_archive_file.s")
-    @mock.patch("celery.chord.delay")
+    @mock.patch("documents.tasks.consume_file.delay")
     def test_rotate_non_pdf(
         self,
-        mock_chord,
-        mock_update_document,
-        mock_update_documents,
+        mock_consume_delay,
     ):
         """
         GIVEN:
@@ -856,14 +853,16 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             output_str = cm.output[1]
             expected_str = "Document 4 is not a PDF, skipping rotation"
             self.assertIn(expected_str, output_str)
-            self.assertEqual(mock_update_document.call_count, 1)
-            mock_update_documents.assert_called_once()
-            mock_chord.assert_called_once()
+            self.assertEqual(mock_consume_delay.call_count, 1)
+            consumable, overrides = mock_consume_delay.call_args[0]
+            self.assertEqual(consumable.head_version_id, self.doc2.id)
+            self.assertIsNotNone(overrides)
             self.assertEqual(result, "OK")
 
-    @mock.patch("documents.tasks.update_document_content_maybe_archive_file.delay")
+    @mock.patch("documents.tasks.consume_file.delay")
     @mock.patch("pikepdf.Pdf.save")
-    def test_delete_pages(self, mock_pdf_save, mock_update_archive_file):
+    @mock.patch("documents.data_models.magic.from_file", return_value="application/pdf")
+    def test_delete_pages(self, mock_magic, mock_pdf_save, mock_consume_delay):
         """
         GIVEN:
             - Existing documents
@@ -871,24 +870,22 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             - Delete pages action is called with 1 document and 2 pages
         THEN:
             - Save should be called once
-            - Archive file should be updated once
-            - The document's page_count should be reduced by the number of deleted pages
+            - A new version should be enqueued via consume_file
         """
         doc_ids = [self.doc2.id]
-        initial_page_count = self.doc2.page_count
         pages = [1, 3]
         result = bulk_edit.delete_pages(doc_ids, pages)
         mock_pdf_save.assert_called_once()
-        mock_update_archive_file.assert_called_once()
+        mock_consume_delay.assert_called_once()
+        consumable, overrides = mock_consume_delay.call_args[0]
+        self.assertEqual(consumable.head_version_id, self.doc2.id)
+        self.assertTrue(str(consumable.original_file).endswith("_pages_deleted.pdf"))
+        self.assertIsNotNone(overrides)
         self.assertEqual(result, "OK")
 
-        expected_page_count = initial_page_count - len(pages)
-        self.doc2.refresh_from_db()
-        self.assertEqual(self.doc2.page_count, expected_page_count)
-
-    @mock.patch("documents.tasks.update_document_content_maybe_archive_file.delay")
+    @mock.patch("documents.tasks.consume_file.delay")
     @mock.patch("pikepdf.Pdf.save")
-    def test_delete_pages_with_error(self, mock_pdf_save, mock_update_archive_file):
+    def test_delete_pages_with_error(self, mock_pdf_save, mock_consume_delay):
         """
         GIVEN:
             - Existing documents
@@ -897,7 +894,7 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             - PikePDF raises an error
         THEN:
             - Save should be called once
-            - Archive file should not be updated
+            - No new version should be enqueued
         """
         mock_pdf_save.side_effect = Exception("Error saving PDF")
         doc_ids = [self.doc2.id]
@@ -908,7 +905,7 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             error_str = cm.output[0]
             expected_str = "Error deleting pages from document"
             self.assertIn(expected_str, error_str)
-            mock_update_archive_file.assert_not_called()
+            mock_consume_delay.assert_not_called()
 
     @mock.patch("documents.bulk_edit.group")
     @mock.patch("documents.tasks.consume_file.s")
@@ -968,21 +965,18 @@ class TestPDFActions(DirectoriesMixin, TestCase):
         self.assertEqual(result, "OK")
         mock_chord.assert_called_once()
 
-    @mock.patch("documents.tasks.update_document_content_maybe_archive_file.delay")
-    def test_edit_pdf_with_update_document(self, mock_update_document):
+    @mock.patch("documents.tasks.consume_file.delay")
+    def test_edit_pdf_with_update_document(self, mock_consume_delay):
         """
         GIVEN:
             - A single existing PDF document
         WHEN:
             - edit_pdf is called with update_document=True and a single output
         THEN:
-            - The original document is updated in-place
-            - The update_document_content_maybe_archive_file task is triggered
+            - A version update is enqueued targeting the existing document
         """
         doc_ids = [self.doc2.id]
         operations = [{"page": 1}, {"page": 2}]
-        original_checksum = self.doc2.checksum
-        original_page_count = self.doc2.page_count
 
         result = bulk_edit.edit_pdf(
             doc_ids,
@@ -992,10 +986,11 @@ class TestPDFActions(DirectoriesMixin, TestCase):
         )
 
         self.assertEqual(result, "OK")
-        self.doc2.refresh_from_db()
-        self.assertNotEqual(self.doc2.checksum, original_checksum)
-        self.assertNotEqual(self.doc2.page_count, original_page_count)
-        mock_update_document.assert_called_once_with(document_id=self.doc2.id)
+        mock_consume_delay.assert_called_once()
+        consumable, overrides = mock_consume_delay.call_args[0]
+        self.assertEqual(consumable.head_version_id, self.doc2.id)
+        self.assertTrue(str(consumable.original_file).endswith("_edited.pdf"))
+        self.assertIsNotNone(overrides)
 
     @mock.patch("documents.bulk_edit.group")
     @mock.patch("documents.tasks.consume_file.s")

src/documents/tests/test_management_consumer.py

@@ -209,26 +209,6 @@ class TestConsumer(DirectoriesMixin, ConsumerThreadMixin, TransactionTestCase):
         # assert that we have an error logged with this invalid file.
         error_logger.assert_called_once()
 
-    @mock.patch("documents.management.commands.document_consumer.logger.warning")
-    def test_permission_error_on_prechecks(self, warning_logger):
-        filepath = Path(self.dirs.consumption_dir) / "selinux.txt"
-        filepath.touch()
-
-        original_stat = Path.stat
-
-        def raising_stat(self, *args, **kwargs):
-            if self == filepath:
-                raise PermissionError("Permission denied")
-            return original_stat(self, *args, **kwargs)
-
-        with mock.patch("pathlib.Path.stat", new=raising_stat):
-            document_consumer._consume(filepath)
-
-        warning_logger.assert_called_once()
-        (args, _) = warning_logger.call_args
-        self.assertIn("Permission denied", args[0])
-        self.consume_file_mock.assert_not_called()
-
     @override_settings(CONSUMPTION_DIR="does_not_exist")
     def test_consumption_directory_invalid(self):
         self.assertRaises(CommandError, call_command, "document_consumer", "--oneshot")

src/documents/views.py

@@ -147,6 +147,7 @@ from documents.serialisers import CustomFieldSerializer
 from documents.serialisers import DocumentListSerializer
 from documents.serialisers import DocumentSerializer
 from documents.serialisers import DocumentTypeSerializer
+from documents.serialisers import DocumentVersionSerializer
 from documents.serialisers import NotesSerializer
 from documents.serialisers import PostDocumentSerializer
 from documents.serialisers import RunTaskViewSerializer
@@ -567,7 +568,7 @@ class DocumentViewSet(
     GenericViewSet,
 ):
     model = Document
-    queryset = Document.objects.annotate(num_notes=Count("notes"))
+    queryset = Document.objects.all()
     serializer_class = DocumentSerializer
     pagination_class = StandardPagination
     permission_classes = (IsAuthenticated, PaperlessObjectPermissions)
@@ -596,7 +597,8 @@ class DocumentViewSet(
 
     def get_queryset(self):
         return (
-            Document.objects.distinct()
+            Document.objects.filter(head_version__isnull=True)
+            .distinct()
             .order_by("-created")
             .annotate(num_notes=Count("notes"))
             .select_related("correspondent", "storage_path", "document_type", "owner")
@@ -658,18 +660,55 @@ class DocumentViewSet(
             and request.query_params["original"] == "true"
         )
 
+    def _resolve_file_doc(self, head_doc: Document, request):
+        version_param = request.query_params.get("version")
+        if version_param:
+            try:
+                version_id = int(version_param)
+            except (TypeError, ValueError):
+                raise NotFound("Invalid version parameter")
+            try:
+                candidate = Document.global_objects.select_related("owner").get(
+                    id=version_id,
+                )
+            except Document.DoesNotExist:
+                raise Http404
+            if candidate.id != head_doc.id and candidate.head_version_id != head_doc.id:
+                raise Http404
+            return candidate
+        latest = head_doc.versions.order_by("id").last()
+        return latest or head_doc
+
     def file_response(self, pk, request, disposition):
-        doc = Document.global_objects.select_related("owner").get(id=pk)
+        request_doc = Document.global_objects.select_related("owner").get(id=pk)
+        head_doc = (
+            request_doc
+            if request_doc.head_version_id is None
+            else Document.global_objects.select_related("owner").get(
+                id=request_doc.head_version_id,
+            )
+        )
         if request.user is not None and not has_perms_owner_aware(
             request.user,
             "view_document",
-            doc,
+            head_doc,
         ):
             return HttpResponseForbidden("Insufficient permissions")
+        # If a version is explicitly requested, use it. Otherwise:
+        # - if pk is a head document: serve newest version
+        # - if pk is a version: serve that version
+        if "version" in request.query_params:
+            file_doc = self._resolve_file_doc(head_doc, request)
+        else:
+            file_doc = (
+                self._resolve_file_doc(head_doc, request)
+                if request_doc.head_version_id is None
+                else request_doc
+            )
         return serve_file(
-            doc=doc,
+            doc=file_doc,
             use_archive=not self.original_requested(request)
-            and doc.has_archive_version,
+            and file_doc.has_archive_version,
             disposition=disposition,
         )
 
@@ -704,16 +743,33 @@ class DocumentViewSet(
     )
     def metadata(self, request, pk=None):
         try:
-            doc = Document.objects.select_related("owner").get(pk=pk)
+            request_doc = Document.objects.select_related("owner").get(pk=pk)
+            head_doc = (
+                request_doc
+                if request_doc.head_version_id is None
+                else Document.objects.select_related("owner").get(
+                    id=request_doc.head_version_id,
+                )
+            )
             if request.user is not None and not has_perms_owner_aware(
                 request.user,
                 "view_document",
-                doc,
+                head_doc,
             ):
                 return HttpResponseForbidden("Insufficient permissions")
         except Document.DoesNotExist:
             raise Http404
 
+        # Choose the effective document (newest version by default, or explicit via ?version=)
+        if "version" in request.query_params:
+            doc = self._resolve_file_doc(head_doc, request)
+        else:
+            doc = (
+                self._resolve_file_doc(head_doc, request)
+                if request_doc.head_version_id is None
+                else request_doc
+            )
+
         document_cached_metadata = get_metadata_cache(doc.pk)
 
         archive_metadata = None
@@ -815,8 +871,36 @@ class DocumentViewSet(
     )
     def preview(self, request, pk=None):
         try:
-            response = self.file_response(pk, request, "inline")
-            return response
+            request_doc = Document.objects.select_related("owner").get(id=pk)
+            head_doc = (
+                request_doc
+                if request_doc.head_version_id is None
+                else Document.objects.select_related("owner").get(
+                    id=request_doc.head_version_id,
+                )
+            )
+            if request.user is not None and not has_perms_owner_aware(
+                request.user,
+                "view_document",
+                head_doc,
+            ):
+                return HttpResponseForbidden("Insufficient permissions")
+
+            if "version" in request.query_params:
+                file_doc = self._resolve_file_doc(head_doc, request)
+            else:
+                file_doc = (
+                    self._resolve_file_doc(head_doc, request)
+                    if request_doc.head_version_id is None
+                    else request_doc
+                )
+
+            return serve_file(
+                doc=file_doc,
+                use_archive=not self.original_requested(request)
+                and file_doc.has_archive_version,
+                disposition="inline",
+            )
         except (FileNotFoundError, Document.DoesNotExist):
             raise Http404
 
@@ -825,17 +909,32 @@ class DocumentViewSet(
     @method_decorator(last_modified(thumbnail_last_modified))
     def thumb(self, request, pk=None):
         try:
-            doc = Document.objects.select_related("owner").get(id=pk)
+            request_doc = Document.objects.select_related("owner").get(id=pk)
+            head_doc = (
+                request_doc
+                if request_doc.head_version_id is None
+                else Document.objects.select_related("owner").get(
+                    id=request_doc.head_version_id,
+                )
+            )
             if request.user is not None and not has_perms_owner_aware(
                 request.user,
                 "view_document",
-                doc,
+                head_doc,
             ):
                 return HttpResponseForbidden("Insufficient permissions")
-            if doc.storage_type == Document.STORAGE_TYPE_GPG:
-                handle = GnuPG.decrypted(doc.thumbnail_file)
+            if "version" in request.query_params:
+                file_doc = self._resolve_file_doc(head_doc, request)
             else:
-                handle = doc.thumbnail_file
+                file_doc = (
+                    self._resolve_file_doc(head_doc, request)
+                    if request_doc.head_version_id is None
+                    else request_doc
+                )
+            if file_doc.storage_type == Document.STORAGE_TYPE_GPG:
+                handle = GnuPG.decrypted(file_doc.thumbnail_file)
+            else:
+                handle = file_doc.thumbnail_file
 
             return HttpResponse(handle, content_type="image/webp")
         except (FileNotFoundError, Document.DoesNotExist):
@@ -1103,6 +1202,56 @@ class DocumentViewSet(
                 "Error emailing document, check logs for more detail.",
             )
 
+    @action(methods=["post"], detail=True)
+    def update_version(self, request, pk=None):
+        serializer = DocumentVersionSerializer(data=request.data)
+        serializer.is_valid(raise_exception=True)
+
+        try:
+            doc = Document.objects.select_related("owner").get(pk=pk)
+            if request.user is not None and not has_perms_owner_aware(
+                request.user,
+                "change_document",
+                doc,
+            ):
+                return HttpResponseForbidden("Insufficient permissions")
+        except Document.DoesNotExist:
+            raise Http404
+
+        try:
+            doc_name, doc_data = serializer.validated_data.get("document")
+
+            t = int(mktime(datetime.now().timetuple()))
+
+            settings.SCRATCH_DIR.mkdir(parents=True, exist_ok=True)
+
+            temp_file_path = Path(tempfile.mkdtemp(dir=settings.SCRATCH_DIR)) / Path(
+                pathvalidate.sanitize_filename(doc_name),
+            )
+
+            temp_file_path.write_bytes(doc_data)
+
+            os.utime(temp_file_path, times=(t, t))
+
+            input_doc = ConsumableDocument(
+                source=DocumentSource.ApiUpload,
+                original_file=temp_file_path,
+                head_version_id=doc.pk,
+            )
+
+            async_task = consume_file.delay(
+                input_doc,
+            )
+            logger.debug(
+                f"Updated document {doc.id} with new version",
+            )
+            return Response(async_task.id)
+        except Exception as e:
+            logger.warning(f"An error occurred updating document: {e!s}")
+            return HttpResponseServerError(
+                "Error updating document, check logs for more detail.",
+            )
+
 
 @extend_schema_view(
     list=extend_schema(

src/paperless/settings.py

@@ -322,7 +322,6 @@ INSTALLED_APPS = [
     "paperless_tesseract.apps.PaperlessTesseractConfig",
     "paperless_text.apps.PaperlessTextConfig",
     "paperless_mail.apps.PaperlessMailConfig",
-    "paperless_remote.apps.PaperlessRemoteParserConfig",
     "django.contrib.admin",
     "rest_framework",
     "rest_framework.authtoken",
@@ -923,7 +922,7 @@ CELERY_ACCEPT_CONTENT = ["application/json", "application/x-python-serialize"]
 CELERY_BEAT_SCHEDULE = _parse_beat_schedule()
 
 # https://docs.celeryq.dev/en/stable/userguide/configuration.html#beat-schedule-filename
-CELERY_BEAT_SCHEDULE_FILENAME = str(DATA_DIR / "celerybeat-schedule.db")
+CELERY_BEAT_SCHEDULE_FILENAME = DATA_DIR / "celerybeat-schedule.db"
 
 
 # Cachalot: Database read cache.
@@ -1390,10 +1389,3 @@ WEBHOOKS_ALLOW_INTERNAL_REQUESTS = __get_boolean(
     "PAPERLESS_WEBHOOKS_ALLOW_INTERNAL_REQUESTS",
     "true",
 )
-
-###############################################################################
-# Remote Parser                                                               #
-###############################################################################
-REMOTE_OCR_ENGINE = os.getenv("PAPERLESS_REMOTE_OCR_ENGINE")
-REMOTE_OCR_API_KEY = os.getenv("PAPERLESS_REMOTE_OCR_API_KEY")
-REMOTE_OCR_ENDPOINT = os.getenv("PAPERLESS_REMOTE_OCR_ENDPOINT")

src/paperless_remote/__init__.py

@@ -1,4 +0,0 @@
-# this is here so that django finds the checks.
-from paperless_remote.checks import check_remote_parser_configured
-
-__all__ = ["check_remote_parser_configured"]

src/paperless_remote/apps.py

@@ -1,14 +0,0 @@
-from django.apps import AppConfig
-
-from paperless_remote.signals import remote_consumer_declaration
-
-
-class PaperlessRemoteParserConfig(AppConfig):
-    name = "paperless_remote"
-
-    def ready(self):
-        from documents.signals import document_consumer_declaration
-
-        document_consumer_declaration.connect(remote_consumer_declaration)
-
-        AppConfig.ready(self)

src/paperless_remote/checks.py

@@ -1,17 +0,0 @@
-from django.conf import settings
-from django.core.checks import Error
-from django.core.checks import register
-
-
-@register()
-def check_remote_parser_configured(app_configs, **kwargs):
-    if settings.REMOTE_OCR_ENGINE == "azureai" and not (
-        settings.REMOTE_OCR_ENDPOINT and settings.REMOTE_OCR_API_KEY
-    ):
-        return [
-            Error(
-                "Azure AI remote parser requires endpoint and API key to be configured.",
-            ),
-        ]
-
-    return []

src/paperless_remote/parsers.py

@@ -1,113 +0,0 @@
-from pathlib import Path
-
-from django.conf import settings
-
-from paperless_tesseract.parsers import RasterisedDocumentParser
-
-
-class RemoteEngineConfig:
-    def __init__(
-        self,
-        engine: str,
-        api_key: str | None = None,
-        endpoint: str | None = None,
-    ):
-        self.engine = engine
-        self.api_key = api_key
-        self.endpoint = endpoint
-
-    def engine_is_valid(self):
-        valid = self.engine in ["azureai"] and self.api_key is not None
-        if self.engine == "azureai":
-            valid = valid and self.endpoint is not None
-        return valid
-
-
-class RemoteDocumentParser(RasterisedDocumentParser):
-    """
-    This parser uses a remote OCR engine to parse documents. Currently, it supports Azure AI Vision
-    as this is the only service that provides a remote OCR API with text-embedded PDF output.
-    """
-
-    logging_name = "paperless.parsing.remote"
-
-    def get_settings(self) -> RemoteEngineConfig:
-        """
-        Returns the configuration for the remote OCR engine, loaded from Django settings.
-        """
-        return RemoteEngineConfig(
-            engine=settings.REMOTE_OCR_ENGINE,
-            api_key=settings.REMOTE_OCR_API_KEY,
-            endpoint=settings.REMOTE_OCR_ENDPOINT,
-        )
-
-    def supported_mime_types(self):
-        if self.settings.engine_is_valid():
-            return {
-                "application/pdf": ".pdf",
-                "image/png": ".png",
-                "image/jpeg": ".jpg",
-                "image/tiff": ".tiff",
-                "image/bmp": ".bmp",
-                "image/gif": ".gif",
-                "image/webp": ".webp",
-            }
-        else:
-            return {}
-
-    def azure_ai_vision_parse(
-        self,
-        file: Path,
-    ) -> str | None:
-        """
-        Uses Azure AI Vision to parse the document and return the text content.
-        It requests a searchable PDF output with embedded text.
-        The PDF is saved to the archive_path attribute.
-        Returns the text content extracted from the document.
-        If the parsing fails, it returns None.
-        """
-        from azure.ai.documentintelligence import DocumentIntelligenceClient
-        from azure.ai.documentintelligence.models import AnalyzeDocumentRequest
-        from azure.ai.documentintelligence.models import AnalyzeOutputOption
-        from azure.ai.documentintelligence.models import DocumentContentFormat
-        from azure.core.credentials import AzureKeyCredential
-
-        client = DocumentIntelligenceClient(
-            endpoint=self.settings.endpoint,
-            credential=AzureKeyCredential(self.settings.api_key),
-        )
-
-        with file.open("rb") as f:
-            analyze_request = AnalyzeDocumentRequest(bytes_source=f.read())
-            poller = client.begin_analyze_document(
-                model_id="prebuilt-read",
-                body=analyze_request,
-                output_content_format=DocumentContentFormat.TEXT,
-                output=[AnalyzeOutputOption.PDF],  # request searchable PDF output
-                content_type="application/json",
-            )
-
-        poller.wait()
-        result_id = poller.details["operation_id"]
-        result = poller.result()
-
-        # Download the PDF with embedded text
-        self.archive_path = self.tempdir / "archive.pdf"
-        with self.archive_path.open("wb") as f:
-            for chunk in client.get_analyze_result_pdf(
-                model_id="prebuilt-read",
-                result_id=result_id,
-            ):
-                f.write(chunk)
-
-        client.close()
-        return result.content
-
-    def parse(self, document_path: Path, mime_type, file_name=None):
-        if not self.settings.engine_is_valid():
-            self.log.warning(
-                "No valid remote parser engine is configured, content will be empty.",
-            )
-            self.text = ""
-        elif self.settings.engine == "azureai":
-            self.text = self.azure_ai_vision_parse(document_path)

src/paperless_remote/signals.py

@@ -1,18 +0,0 @@
-def get_parser(*args, **kwargs):
-    from paperless_remote.parsers import RemoteDocumentParser
-
-    return RemoteDocumentParser(*args, **kwargs)
-
-
-def get_supported_mime_types():
-    from paperless_remote.parsers import RemoteDocumentParser
-
-    return RemoteDocumentParser(None).supported_mime_types()
-
-
-def remote_consumer_declaration(sender, **kwargs):
-    return {
-        "parser": get_parser,
-        "weight": 5,
-        "mime_types": get_supported_mime_types(),
-    }

src/paperless_remote/tests/test_checks.py

@@ -1,24 +0,0 @@
-from unittest import TestCase
-
-from django.test import override_settings
-
-from paperless_remote import check_remote_parser_configured
-
-
-class TestChecks(TestCase):
-    @override_settings(REMOTE_OCR_ENGINE=None)
-    def test_no_engine(self):
-        msgs = check_remote_parser_configured(None)
-        self.assertEqual(len(msgs), 0)
-
-    @override_settings(REMOTE_OCR_ENGINE="azureai")
-    @override_settings(REMOTE_OCR_API_KEY="somekey")
-    @override_settings(REMOTE_OCR_ENDPOINT=None)
-    def test_azure_no_endpoint(self):
-        msgs = check_remote_parser_configured(None)
-        self.assertEqual(len(msgs), 1)
-        self.assertTrue(
-            msgs[0].msg.startswith(
-                "Azure AI remote parser requires endpoint and API key to be configured.",
-            ),
-        )

src/paperless_remote/tests/test_parser.py

@@ -1,101 +0,0 @@
-import uuid
-from pathlib import Path
-from unittest import mock
-
-from django.test import TestCase
-from django.test import override_settings
-
-from documents.tests.utils import DirectoriesMixin
-from documents.tests.utils import FileSystemAssertsMixin
-from paperless_remote.parsers import RemoteDocumentParser
-from paperless_remote.signals import get_parser
-
-
-class TestParser(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
-    SAMPLE_FILES = Path(__file__).resolve().parent / "samples"
-
-    def assertContainsStrings(self, content: str, strings: list[str]):
-        # Asserts that all strings appear in content, in the given order.
-        indices = []
-        for s in strings:
-            if s in content:
-                indices.append(content.index(s))
-            else:
-                self.fail(f"'{s}' is not in '{content}'")
-        self.assertListEqual(indices, sorted(indices))
-
-    @mock.patch("paperless_tesseract.parsers.run_subprocess")
-    @mock.patch("azure.ai.documentintelligence.DocumentIntelligenceClient")
-    def test_get_text_with_azure(self, mock_client_cls, mock_subprocess):
-        # Arrange mock Azure client
-        mock_client = mock.Mock()
-        mock_client_cls.return_value = mock_client
-
-        # Simulate poller result and its `.details`
-        mock_poller = mock.Mock()
-        mock_poller.wait.return_value = None
-        mock_poller.details = {"operation_id": "fake-op-id"}
-        mock_client.begin_analyze_document.return_value = mock_poller
-        mock_poller.result.return_value.content = "This is a test document."
-
-        # Return dummy PDF bytes
-        mock_client.get_analyze_result_pdf.return_value = [
-            b"%PDF-",
-            b"1.7 ",
-            b"FAKEPDF",
-        ]
-
-        # Simulate pdftotext by writing dummy text to sidecar file
-        def fake_run(cmd, *args, **kwargs):
-            with Path(cmd[-1]).open("w", encoding="utf-8") as f:
-                f.write("This is a test document.")
-
-        mock_subprocess.side_effect = fake_run
-
-        with override_settings(
-            REMOTE_OCR_ENGINE="azureai",
-            REMOTE_OCR_API_KEY="somekey",
-            REMOTE_OCR_ENDPOINT="https://endpoint.cognitiveservices.azure.com",
-        ):
-            parser = get_parser(uuid.uuid4())
-            parser.parse(
-                self.SAMPLE_FILES / "simple-digital.pdf",
-                "application/pdf",
-            )
-
-            self.assertContainsStrings(
-                parser.text.strip(),
-                ["This is a test document."],
-            )
-
-    @override_settings(
-        REMOTE_OCR_ENGINE="azureai",
-        REMOTE_OCR_API_KEY="key",
-        REMOTE_OCR_ENDPOINT="https://endpoint.cognitiveservices.azure.com",
-    )
-    def test_supported_mime_types_valid_config(self):
-        parser = RemoteDocumentParser(uuid.uuid4())
-        expected_types = {
-            "application/pdf": ".pdf",
-            "image/png": ".png",
-            "image/jpeg": ".jpg",
-            "image/tiff": ".tiff",
-            "image/bmp": ".bmp",
-            "image/gif": ".gif",
-            "image/webp": ".webp",
-        }
-        self.assertEqual(parser.supported_mime_types(), expected_types)
-
-    def test_supported_mime_types_invalid_config(self):
-        parser = get_parser(uuid.uuid4())
-        self.assertEqual(parser.supported_mime_types(), {})
-
-    @override_settings(
-        REMOTE_OCR_ENGINE=None,
-        REMOTE_OCR_API_KEY=None,
-        REMOTE_OCR_ENDPOINT=None,
-    )
-    def test_parse_with_invalid_config(self):
-        parser = get_parser(uuid.uuid4())
-        parser.parse(self.SAMPLE_FILES / "simple-digital.pdf", "application/pdf")
-        self.assertEqual(parser.text, "")

uv.lock

@@ -95,34 +95,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/af/cc/55a32a2c98022d88812b5986d2a92c4ff3ee087e83b712ebc703bba452bf/Automat-24.8.1-py3-none-any.whl", hash = "sha256:bf029a7bc3da1e2c24da2343e7598affaa9f10bf0ab63ff808566ce90551e02a", size = 42585, upload-time = "2024-08-19T17:31:56.729Z" },
 ]
 
-[[package]]
-name = "azure-ai-documentintelligence"
-version = "1.0.2"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "azure-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-    { name = "isodate", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-    { name = "typing-extensions", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/44/7b/8115cd713e2caa5e44def85f2b7ebd02a74ae74d7113ba20bdd41fd6dd80/azure_ai_documentintelligence-1.0.2.tar.gz", hash = "sha256:4d75a2513f2839365ebabc0e0e1772f5601b3a8c9a71e75da12440da13b63484", size = 170940 }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/d9/75/c9ec040f23082f54ffb1977ff8f364c2d21c79a640a13d1c1809e7fd6b1a/azure_ai_documentintelligence-1.0.2-py3-none-any.whl", hash = "sha256:e1fb446abbdeccc9759d897898a0fe13141ed29f9ad11fc705f951925822ed59", size = 106005 },
-]
-
-[[package]]
-name = "azure-core"
-version = "1.33.0"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "requests", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-    { name = "six", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-    { name = "typing-extensions", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/75/aa/7c9db8edd626f1a7d99d09ef7926f6f4fb34d5f9fa00dc394afdfe8e2a80/azure_core-1.33.0.tar.gz", hash = "sha256:f367aa07b5e3005fec2c1e184b882b0b039910733907d001c20fb08ebb8c0eb9", size = 295633 }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/07/b7/76b7e144aa53bd206bf1ce34fa75350472c3f69bf30e5c8c18bc9881035d/azure_core-1.33.0-py3-none-any.whl", hash = "sha256:9b5b6d0223a1d38c37500e6971118c1e0f13f54951e6893968b38910bc9cda8f", size = 207071 },
-]
-
 [[package]]
 name = "babel"
 version = "2.17.0"
@@ -1440,15 +1412,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/c7/fc/4e5a141c3f7c7bed550ac1f69e599e92b6be449dd4677ec09f325cad0955/inotifyrecursive-0.3.5-py3-none-any.whl", hash = "sha256:7e5f4a2e1dc2bef0efa3b5f6b339c41fb4599055a2b54909d020e9e932cc8d2f", size = 8009, upload-time = "2020-11-20T12:38:46.981Z" },
 ]
 
-[[package]]
-name = "isodate"
-version = "0.7.2"
-source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/54/4d/e940025e2ce31a8ce1202635910747e5a87cc3a6a6bb2d00973375014749/isodate-0.7.2.tar.gz", hash = "sha256:4cd1aa0f43ca76f4a6c6c0292a85f40b35ec2e43e315b59f06e6d32171a953e6", size = 29705 }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/15/aa/0aca39a37d3c7eb941ba736ede56d689e7be91cab5d9ca846bde3999eba6/isodate-0.7.2-py3-none-any.whl", hash = "sha256:28009937d8031054830160fce6d409ed342816b543597cece116d966c6d99e15", size = 22320 },
-]
-
 [[package]]
 name = "jinja2"
 version = "3.1.6"
@@ -2069,7 +2032,6 @@ name = "paperless-ngx"
 version = "2.18.4"
 source = { virtual = "." }
 dependencies = [
-    { name = "azure-ai-documentintelligence", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
     { name = "babel", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
     { name = "bleach", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
     { name = "celery", extra = ["redis"], marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
@@ -2207,7 +2169,6 @@ typing = [
 
 [package.metadata]
 requires-dist = [
-    { name = "azure-ai-documentintelligence", specifier = ">=1.0.2" },
     { name = "babel", specifier = ">=2.17" },
     { name = "bleach", specifier = "~=6.2.0" },
     { name = "celery", extras = ["redis"], specifier = "~=5.5.1" },