Fix: use a root document's effective content for the LLM index and prompts

The search index and the document views use the newest version's content as a root document's content, but everything the LLM side reads used the root's own content: the text embedded in the LLM index, the content in the classification prompt, the context blocks from similar documents, and the text that is searched for similar documents. After a document was replaced by a new version the LLM kept answering from the old text. Read get_effective_content() in those four places. The index update annotates the content in the query like the search index does, and the similar-document lookup annotates it for the documents it fetches, so no query is made per document. Entries already in the LLM index keep the old text until the document next changes or the index is rebuilt.
Fix: index document versions as their root document
2026-10-20 06:48:16 +00:00 · 2026-10-08 15:22:03 -07:00 · 2026-10-08 15:22:02 -07:00 · 2026-10-08 15:21:45 -07:00 · 2026-10-08 15:21:28 -07:00
178 changed files with 32438 additions and 32638 deletions
@@ -66,7 +66,7 @@ services:
    command:
      - "gotenberg"
      - "--chromium-disable-javascript=true"
-      - "--chromium-allow-list=^file:///tmp/.*"
+      - "--chromium-allow-list=file:///tmp/.*"
  tika:
    image: docker.io/apache/tika:3.3.1.0
    restart: unless-stopped
@@ -31,19 +31,9 @@ body:
      value: |
        #### 🤖 Use of AI tools

-        If this report was generated, in whole or in part, by an AI tool or agent you **must** say so using the field below. Failing to clearly disclose this is a violation of our [Code of Conduct](https://github.com/paperless-ngx/paperless-ngx/blob/main/CODE_OF_CONDUCT.md).
+        If this report was generated, in whole or in part, by an AI tool or agent you **must** clearly say so in the description below. Failing to clearly disclose this is a violation of our [Code of Conduct](https://github.com/paperless-ngx/paperless-ngx/blob/main/CODE_OF_CONDUCT.md).

        Reports written by an AI agent must describe the _observed behavior only_: what happened, how to reproduce it, logs etc. Do **not** include code analysis, speculation about the root cause, or suggest code fixes.
-  - type: dropdown
-    id: ai-usage
-    attributes:
-      label: AI usage
-      description: Was any part of this report written by an AI tool or agent?
-      options:
-        - No AI tools were used
-        - Yes, AI tools helped write this report
-    validations:
-      required: true
  - type: textarea
    id: description
    attributes:
@@ -43,12 +43,6 @@ jobs:
            const haystack = `${issue.title}\n${issue.body ?? ''}`;
            const honeypot = haystack.toUpperCase().includes('ASLOP-PR-VERIFY');

-            // The form's required "AI usage" dropdown renders as a `### AI usage` section. An agent
-            // that follows the template trips the honeypot even when it discloses, so only treat
-            // the honeypot as undisclosed AI use when the reporter did not answer "Yes" here.
-            const aiUsage = (issue.body ?? '').match(/^###\s+AI usage\s*\n+\s*(.+?)\s*$/m)?.[1] ?? '';
-            const disclosedAI = aiUsage.toLowerCase().startsWith('yes');
-
            // Issues opened through the form always get the template's default labels. GitHub
            // applies them a second or two *after* creation, so re-read them instead of trusting
            // the webhook payload, and retry before concluding that there are none.
@@ -61,30 +55,19 @@ jobs:
              if (labels.some((label) => templateLabels.includes(label))) break;
            }
            const bypassedTemplate = !labels.some((label) => templateLabels.includes(label));
-            core.info(
-              `Labels: [${labels.join(', ')}], honeypot: ${honeypot}, AI usage: "${aiUsage}", ` +
-                `bypassed template: ${bypassedTemplate}`
-            );
+            core.info(`Labels: [${labels.join(', ')}], honeypot: ${honeypot}, bypassed template: ${bypassedTemplate}`);

-            if (!bypassedTemplate && (disclosedAI || !honeypot)) {
-              if (disclosedAI) {
-                core.info('AI use disclosed, labeling only');
-                await github.rest.issues.addLabels({ ...common, labels: ['ai'] });
-              } else {
-                core.info('No slop signals found');
-              }
+            if (!honeypot && !bypassedTemplate) {
+              core.info('No slop signals found');
              return;
            }

+            // The honeypot is only ever tripped deliberately, so that message can name the cause.
            // A missing template label only tells us the issue did not come from the form, which
-            // is reason enough to close it, but not proof of how it was written. It takes priority:
-            // a hand-built body can carry the honeypot *and* a disclosure, and the form message is
-            // the only one that is accurate for both. Reaching the honeypot message means the issue
-            // came from the form without disclosing AI usage, so that message can name the cause.
-            const body = honeypot && !bypassedTemplate
+            // is reason enough to close it, but not proof of how it was written.
+            const body = honeypot
              ? 'This issue was automatically closed because it contains a marker that is only visible to ' +
-                'automated tools, which indicates it was generated by an AI agent, but the "AI usage" field does not ' +
-                'disclose this.\n\n' +
+                'automated tools, which indicates it was generated by an AI agent without being disclosed as such.\n\n' +
                `Please see our [contributing guidelines](${contributing}) and [Code of Conduct](${codeOfConduct}). ` +
                'You are welcome to open a new issue that describes the problem you observed in your own words.'
              : 'This issue was automatically closed because it was not opened using our bug report form. ' +
@@ -38,6 +38,16 @@ jobs:
        uses: actions/labeler@bf12e9b00b37c5c0ca2b87b79b2daf7891dbda13 # v7.0.0
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
+      - name: Label by size
+        uses: Gascon1/pr-size-labeler@deff8ed00a76639a7c0f197525bafa3350ba4c36 # v1.3.0
+        with:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          xs_label: 'small-change'
+          xs_diff: '9'
+          s_label: 'non-trivial'
+          s_diff: '99999'
+          fail_if_xl: 'false'
+          excluded_files: /\.lock$/ /\.txt$/ ^src-ui/pnpm-lock\.yaml$ ^src-ui/messages\.xlf$ ^src/locale/en_US/LC_MESSAGES/django\.po$
      - name: Label by PR title
        uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
        with:
@@ -43,17 +43,26 @@ Once you have submitted a **P**ull **R**equest it will be reviewed, approved, an

 Important: 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. Instead of opening a PR which does not meet this requirement, please open a feature request instead, to gather feedback from both users and the project maintainers.

-## PR Review Process
+## Non-Trivial Requests

-Our community review process for PRs is the following:
+PRs deemed `non-trivial` will go through a stricter review process before being merged into `dev`. This is to ensure code quality and complete functionality (free of side effects).
+
+Examples of `non-trivial` PRs might include:
+
+- Additional features
+- Large changes to many distinct files
+- Breaking or deprecation of existing features
+
+Our community review process for `non-trivial` PRs is the following:

 1. Must pass usual automated code tests and formatting checks.
 2. The PR will be assigned and pinged to the appropriately experienced team (i.e. @paperless-ngx/backend for backend changes).
 3. Development team will check and test code manually (possibly over several days).
   - You may be asked to make changes or rebase.
-4. One or more members of the team will approve and finally merge the request into `dev` 🎉.
+   - The team may ask for additional testing done by @paperless-ngx/test
+4. **At least two** members of the team will approve and finally merge the request into `dev` 🎉.

-This process might be slow as community members have different schedules and time to dedicate to the Paperless project. However it ensures community code reviews are thorough and of high quality.
+This process might be slow as community members have different schedules and time to dedicate to the Paperless project. However it ensures community code reviews are as brilliantly thorough as they once were with @jonaswinkler.

 # Use of AI Tools

@@ -21,7 +21,7 @@ Paperless-ngx is a document management system that transforms your physical docu

 Paperless-ngx is the official successor to the original [Paperless](https://github.com/the-paperless-project/paperless) & [Paperless-ng](https://github.com/jonaswinkler/paperless-ng) projects and is designed to distribute the responsibility of advancing and supporting the project among a team of people. [Consider joining us!](#community-support)

-Thanks to the generous folks at [Fyra Stack](https://fyrastack.com), a demo is available at [demo.paperless-ngx.com](https://demo.paperless-ngx.com) using login `demo` / `demo`. _Note: demo content is reset frequently and confidential information should not be uploaded._
+Thanks to the generous folks at [DigitalOcean](https://m.do.co/c/8d70b916d462), a demo is available at [demo.paperless-ngx.com](https://demo.paperless-ngx.com) using login `demo` / `demo`. _Note: demo content is reset frequently and confidential information should not be uploaded._

 - [Features](#features)
 - [Getting started](#getting-started)
@@ -33,9 +33,13 @@ Thanks to the generous folks at [Fyra Stack](https://fyrastack.com), a demo is a
 - [Related Projects](#related-projects)
 - [Important Note](#important-note)

-<p align="right">Demo hosting provided by:<br/>
-  <a href="https://fyrastack.com">
-    <img src="docs/assets/sponsors/fyra-stack.svg" alt="" width="28px" align="absmiddle"> Fyra Stack
+<p align="right">This project is supported by:<br/>
+  <a href="https://m.do.co/c/8d70b916d462" style="padding-top: 4px; display: block;">
+    <picture>
+      <source media="(prefers-color-scheme: dark)" srcset="https://opensource.nyc3.cdn.digitaloceanspaces.com/attribution/assets/SVG/DO_Logo_horizontal_white.svg" width="140px">
+      <source media="(prefers-color-scheme: light)" srcset="https://opensource.nyc3.cdn.digitaloceanspaces.com/attribution/assets/SVG/DO_Logo_horizontal_blue.svg" width="140px">
+      <img src="https://opensource.nyc3.cdn.digitaloceanspaces.com/attribution/assets/SVG/DO_Logo_horizontal_black_.svg" width="140px">
+    </picture>
  </a>
 </p>

@@ -14,7 +14,7 @@ services:
    command:
      - "gotenberg"
      - "--chromium-disable-javascript=true"
-      - "--chromium-allow-list=^file:///tmp/.*"
+      - "--chromium-allow-list=file:///tmp/.*"
      - "--log-level=warn"
      - "--log-format=text"
  tika:
@@ -40,6 +40,7 @@ services:
    volumes:
      - dbdata:/var/lib/mysql
    environment:
+      MARIADB_HOST: paperless
      MARIADB_DATABASE: paperless
      MARIADB_USER: paperless
      MARIADB_PASSWORD: paperless
@@ -78,7 +79,7 @@ services:
    command:
      - "gotenberg"
      - "--chromium-disable-javascript=true"
-      - "--chromium-allow-list=^file:///tmp/.*"
+      - "--chromium-allow-list=file:///tmp/.*"
  tika:
    image: docker.io/apache/tika:3.3.1.0
    restart: unless-stopped
@@ -36,6 +36,7 @@ services:
    volumes:
      - dbdata:/var/lib/mysql
    environment:
+      MARIADB_HOST: paperless
      MARIADB_DATABASE: paperless
      MARIADB_USER: paperless
      MARIADB_PASSWORD: paperless
@@ -74,7 +74,7 @@ services:
    command:
      - "gotenberg"
      - "--chromium-disable-javascript=true"
-      - "--chromium-allow-list=^file:///tmp/.*"
+      - "--chromium-allow-list=file:///tmp/.*"
  tika:
    image: docker.io/apache/tika:3.3.1.0
    restart: unless-stopped
@@ -63,7 +63,7 @@ services:
    command:
      - "gotenberg"
      - "--chromium-disable-javascript=true"
-      - "--chromium-allow-list=^file:///tmp/.*"
+      - "--chromium-allow-list=file:///tmp/.*"
  tika:
    image: docker.io/apache/tika:3.3.1.0
    restart: unless-stopped
@@ -101,32 +101,3 @@ h4 code {
        border-top: .05rem solid var(--md-default-fg-color--lightest);
    }
 }
-
-.md-typeset .demo-sponsor {
-    display: flex;
-    align-items: center;
-    flex-wrap: wrap;
-    gap: .35rem;
-    margin-top: -1rem;
-    margin-bottom: 2rem;
-    color: var(--md-default-fg-color--light);
-    font-size: .65rem;
-}
-
-.md-typeset .demo-sponsor a {
-    display: inline-flex;
-    align-items: center;
-    gap: .25rem;
-    color: inherit;
-    font-weight: 600;
-}
-
-.md-typeset .demo-sponsor a:hover {
-    color: var(--md-accent-fg-color);
-}
-
-.md-typeset .demo-sponsor img {
-    width: 1rem;
-    height: 1rem;
-    flex-shrink: 0;
-}
@@ -1 +0,0 @@
-<svg xmlns="http://www.w3.org/2000/svg" width="666.667" height="666.667" viewBox="0 0 666.667 666.667"><g stroke-width="1.333"><path fill="#fef400" d="m332.66668 418.66668h-84.66667v-85.33334-85.33333h84.66667 84.66667v-85.33334-85.333334h85.33333 85.33334v85.333334 85.33334h-85.33334-85.33333v85.33333 85.33334z"/><path fill="#fe6565" d="m248.00001 589.33335h-170.666674v-170.66667-170.66667h85.333334 85.33334v-85.33334-85.333334h84.66667 84.66667v85.333334 85.33334h85.33333 85.33334v85.33333 85.33334h-84.66667-84.66667v85.33334 85.33333zm84.66667-170.66667h84.66667v-85.33334-85.33333h-84.66667-84.66667v85.33333 85.33334z"/></g></svg>
@@ -1,205 +1,5 @@
 # Changelog

-## paperless-ngx 3.3.2
-
-### Bug Fixes
-
- Fix: copy only direct user grants in DocumentMetadataOverrides.from\_document [@stumpylog](https://github.com/stumpylog) ([#14452](https://github.com/paperless-ngx/paperless-ngx/pull/14452))
- Fix: support combined text/query searches [@shamoon](https://github.com/shamoon) ([#14443](https://github.com/paperless-ngx/paperless-ngx/pull/14443))
- Fix: make exporter derived file names unique when originals differ only by extension [@stumpylog](https://github.com/stumpylog) ([#14429](https://github.com/paperless-ngx/paperless-ngx/pull/14429))
-
-### Dependencies
-
- Chore(deps): Bump filelock from 3.32.6 to 4.0.12 @[dependabot[bot]](https://github.com/apps/dependabot) ([#14336](https://github.com/paperless-ngx/paperless-ngx/pull/14336))
- Chore(deps): Bump the uv group across 1 directory with 3 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14418](https://github.com/paperless-ngx/paperless-ngx/pull/14418))
-
-### All App Changes
-
-<details>
-<summary>5 changes</summary>
-
- Fix: copy only direct user grants in DocumentMetadataOverrides.from\_document [@stumpylog](https://github.com/stumpylog) ([#14452](https://github.com/paperless-ngx/paperless-ngx/pull/14452))
- Chore(deps): Bump filelock from 3.32.6 to 4.0.12 @[dependabot[bot]](https://github.com/apps/dependabot) ([#14336](https://github.com/paperless-ngx/paperless-ngx/pull/14336))
- Fix: support combined text/query searches [@shamoon](https://github.com/shamoon) ([#14443](https://github.com/paperless-ngx/paperless-ngx/pull/14443))
- Chore(deps): Bump the uv group across 1 directory with 3 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14418](https://github.com/paperless-ngx/paperless-ngx/pull/14418))
- Fix: make exporter derived file names unique when originals differ only by extension [@stumpylog](https://github.com/stumpylog) ([#14429](https://github.com/paperless-ngx/paperless-ngx/pull/14429))
-
-</details>
-
-## paperless-ngx 3.3.1
-
-### Bug Fixes
-
- Fix: use a root document's effective content for the LLM index and pr… [@stumpylog](https://github.com/stumpylog) ([#14423](https://github.com/paperless-ngx/paperless-ngx/pull/14423))
- Fix: index document versions as their root document [@stumpylog](https://github.com/stumpylog) ([#14422](https://github.com/paperless-ngx/paperless-ngx/pull/14422))
- Fix: authorize document versions by their root in the single-object permission check [@stumpylog](https://github.com/stumpylog) ([#14421](https://github.com/paperless-ngx/paperless-ngx/pull/14421))
- Fix: Provide better error messages to users for invalid JSON [@stumpylog](https://github.com/stumpylog) ([#14419](https://github.com/paperless-ngx/paperless-ngx/pull/14419))
- Fix: authorize document versions by their root and speed up permission id sets [@stumpylog](https://github.com/stumpylog) ([#14396](https://github.com/paperless-ngx/paperless-ngx/pull/14396))
- Chore: update the db options warning [@shamoon](https://github.com/shamoon) ([#14398](https://github.com/paperless-ngx/paperless-ngx/pull/14398))
- Fix: bulk reprocess latest version for root documents [@shamoon](https://github.com/shamoon) ([#14385](https://github.com/paperless-ngx/paperless-ngx/pull/14385))
- Fix: consistently use root doc for version action permissions [@shamoon](https://github.com/shamoon) ([#14384](https://github.com/paperless-ngx/paperless-ngx/pull/14384))
- Fix: Don't attempt to index created dates which are not representable in the FTS index [@stumpylog](https://github.com/stumpylog) ([#14390](https://github.com/paperless-ngx/paperless-ngx/pull/14390))
- Fix: retry a search index rebuild that was interrupted [@stumpylog](https://github.com/stumpylog) ([#14379](https://github.com/paperless-ngx/paperless-ngx/pull/14379))
-
-### Documentation
-
- Chore: silence gotenberg startup warning [@slankes](https://github.com/slankes) ([#14405](https://github.com/paperless-ngx/paperless-ngx/pull/14405))
- Documentation: remove MARIADB\_HOST from MariaDB compose files [@shamoon](https://github.com/shamoon) ([#14378](https://github.com/paperless-ngx/paperless-ngx/pull/14378))
- Documentation: Updates the troubleshooting with 2 common database collation issues [@stumpylog](https://github.com/stumpylog) ([#14370](https://github.com/paperless-ngx/paperless-ngx/pull/14370))
-
-### Maintenance
-
- Chore: Upgrade runners to Ubuntu 26.04 [@stumpylog](https://github.com/stumpylog) ([#14347](https://github.com/paperless-ngx/paperless-ngx/pull/14347))
-
-### Dependencies
-
- Chore(deps): Bump the npm\_and\_yarn group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14417](https://github.com/paperless-ngx/paperless-ngx/pull/14417))
-
-### All App Changes
-
-<details>
-<summary>12 changes</summary>
-
- Chore: share the authorizing owner rule between the id sets and the trash filter [@stumpylog](https://github.com/stumpylog) ([#14425](https://github.com/paperless-ngx/paperless-ngx/pull/14425))
- Fix: use a root document's effective content for the LLM index and pr… [@stumpylog](https://github.com/stumpylog) ([#14423](https://github.com/paperless-ngx/paperless-ngx/pull/14423))
- Fix: index document versions as their root document [@stumpylog](https://github.com/stumpylog) ([#14422](https://github.com/paperless-ngx/paperless-ngx/pull/14422))
- Fix: authorize document versions by their root in the single-object permission check [@stumpylog](https://github.com/stumpylog) ([#14421](https://github.com/paperless-ngx/paperless-ngx/pull/14421))
- Fix: Provide better error messages to users for invalid JSON [@stumpylog](https://github.com/stumpylog) ([#14419](https://github.com/paperless-ngx/paperless-ngx/pull/14419))
- Chore(deps): Bump the npm\_and\_yarn group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14417](https://github.com/paperless-ngx/paperless-ngx/pull/14417))
- Fix: authorize document versions by their root and speed up permission id sets [@stumpylog](https://github.com/stumpylog) ([#14396](https://github.com/paperless-ngx/paperless-ngx/pull/14396))
- Chore: update the db options warning [@shamoon](https://github.com/shamoon) ([#14398](https://github.com/paperless-ngx/paperless-ngx/pull/14398))
- Fix: bulk reprocess latest version for root documents [@shamoon](https://github.com/shamoon) ([#14385](https://github.com/paperless-ngx/paperless-ngx/pull/14385))
- Fix: consistently use root doc for version action permissions [@shamoon](https://github.com/shamoon) ([#14384](https://github.com/paperless-ngx/paperless-ngx/pull/14384))
- Fix: Don't attempt to index created dates which are not representable in the FTS index [@stumpylog](https://github.com/stumpylog) ([#14390](https://github.com/paperless-ngx/paperless-ngx/pull/14390))
- Fix: retry a search index rebuild that was interrupted [@stumpylog](https://github.com/stumpylog) ([#14379](https://github.com/paperless-ngx/paperless-ngx/pull/14379))
-
-</details>
-
-## paperless-ngx 3.3.0
-
-### Features / Enhancements
-
- Enhancement: more control over suggestion requests [@shamoon](https://github.com/shamoon) ([#14258](https://github.com/paperless-ngx/paperless-ngx/pull/14258))
- Chorehancement: set manifest CORS for credentials [@shamoon](https://github.com/shamoon) ([#14307](https://github.com/paperless-ngx/paperless-ngx/pull/14307))
- Enhancement: include Django admin with 2FA [@shamoon](https://github.com/shamoon) ([#14270](https://github.com/paperless-ngx/paperless-ngx/pull/14270))
- Feature: propagate resolved secrets to interactive container shells [@stumpylog](https://github.com/stumpylog) ([#14254](https://github.com/paperless-ngx/paperless-ngx/pull/14254))
- Enhancement: support separate embedding API key [@furkanural](https://github.com/furkanural) ([#14067](https://github.com/paperless-ngx/paperless-ngx/pull/14067))
- Enhancement: support passthrough extra params for LLMs [@shamoon](https://github.com/shamoon) ([#14202](https://github.com/paperless-ngx/paperless-ngx/pull/14202))
- Feature: store barcode contents, list and search them [@jurassicparkicecream](https://github.com/jurassicparkicecream) ([#14276](https://github.com/paperless-ngx/paperless-ngx/pull/14276))
-
-### Bug Fixes
-
- Fix: Set the ProcessedMail owner based on the rule owner in all cases [@stumpylog](https://github.com/stumpylog) ([#14356](https://github.com/paperless-ngx/paperless-ngx/pull/14356))
- Fix: Ensure log rotation settings are converted to integers [@stumpylog](https://github.com/stumpylog) ([#14343](https://github.com/paperless-ngx/paperless-ngx/pull/14343))
- Fix: Wrap apt calls into a retry so we can ideally jump a slow mirror [@stumpylog](https://github.com/stumpylog) ([#14344](https://github.com/paperless-ngx/paperless-ngx/pull/14344))
- Fix: ship pdf.js CMaps so CJK documents render in the viewer [@MrOggy85](https://github.com/MrOggy85) ([#14318](https://github.com/paperless-ngx/paperless-ngx/pull/14318))
- Fix: use version page\_count for versioned document [@shamoon](https://github.com/shamoon) ([#14280](https://github.com/paperless-ngx/paperless-ngx/pull/14280))
- Fix: allow pointer events for pdf links in pngx viewer [@shamoon](https://github.com/shamoon) ([#14264](https://github.com/paperless-ngx/paperless-ngx/pull/14264))
- Fix: During a move to the trash directory, only attempt to copy metadata [@stumpylog](https://github.com/stumpylog) ([#14250](https://github.com/paperless-ngx/paperless-ngx/pull/14250))
- Fix: convert file mtime to the configured time zone directly [@stumpylog](https://github.com/stumpylog) ([#14249](https://github.com/paperless-ngx/paperless-ngx/pull/14249))
- Fix: ensure documentDeleted subscription is discarded [@shamoon](https://github.com/shamoon) ([#14247](https://github.com/paperless-ngx/paperless-ngx/pull/14247))
- Chore: Fix bugs in the test suite [@stumpylog](https://github.com/stumpylog) ([#14244](https://github.com/paperless-ngx/paperless-ngx/pull/14244))
- Fix: ensure bulk operations are checked against version root [@shamoon](https://github.com/shamoon) ([#14246](https://github.com/paperless-ngx/paperless-ngx/pull/14246))
- Fix: indexing after document-added workflows signal [@shamoon](https://github.com/shamoon) ([#14242](https://github.com/paperless-ngx/paperless-ngx/pull/14242))
- Fix: Record full tag and custom field lists in bulk edit audit log [@stumpylog](https://github.com/stumpylog) ([#14236](https://github.com/paperless-ngx/paperless-ngx/pull/14236))
- Chore: update pikepdf for ocrmypdf requirement [@shamoon](https://github.com/shamoon) ([#14235](https://github.com/paperless-ngx/paperless-ngx/pull/14235))
- Fix: handle legacy bulk edit split page range with missing page\_count [@shamoon](https://github.com/shamoon) ([#14212](https://github.com/paperless-ngx/paperless-ngx/pull/14212))
- Fix: ignore invalid EXIF orientation when generating image archives [@zhzy0077](https://github.com/zhzy0077) ([#14203](https://github.com/paperless-ngx/paperless-ngx/pull/14203))
-
-### Documentation
-
- Documentation: correct duplicates info [@shamoon](https://github.com/shamoon) ([#14243](https://github.com/paperless-ngx/paperless-ngx/pull/14243))
-
-### Maintenance
-
- Chore(deps): Bump the actions group across 1 directory with 4 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14332](https://github.com/paperless-ngx/paperless-ngx/pull/14332))
- Fix: Wrap apt calls into a retry so we can ideally jump a slow mirror [@stumpylog](https://github.com/stumpylog) ([#14344](https://github.com/paperless-ngx/paperless-ngx/pull/14344))
- Chore(deps): Bump the actions group across 1 directory with 10 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14301](https://github.com/paperless-ngx/paperless-ngx/pull/14301))
-
-### Dependencies
-
-<details>
-<summary>27 changes</summary>
-
- Chore(deps): Bump django-filter from 25.2 to 26.1 @[dependabot[bot]](https://github.com/apps/dependabot) ([#14337](https://github.com/paperless-ngx/paperless-ngx/pull/14337))
- Chore(deps): Bump the utilities-patch group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14338](https://github.com/paperless-ngx/paperless-ngx/pull/14338))
- Chore(deps): Bump the pre-commit-dependencies group across 1 directory with 3 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14351](https://github.com/paperless-ngx/paperless-ngx/pull/14351))
- Chore(deps-dev): Bump types-channels from 4.3.0.20260408 to 4.3.0.20260518 @[dependabot[bot]](https://github.com/apps/dependabot) ([#14335](https://github.com/paperless-ngx/paperless-ngx/pull/14335))
- docker(deps): Bump astral-sh/uv from 0.12.20-python3.14-trixie-slim to 0.12.23-python3.14-trixie-slim @[dependabot[bot]](https://github.com/apps/dependabot) ([#14327](https://github.com/paperless-ngx/paperless-ngx/pull/14327))
- Chore(deps): Bump the actions group across 1 directory with 4 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14332](https://github.com/paperless-ngx/paperless-ngx/pull/14332))
- Chore(deps): Bump the uv group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14325](https://github.com/paperless-ngx/paperless-ngx/pull/14325))
- Chore(deps): Bump the frontend-angular-dependencies group across 1 directory with 13 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14329](https://github.com/paperless-ngx/paperless-ngx/pull/14329))
- Chore(deps-dev): Bump prettier from 3.9.8 to 3.9.9 in /src-ui @[dependabot[bot]](https://github.com/apps/dependabot) ([#14331](https://github.com/paperless-ngx/paperless-ngx/pull/14331))
- Chore(deps-dev): Bump the frontend-eslint-dependencies group across 1 directory with 3 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14330](https://github.com/paperless-ngx/paperless-ngx/pull/14330))
- Chore(deps-dev): Bump zensical from 0.0.64 to 0.0.65 in the development group @[dependabot[bot]](https://github.com/apps/dependabot) ([#14326](https://github.com/paperless-ngx/paperless-ngx/pull/14326))
- Chore(deps): Bump the uv group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14314](https://github.com/paperless-ngx/paperless-ngx/pull/14314))
- Chore(deps): Bump the utilities-minor group across 1 directory with 7 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14305](https://github.com/paperless-ngx/paperless-ngx/pull/14305))
- docker-compose(deps): bump greenmail/standalone from 2.1.13 to 2.1.14 in /docker/compose @[dependabot[bot]](https://github.com/apps/dependabot) ([#14281](https://github.com/paperless-ngx/paperless-ngx/pull/14281))
- docker(deps): Bump astral-sh/uv from 0.12.16-python3.14-trixie-slim to 0.12.20-python3.14-trixie-slim @[dependabot[bot]](https://github.com/apps/dependabot) ([#14282](https://github.com/paperless-ngx/paperless-ngx/pull/14282))
- Chore(deps): Bump the pre-commit-dependencies group across 1 directory with 3 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14283](https://github.com/paperless-ngx/paperless-ngx/pull/14283))
- Chore(deps): Bump the utilities-patch group across 1 directory with 6 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14297](https://github.com/paperless-ngx/paperless-ngx/pull/14297))
- Chore(deps): Bump the actions group across 1 directory with 10 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14301](https://github.com/paperless-ngx/paperless-ngx/pull/14301))
- Chore(deps-dev): Bump the frontend-jest-dependencies group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14286](https://github.com/paperless-ngx/paperless-ngx/pull/14286))
- Chore(deps-dev): Bump eslint from 10.10.0 to 10.11.0 in /src-ui in the frontend-eslint-dependencies group across 1 directory @[dependabot[bot]](https://github.com/apps/dependabot) ([#14287](https://github.com/paperless-ngx/paperless-ngx/pull/14287))
- Chore(deps-dev): Bump @types/node from 26.5.0 to 26.6.2 in /src-ui @[dependabot[bot]](https://github.com/apps/dependabot) ([#14288](https://github.com/paperless-ngx/paperless-ngx/pull/14288))
- Chore(deps-dev): Bump prettier from 3.9.6 to 3.9.8 in /src-ui @[dependabot[bot]](https://github.com/apps/dependabot) ([#14289](https://github.com/paperless-ngx/paperless-ngx/pull/14289))
- Chore(deps): Bump the frontend-angular-dependencies group across 1 directory with 10 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14285](https://github.com/paperless-ngx/paperless-ngx/pull/14285))
- Chore: replace bleach with turbohtml [@gaborbernat](https://github.com/gaborbernat) ([#14269](https://github.com/paperless-ngx/paperless-ngx/pull/14269))
- Chore(deps): Bump autobahn from 25.12.2 to 26.7.1 in the uv group across 1 directory @[dependabot[bot]](https://github.com/apps/dependabot) ([#14231](https://github.com/paperless-ngx/paperless-ngx/pull/14231))
- Chore: update pikepdf for ocrmypdf requirement [@shamoon](https://github.com/shamoon) ([#14235](https://github.com/paperless-ngx/paperless-ngx/pull/14235))
- Chore(deps): Bump the pre-commit-dependencies group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14133](https://github.com/paperless-ngx/paperless-ngx/pull/14133))
-
-</details>
-
-### All App Changes
-
-<details>
-<summary>41 changes</summary>
-
- Feature: store barcode contents, list and search them [@jurassicparkicecream](https://github.com/jurassicparkicecream) ([#14276](https://github.com/paperless-ngx/paperless-ngx/pull/14276))
- Fix: Set the ProcessedMail owner based on the rule owner in all cases [@stumpylog](https://github.com/stumpylog) ([#14356](https://github.com/paperless-ngx/paperless-ngx/pull/14356))
- Chore(deps): Bump django-filter from 25.2 to 26.1 @[dependabot[bot]](https://github.com/apps/dependabot) ([#14337](https://github.com/paperless-ngx/paperless-ngx/pull/14337))
- Chore(deps): Bump the utilities-patch group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14338](https://github.com/paperless-ngx/paperless-ngx/pull/14338))
- Chore(deps-dev): Bump types-channels from 4.3.0.20260408 to 4.3.0.20260518 @[dependabot[bot]](https://github.com/apps/dependabot) ([#14335](https://github.com/paperless-ngx/paperless-ngx/pull/14335))
- Chore(deps): Bump the uv group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14325](https://github.com/paperless-ngx/paperless-ngx/pull/14325))
- Fix: Ensure log rotation settings are converted to integers [@stumpylog](https://github.com/stumpylog) ([#14343](https://github.com/paperless-ngx/paperless-ngx/pull/14343))
- Chore(deps): Bump the frontend-angular-dependencies group across 1 directory with 13 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14329](https://github.com/paperless-ngx/paperless-ngx/pull/14329))
- Chore(deps-dev): Bump prettier from 3.9.8 to 3.9.9 in /src-ui @[dependabot[bot]](https://github.com/apps/dependabot) ([#14331](https://github.com/paperless-ngx/paperless-ngx/pull/14331))
- Chore(deps-dev): Bump the frontend-eslint-dependencies group across 1 directory with 3 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14330](https://github.com/paperless-ngx/paperless-ngx/pull/14330))
- Fix: ship pdf.js CMaps so CJK documents render in the viewer [@MrOggy85](https://github.com/MrOggy85) ([#14318](https://github.com/paperless-ngx/paperless-ngx/pull/14318))
- Chore(deps-dev): Bump zensical from 0.0.64 to 0.0.65 in the development group @[dependabot[bot]](https://github.com/apps/dependabot) ([#14326](https://github.com/paperless-ngx/paperless-ngx/pull/14326))
- Enhancement: more control over suggestion requests [@shamoon](https://github.com/shamoon) ([#14258](https://github.com/paperless-ngx/paperless-ngx/pull/14258))
- Chore(deps): Bump the uv group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14314](https://github.com/paperless-ngx/paperless-ngx/pull/14314))
- Chore: anchor admin url pattern [@shamoon](https://github.com/shamoon) ([#14316](https://github.com/paperless-ngx/paperless-ngx/pull/14316))
- Chorehancement: set manifest CORS for credentials [@shamoon](https://github.com/shamoon) ([#14307](https://github.com/paperless-ngx/paperless-ngx/pull/14307))
- Chore(deps): Bump the utilities-minor group across 1 directory with 7 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14305](https://github.com/paperless-ngx/paperless-ngx/pull/14305))
- Chore(deps): Bump the utilities-patch group across 1 directory with 6 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14297](https://github.com/paperless-ngx/paperless-ngx/pull/14297))
- Chore(deps-dev): Bump the frontend-jest-dependencies group across 1 directory with 2 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14286](https://github.com/paperless-ngx/paperless-ngx/pull/14286))
- Chore(deps-dev): Bump eslint from 10.10.0 to 10.11.0 in /src-ui in the frontend-eslint-dependencies group across 1 directory @[dependabot[bot]](https://github.com/apps/dependabot) ([#14287](https://github.com/paperless-ngx/paperless-ngx/pull/14287))
- Chore(deps-dev): Bump @types/node from 26.5.0 to 26.6.2 in /src-ui @[dependabot[bot]](https://github.com/apps/dependabot) ([#14288](https://github.com/paperless-ngx/paperless-ngx/pull/14288))
- Chore(deps-dev): Bump prettier from 3.9.6 to 3.9.8 in /src-ui @[dependabot[bot]](https://github.com/apps/dependabot) ([#14289](https://github.com/paperless-ngx/paperless-ngx/pull/14289))
- Chore(deps): Bump the frontend-angular-dependencies group across 1 directory with 10 updates @[dependabot[bot]](https://github.com/apps/dependabot) ([#14285](https://github.com/paperless-ngx/paperless-ngx/pull/14285))
- Fix: use version page\_count for versioned document [@shamoon](https://github.com/shamoon) ([#14280](https://github.com/paperless-ngx/paperless-ngx/pull/14280))
- Chore: replace bleach with turbohtml [@gaborbernat](https://github.com/gaborbernat) ([#14269](https://github.com/paperless-ngx/paperless-ngx/pull/14269))
- Enhancement: include Django admin with 2FA [@shamoon](https://github.com/shamoon) ([#14270](https://github.com/paperless-ngx/paperless-ngx/pull/14270))
- Fix: allow pointer events for pdf links in pngx viewer [@shamoon](https://github.com/shamoon) ([#14264](https://github.com/paperless-ngx/paperless-ngx/pull/14264))
- Feature: propagate resolved secrets to interactive container shells [@stumpylog](https://github.com/stumpylog) ([#14254](https://github.com/paperless-ngx/paperless-ngx/pull/14254))
- Fix: During a move to the trash directory, only attempt to copy metadata [@stumpylog](https://github.com/stumpylog) ([#14250](https://github.com/paperless-ngx/paperless-ngx/pull/14250))
- Fix: convert file mtime to the configured time zone directly [@stumpylog](https://github.com/stumpylog) ([#14249](https://github.com/paperless-ngx/paperless-ngx/pull/14249))
- Fix: ensure documentDeleted subscription is discarded [@shamoon](https://github.com/shamoon) ([#14247](https://github.com/paperless-ngx/paperless-ngx/pull/14247))
- Chore: Fix bugs in the test suite [@stumpylog](https://github.com/stumpylog) ([#14244](https://github.com/paperless-ngx/paperless-ngx/pull/14244))
- Fix: ensure bulk operations are checked against version root [@shamoon](https://github.com/shamoon) ([#14246](https://github.com/paperless-ngx/paperless-ngx/pull/14246))
- Enhancement: support separate embedding API key [@furkanural](https://github.com/furkanural) ([#14067](https://github.com/paperless-ngx/paperless-ngx/pull/14067))
- Enhancement: support passthrough extra params for LLMs [@shamoon](https://github.com/shamoon) ([#14202](https://github.com/paperless-ngx/paperless-ngx/pull/14202))
- Chore(deps): Bump autobahn from 25.12.2 to 26.7.1 in the uv group across 1 directory @[dependabot[bot]](https://github.com/apps/dependabot) ([#14231](https://github.com/paperless-ngx/paperless-ngx/pull/14231))
- Fix: indexing after document-added workflows signal [@shamoon](https://github.com/shamoon) ([#14242](https://github.com/paperless-ngx/paperless-ngx/pull/14242))
- Fix: Record full tag and custom field lists in bulk edit audit log [@stumpylog](https://github.com/stumpylog) ([#14236](https://github.com/paperless-ngx/paperless-ngx/pull/14236))
- Chore: update pikepdf for ocrmypdf requirement [@shamoon](https://github.com/shamoon) ([#14235](https://github.com/paperless-ngx/paperless-ngx/pull/14235))
- Fix: handle legacy bulk edit split page range with missing page\_count [@shamoon](https://github.com/shamoon) ([#14212](https://github.com/paperless-ngx/paperless-ngx/pull/14212))
- Fix: ignore invalid EXIF orientation when generating image archives [@zhzy0077](https://github.com/zhzy0077) ([#14203](https://github.com/paperless-ngx/paperless-ngx/pull/14203))
-
-</details>
-
 ## paperless-ngx 3.2.1

 ### Bug Fixes
@@ -76,9 +76,7 @@ is not supported by any of the available parsers.

 **A:** Not by default. As of v3, a file whose contents match an existing document is still
 consumed, and the duplicate is flagged in the UI — open the document and check the
-**Duplicates** tab to review documents that share the same content, or filter the document
-list by **Duplicates** to find all of them (see
-[Duplicate documents](usage.md#duplicate-documents)). If you prefer the old
+**Duplicates** tab to review documents that share the same content. If you prefer the old
 behavior of rejecting duplicates during consumption, set
 [`PAPERLESS_CONSUMER_DELETE_DUPLICATES`](configuration.md#PAPERLESS_CONSUMER_DELETE_DUPLICATES)
 to `true`.
@@ -12,10 +12,10 @@ physical documents into a searchable online archive so you can keep, well, _less
 [Get started](setup.md){ .md-button .md-button--primary .index-callout }
 [Demo](https://demo.paperless-ngx.com){ .md-button .md-button--secondary target=\_blank }

-<div class="demo-sponsor">
-  <span>Demo hosting by</span>
-  <a href="https://fyrastack.com" target="_blank" rel="noopener">
-    <img src="assets/sponsors/fyra-stack.svg" alt="" class="no-lightbox" width="20" height="20"> Fyra Stack
+<div style="display: flex; justify-content: end; margin-top: -1.5rem;">
+  <a href="https://m.do.co/c/8d70b916d462" target="_blank">
+    <img src="https://opensource.nyc3.cdn.digitaloceanspaces.com/attribution/assets/PoweredByDO/DO_Powered_by_Badge_white.svg#only-dark" class="no-lightbox" width="150px">
+    <img src="https://opensource.nyc3.cdn.digitaloceanspaces.com/attribution/assets/PoweredByDO/DO_Powered_by_Badge_black.svg#only-light" class="no-lightbox" width="150px">
  </a>
 </div>

@@ -129,7 +129,7 @@ change in the `docker-compose.yml` file:
 command:
  - 'gotenberg'
  - '--chromium-disable-javascript=true'
-  - '--chromium-allow-list=^file:///tmp/.*'
+  - '--chromium-allow-list=file:///tmp/.*'
  - '--api-timeout=60s'
 ```

@@ -272,65 +272,6 @@ This error can occur in installations which have upgraded from a version of Pape
 $ python3 manage.py convert_mariadb_uuid
 ```

-## MariaDB/MySQL error "Illegal mix of collations"
-
-Consumption or other operations fail with an error like:
-
-```
-(1267, "Illegal mix of collations (utf8mb4_general_ci,IMPLICIT) and
-(utf8mb4_unicode_ci,IMPLICIT) for operation '='")
-```
-
-This happens when the tables in your database do not all use the same
-collation. It is most often seen on databases that existed before a
-MariaDB/MySQL upgrade: older tables keep their original collation, while
-tables created afterwards use the new server default.
-
-To work around it, set the collation your existing tables use (the one in the error
-that is not `utf8mb4_unicode_ci`) with
-[`PAPERLESS_DB_OPTIONS`](configuration.md#PAPERLESS_DB_OPTIONS):
-
-```bash
-PAPERLESS_DB_OPTIONS="collation=utf8mb4_general_ci"
-```
-
-To fix it permanently, back up your database, then convert the database and
-each table to a single collation and remove the override:
-
-```sql
-ALTER DATABASE paperless CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci;
-ALTER TABLE <table_name> CONVERT TO CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci;
-```
-
-## PostgreSQL warns about a "collation version mismatch"
-
-The PostgreSQL log shows a warning like:
-
-```
-WARNING:  database "paperless" has a collation version mismatch
-DETAIL:  The database was created using collation version 2.36, but the operating system provides version 2.41.
-HINT:  Rebuild all objects in this database that use the default collation and run ALTER DATABASE paperless REFRESH COLLATION VERSION, or build PostgreSQL with the right library version.
-```
-
-This comes from PostgreSQL, not Paperless-ngx. The `glibc` version that PostgreSQL
-runs against changed, and the existing database was created with an older one. With
-Docker, `glibc` comes from the PostgreSQL image's Debian base, not the host, so this
-commonly happens when a new image is pulled after its base Debian release changed.
-If PostgreSQL is installed directly on a host, it uses the host's `glibc` instead.
-
-The warning is not an error and Paperless-ngx keeps working, but the database's text
-indexes may be built with outdated sorting rules. To resolve it, back up your
-database, then connect to it (for example with `psql -U paperless -d paperless`
-inside the database container) and run:
-
-```sql
-REINDEX DATABASE paperless;
-ALTER DATABASE paperless REFRESH COLLATION VERSION;
-```
-
-To avoid this in the future, pin your PostgreSQL image to a specific Debian release,
-for example `postgres:18-trixie`, rather than `postgres:18`.
-
 ## Platform-Specific Deployment Troubleshooting

 A user-maintained wiki page is available to help troubleshoot issues that may arise when trying to deploy Paperless-ngx on specific platforms, for example SELinux. Please see [the wiki](https://github.com/paperless-ngx/paperless-ngx/wiki/Platform%E2%80%90Specific-Troubleshooting).
@@ -299,18 +299,19 @@ for details.
 ### Duplicate documents

 By default, Paperless-ngx **does not reject duplicates**. If you consume a file whose
-contents match an existing document (same original or archive checksum), the new copy is
-still consumed and a warning is logged.
+contents exactly match an existing document (same checksum), the new copy is still
+consumed and a warning is logged. The task entry for the upload also flags that a
+duplicate was detected and links to the existing document(s).

-When a document has duplicates, a **Duplicates** tab appears on its detail page, listing
-the other documents you can view that share the same content (including any in the trash).
-To find all documents with duplicates, choose **Duplicates** in the document list's text
-filter dropdown, or use `has_duplicates=true` in the REST API.
+To review duplicates, open a document and switch to the **Duplicates** tab on the
+document detail page. It lists other documents that share the same content, including any
+that are in the trash (shown with a badge), and links to each so you can decide which to
+keep.

 If you would rather reject duplicates at consumption time (the pre-v3 behavior), set
 [`PAPERLESS_CONSUMER_DELETE_DUPLICATES`](configuration.md#PAPERLESS_CONSUMER_DELETE_DUPLICATES)
 to `true`. The duplicate file is then deleted instead of consumed, and the task fails with
-a "Document already exists" message linking to the existing document.
+a "document already exists" message.

 ## Document Suggestions

@@ -1,6 +1,6 @@
 [project]
 name = "paperless-ngx"
-version = "3.3.3"
+version = "3.2.1"
 description = """\
  A community-supported supercharged document management system: scan, index and archive all your physical documents\
  """
@@ -43,7 +43,7 @@ dependencies = [
  "drf-spectacular~=0.30",
  "drf-spectacular-sidecar>=2026.7.1,<2026.10",
  "drf-writable-nested~=0.7.1",
-  "filelock>=3.32,<4.1",
+  "filelock~=3.32.0",
  "flower>=2.0.1,<2.3",
  "gotenberg-client[httpx]~=1.0",
  "httpx-oauth~=0.17",
@@ -2,5 +2,5 @@

 docker run -p 5432:5432 -e POSTGRES_PASSWORD=password -v paperless_pgdata:/var/lib/postgresql/data -d postgres:15
 docker run -d -p 6379:6379 redis:latest
-docker run -p 3000:3000 -d gotenberg/gotenberg:7.8 gotenberg --chromium-disable-javascript=true --chromium-allow-list="^file:///tmp/.*"
+docker run -p 3000:3000 -d gotenberg/gotenberg:7.8 gotenberg --chromium-disable-javascript=true --chromium-allow-list="file:///tmp/.*"
 docker run -p 9998:9998 -d docker.io/apache/tika:latest
@@ -15,8 +15,6 @@ module.exports = {
        tsconfig: '<rootDir>/tsconfig.spec.json',
        stringifyContentPathRegex: '\\.(html|svg)$',
        useESM: true,
-        // Disable the builder's deprecated override; tsconfig.spec.json enables isolation.
-        isolatedModules: false,
      },
    ],
  },
@@ -1,6 +1,6 @@
 {
  "name": "paperless-ngx-ui",
-  "version": "3.3.3",
+  "version": "3.2.1",
  "scripts": {
    "preinstall": "npx only-allow pnpm",
    "ng": "ng",
@@ -102,34 +102,25 @@
    </button>
  }

-  @if (selectedSection !== TaskSection.All && totalTasks() > pageSize) {
-    <ngb-pagination
-      [pageSize]="pageSize"
-      [collectionSize]="totalTasks()"
-      [page]="page"
-      [maxSize]="5"
-      [rotate]="true"
-      size="sm"
-      aria-label="Tasks pagination"
-      (pageChange)="setPage($event)">
-    </ngb-pagination>
-  }
+  <ngb-pagination
+    [pageSize]="pageSize"
+    [collectionSize]="totalTasks()"
+    [page]="page"
+    [maxSize]="5"
+    [rotate]="true"
+    size="sm"
+    aria-label="Tasks pagination"
+    (pageChange)="setPage($event)">
+  </ngb-pagination>
 </div>

 <ng-template let-tasks="tasks" let-section="section" #tasksTemplate>
-  <div class="section-header mb-2">
-    <h5 class="mb-0">{{ sectionLabel(section) }}</h5>
-    @let count = sectionCount(section);
-    @let hasMore = selectedSection === TaskSection.All && count > tasks.length;
-    <div class="small text-muted">
-      @if (hasMore) {
-        <ng-container i18n>{{tasks.length}} of {{count}} tasks</ng-container>
-        <button class="btn btn-link show-all p-0 ms-2 align-baseline" type="button" (click)="setSection(section)">
-          <ng-container i18n>Show all</ng-container><i-bs class="ms-1" width="1em" height="1em" name="arrow-right"></i-bs>
-        </button>
-      } @else {
-        <ng-container i18n>{count, plural, =1 {1 task} other {{{count}} tasks}}</ng-container>
-      }
+  <div class="section-header d-flex align-items-center justify-content-between mb-2">
+    <div>
+      <h5 class="mb-0">{{ sectionLabel(section) }}</h5>
+      <div class="small text-muted">
+        <ng-container i18n>{tasks.length, plural, =1 {1 task} other {{{tasks.length}} tasks}}</ng-container>
+      </div>
    </div>
  </div>

@@ -178,9 +169,10 @@
              <div>{{ taskDisplayName(task) }}</div>
              <div class="small text-muted">
                @if (taskShowsSeparateTypeLabel(task)) {
-                  <span class="me-2">{{ task.task_type_display }}</span>
+                  <span>{{ task.task_type_display }}</span>
+                  <span class="mx-1">&bull;</span>
                }
-                <span class="text-nowrap"><i-bs class="me-1" width="1em" height="1em" name="box-arrow-in-right"></i-bs>{{ task.trigger_source_display }}</span>
+                <span>{{ task.trigger_source_display }}</span>
              </div>
            </td>
            <td class="d-none d-lg-table-cell created-col">{{ task.date_created | customDate:'short' }}</td>
@@ -72,7 +72,3 @@ pre {
 tbody tr:nth-last-child(2) td {
    border-bottom: none !important;
 }
-
-.section-header .show-all {
-    font-size: inherit;
-}
@@ -223,6 +223,16 @@ describe('TasksComponent', () => {
      )
      .flush(paginatedTasks)

+    httpTestingController
+      .expectOne(
+        (req) =>
+          req.url === `${environment.apiBaseUrl}tasks/` &&
+          req.params.get('acknowledged') === 'false' &&
+          req.params.get('page_size') === '25' &&
+          req.params.get('page') === '1'
+      )
+      .flush(paginatedTasks)
+
    httpTestingController
      .expectOne(
        (req) =>
@@ -231,30 +241,8 @@ describe('TasksComponent', () => {
          !req.params.has('status')
      )
      .flush(sectionCountResponse)
-
-    flushTaskLists(10)
  })

-  // answers task list requests with the fixture tasks matching their status filter
-  function flushTaskLists(pageSize: number, count?: number) {
-    httpTestingController
-      .match(
-        (req) =>
-          req.url === `${environment.apiBaseUrl}tasks/` &&
-          req.params.get('page_size') === `${pageSize}`
-      )
-      .forEach((req) => {
-        const statuses = req.request.params.getAll('status')
-        const results = tasks.filter((task) => statuses.includes(task.status))
-        req.flush({ count: count ?? results.length, results })
-      })
-  }
-
-  function selectSection(section: TaskSection, count: number) {
-    component.setSection(section)
-    flushTaskLists(25, count)
-  }
-
  it('should display task sections with counts', () => {
    expect(component.selectedSection).toBe(TaskSection.All)
    expect(component.selectedTaskType).toBeNull()
@@ -350,7 +338,6 @@ describe('TasksComponent', () => {
  })

  it('should render pagination controls next to the task filter', () => {
-    selectSection(TaskSection.Completed, 30)
    fixture.detectChanges()

    const controls = fixture.debugElement.query(By.css('.task-controls'))
@@ -361,51 +348,6 @@ describe('TasksComponent', () => {
    expect(pagination).not.toBeNull()
  })

-  it('should only paginate a section with more than one page of tasks', () => {
-    const pagination = () =>
-      fixture.debugElement.query(By.css('ngb-pagination'))
-
-    fixture.detectChanges()
-    expect(pagination()).toBeNull()
-
-    selectSection(TaskSection.Completed, 2)
-    fixture.detectChanges()
-    expect(pagination()).toBeNull()
-  })
-
-  it('should load the all tasks overview one section at a time', () => {
-    expect(component.tasksForSection(TaskSection.NeedsAttention)).toHaveLength(
-      2
-    )
-    expect(component.tasksForSection(TaskSection.InProgress)).toHaveLength(3)
-    expect(component.tasksForSection(TaskSection.Completed)).toHaveLength(2)
-    expect(component.totalTasks()).toBe(7)
-  })
-
-  it('should show section totals and link truncated sections to their full list', () => {
-    component.sectionCounts.update((counts) => ({
-      ...counts,
-      [TaskSection.Completed]: 27,
-    }))
-    fixture.detectChanges()
-
-    const header = (label: string) =>
-      fixture.debugElement
-        .queryAll(By.css('.section-header'))
-        .find((h) => h.nativeElement.textContent.includes(label))
-
-    expect(header('Needs attention').nativeElement.textContent).toContain(
-      '2 tasks'
-    )
-    expect(header('Needs attention').query(By.css('button'))).toBeNull()
-
-    expect(header('Recently completed').nativeElement.textContent).toContain(
-      '2 of 27 tasks'
-    )
-    header('Recently completed').query(By.css('button')).nativeElement.click()
-    expect(component.selectedSection).toBe(TaskSection.Completed)
-  })
-
  it('should apply the selected section to the server-side task query', () => {
    component.setSection(TaskSection.NeedsAttention)

@@ -426,56 +368,55 @@ describe('TasksComponent', () => {
  it('should apply task type and trigger source filters to the server-side task query', () => {
    component.setTaskType(PaperlessTaskType.SanityCheck)

-    expect(
-      httpTestingController.match(
+    httpTestingController
+      .expectOne(
        (request) =>
          request.url === `${environment.apiBaseUrl}tasks/` &&
-          request.params.get('page_size') === '10' &&
+          request.params.get('page_size') === '25' &&
          request.params.get('task_type') === PaperlessTaskType.SanityCheck
      )
-    ).toHaveLength(3)
+      .flush({ count: 1, results: [tasks[6]] })

    component.setTriggerSource(PaperlessTaskTriggerSource.System)

-    expect(
-      httpTestingController.match(
+    httpTestingController
+      .expectOne(
        (request) =>
          request.url === `${environment.apiBaseUrl}tasks/` &&
-          request.params.get('page_size') === '10' &&
+          request.params.get('page_size') === '25' &&
          request.params.get('task_type') === PaperlessTaskType.SanityCheck &&
          request.params.get('trigger_source') ===
            PaperlessTaskTriggerSource.System
      )
-    ).toHaveLength(3)
+      .flush({ count: 1, results: [tasks[6]] })
  })

  it('should apply text filters to the server-side task query', () => {
    component.filterText = 'invoice'
    jest.advanceTimersByTime(150)

-    expect(
-      httpTestingController.match(
+    httpTestingController
+      .expectOne(
        (request) =>
          request.url === `${environment.apiBaseUrl}tasks/` &&
-          request.params.get('page_size') === '10' &&
+          request.params.get('page_size') === '25' &&
          request.params.get('name') === 'invoice'
      )
-    ).toHaveLength(3)
+      .flush({ count: 1, results: [tasks[0]] })

    component.setFilterTarget(TaskFilterTargetID.Result)

-    expect(
-      httpTestingController.match(
+    httpTestingController
+      .expectOne(
        (request) =>
          request.url === `${environment.apiBaseUrl}tasks/` &&
-          request.params.get('page_size') === '10' &&
+          request.params.get('page_size') === '25' &&
          request.params.get('result') === 'invoice'
      )
-    ).toHaveLength(3)
+      .flush({ count: 0, results: [] })
  })

  it('should load a different task page when pagination changes', () => {
-    selectSection(TaskSection.Completed, 30)
    component.setPage(2)

    const pageTwoTasks = {
@@ -499,7 +440,6 @@ describe('TasksComponent', () => {
  })

  it('should not replace section counts with current-page counts', () => {
-    selectSection(TaskSection.Completed, 30)
    component.setPage(2)

    httpTestingController
@@ -515,21 +455,11 @@ describe('TasksComponent', () => {
        results: [tasks[0]],
      })

-    expect(component.sectionCount(TaskSection.All)).toBe(7)
    expect(component.sectionCount(TaskSection.NeedsAttention)).toBe(2)
    expect(component.sectionCount(TaskSection.InProgress)).toBe(3)
    expect(component.sectionCount(TaskSection.Completed)).toBe(2)
  })

-  it('should stop loading when the overview fails to load', () => {
-    component.setSection(TaskSection.All)
-    httpTestingController
-      .match((req) => req.params.get('page_size') === '10')[0]
-      .flush('', { status: 500, statusText: 'Error' })
-
-    expect(component.loading()).toBe(false)
-  })
-
  it('should expose stable task type options and disable empty ones', () => {
    expect(component.taskTypeOptions.map((option) => option.value)).toContain(
      PaperlessTaskType.TrainClassifier
@@ -15,9 +15,6 @@ import {
  distinctUntilChanged,
  filter,
  first,
-  forkJoin,
-  map,
-  Observable,
  Subject,
  takeUntil,
  timer,
@@ -28,7 +25,6 @@ import {
  PaperlessTaskTriggerSource,
  PaperlessTaskType,
 } from 'src/app/data/paperless-task'
-import { Results } from 'src/app/data/results'
 import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { CustomDatePipe } from 'src/app/pipes/custom-date.pipe'
 import { TasksService } from 'src/app/services/tasks.service'
@@ -173,7 +169,6 @@ export class TasksComponent
  public expandedTask: number
  public autoRefreshEnabled: boolean = true
  public readonly pageSize = 25
-  public readonly overviewSize = 10
  public page: number = 1
  readonly totalTasks = signal(0)
  readonly sectionCounts = signal<Record<TaskSection, number>>({
@@ -489,6 +484,13 @@ export class TasksComponent
    return this.sectionCounts()[section]
  }

+  private setSectionCount(section: TaskSection, count: number) {
+    this.sectionCounts.update((counts) => ({
+      ...counts,
+      [section]: count,
+    }))
+  }
+
  sectionShowsResults(section: TaskSection): boolean {
    return section !== TaskSection.InProgress
  }
@@ -735,12 +737,21 @@ export class TasksComponent
    this.reloadSectionCounts()

    this.loading.set(true)
-    this.listTasks()
+    this.tasksService
+      .list(
+        this.page,
+        this.pageSize,
+        this.getParamsForSection(this.selectedSection)
+      )
      .pipe(first(), takeUntil(this.unsubscribeNotifier))
      .subscribe({
        next: (result) => {
          this.pagedTasks.set(result.results)
          this.totalTasks.set(result.count)
+          this.setSectionCount(TaskSection.All, result.count)
+          if (this.selectedSection !== TaskSection.All) {
+            this.setSectionCount(this.selectedSection, result.count)
+          }
          this.loading.set(false)
          if (
            this.page > 1 &&
@@ -756,30 +767,4 @@ export class TasksComponent
        },
      })
  }
-
-  private listTasks(): Observable<Results<PaperlessTask>> {
-    if (this.selectedSection !== TaskSection.All) {
-      return this.tasksService.list(
-        this.page,
-        this.pageSize,
-        this.getParamsForSection(this.selectedSection)
-      )
-    }
-
-    // All is an overview of each section to avoid splitting sections across pages
-    return forkJoin(
-      this.sections.map((section) =>
-        this.tasksService.list(
-          1,
-          this.overviewSize,
-          this.getParamsForSection(section)
-        )
-      )
-    ).pipe(
-      map((results) => ({
-        count: results.reduce((sum, result) => sum + result.count, 0),
-        results: results.flatMap((result) => result.results),
-      }))
-    )
-  }
 }
@@ -154,31 +154,6 @@ describe('UsersAndGroupsComponent', () => {
    jest.useRealTimers()
  })

-  it('should logout current user under base href if password changed', () => {
-    jest.useFakeTimers()
-    const base = document.createElement('base')
-    base.href = '/paperless/'
-    document.head.appendChild(base)
-    completeSetup()
-    let modal: NgbModalRef
-    modalService.activeInstances.subscribe((refs) => (modal = refs[0]))
-    component.editUser(users[0])
-    const navSpy = jest
-      .spyOn(navUtils, 'setLocationHref')
-      .mockImplementation(() => {})
-    const editDialog = modal.componentInstance as UserEditDialogComponent
-    editDialog.passwordIsSet = true
-    settingsService.currentUser.set(users[0]) // simulate logged in as same user
-    editDialog.succeeded.emit(users[0])
-    fixture.detectChanges()
-    jest.advanceTimersByTime(2600)
-    expect(navSpy).toHaveBeenCalledWith(
-      `${window.location.origin}/paperless/accounts/logout/?next=/paperless/accounts/login/?next=/paperless/`
-    )
-    base.remove()
-    jest.useRealTimers()
-  })
-
  it('should support edit / create group, show error if needed', () => {
    completeSetup()
    let modal: NgbModalRef
@@ -116,9 +116,8 @@ export class UsersAndGroupsComponent
            $localize`Password has been changed, you will be logged out momentarily.`
          )
          setTimeout(() => {
-            const basePath = new URL(document.baseURI).pathname
            setLocationHref(
-              `${window.location.origin}${basePath}accounts/logout/?next=${basePath}accounts/login/?next=${basePath}`
+              `${window.location.origin}/accounts/logout/?next=/accounts/login/?next=/`
            )
          }, 2500)
        } else {
@@ -66,7 +66,6 @@
 }

 .search-container {
-  min-width: 0;
  padding-left: 1rem;
  padding-right: 1rem;
 }
@@ -500,11 +499,6 @@ main {

 // true center the search with equal flex widths
@media (min-width: 768px) {
-  .search-container {
-    // Prevent input padding and the conditional search button from resizing
-    flex-basis: 55rem;
-  }
-
  .navbar-brand-container,
  .navbar > ul {
    display: flex; // so the brand link stays as wide as its contents
@@ -70,17 +70,6 @@ form {
  max-height: 75vh;
 }

-.dropdown.show {
-  // Hide the search icon when the dropdown is shown
-  i-bs[name="search"] {
-    display: none !important;
-  }
-
-  .form-control {
-    padding-left: 0.5rem;
-  }
-}
-
 .dropdown-item {
  &:has(button:focus) {
    background-color: var(--pngx-bg-darker);
@@ -218,38 +218,6 @@ describe('ProfileEditDialogComponent', () => {
    jest.useRealTimers()
  })

-  it('should logout under base href on save if password changed', () => {
-    jest.useFakeTimers()
-    const base = document.createElement('base')
-    base.href = '/paperless/'
-    document.head.appendChild(base)
-    const getSpy = jest.spyOn(profileService, 'get')
-    getSpy.mockReturnValue(of(profile))
-    const getProvidersSpy = jest.spyOn(
-      profileService,
-      'getSocialAccountProviders'
-    )
-    getProvidersSpy.mockReturnValue(of(socialAccountProviders))
-    component.ngOnInit()
-    component['newPassword'] = 'new*pass'
-    component.form.get('password').patchValue('new*pass')
-    component.form.get('password_confirm').patchValue('new*pass')
-
-    const updateSpy = jest.spyOn(profileService, 'update')
-    updateSpy.mockReturnValue(of(null))
-    const navSpy = jest
-      .spyOn(navUtils, 'setLocationHref')
-      .mockImplementation(() => {})
-    component.save()
-    expect(updateSpy).toHaveBeenCalled()
-    jest.advanceTimersByTime(2600)
-    expect(navSpy).toHaveBeenCalledWith(
-      `${window.location.origin}/paperless/accounts/logout/?next=/paperless/accounts/login/?next=/paperless/`
-    )
-    base.remove()
-    jest.useRealTimers()
-  })
-
  it('should support auth token copy', () => {
    jest.useFakeTimers()
    const getSpy = jest.spyOn(profileService, 'get')
@@ -210,9 +210,8 @@ export class ProfileEditDialogComponent
              $localize`Password has been changed, you will be logged out momentarily.`
            )
            setTimeout(() => {
-              const basePath = new URL(document.baseURI).pathname
              setLocationHref(
-                `${window.location.origin}${basePath}accounts/logout/?next=${basePath}accounts/login/?next=${basePath}`
+                `${window.location.origin}/accounts/logout/?next=/accounts/login/?next=/`
              )
            }, 2500)
          }
@@ -1,6 +1,6 @@
 <div class="d-flex align-items-center">
  <div class="btn-group">
-    <button type="button" class="btn btn-sm btn-outline-primary" (click)="clickSuggest()" [disabled]="disabled() || loading()" [aria-label]="noSuggestions ? 'No suggestions' : 'Suggest'" i18n-aria-label>
+    <button type="button" class="btn btn-sm btn-outline-primary" (click)="clickSuggest()" [disabled]="disabled() || loading() || (suggestions() && !aiEnabled())" [aria-label]="noSuggestions ? 'No suggestions' : 'Suggest'" i18n-aria-label>
      @if (loading()) {
        <div class="spinner-border spinner-border-sm" role="status"></div>
      } @else if (noSuggestions) {
@@ -18,44 +18,46 @@
      }
    </button>

-    <div class="btn-group" ngbDropdown #dropdown="ngbDropdown" [popperOptions]="popperOptions">
-      <button type="button" class="btn btn-sm btn-outline-primary" ngbDropdownToggle [disabled]="disabled() || loading() || !suggestions()" aria-expanded="false" aria-controls="suggestionsDropdown" aria-label="Suggestions dropdown">
-        <span class="visually-hidden" i18n>Show suggestions</span>
-      </button>
+    @if (aiEnabled()) {
+      <div class="btn-group" ngbDropdown #dropdown="ngbDropdown" [popperOptions]="popperOptions">
+        <button type="button" class="btn btn-sm btn-outline-primary" ngbDropdownToggle [disabled]="disabled() || loading() || !suggestions()" aria-expanded="false" aria-controls="suggestionsDropdown" aria-label="Suggestions dropdown">
+          <span class="visually-hidden" i18n>Show suggestions</span>
+        </button>

-      <div ngbDropdownMenu aria-labelledby="suggestionsDropdown" class="shadow suggestions-dropdown">
-        <div class="list-group list-group-flush small pb-0">
-          @if (novelSuggestions === 0 && fieldSuggestions === 0) {
-            <div class="list-group-item text-muted fst-italic">
-              <small class="text-muted small fst-italic" i18n>No novel suggestions</small>
-            </div>
-          }
-          @if (suggestions()?.suggested_tags?.length > 0) {
-            <small class="list-group-item text-uppercase text-muted small"><i-bs class="me-2" name="tags"></i-bs><ng-container i18n>Tags</ng-container></small>
-            @for (tag of suggestions().suggested_tags; track tag) {
-              <button type="button" class="list-group-item list-group-item-action bg-light" (click)="addTag.emit(tag)">{{ tag }}</button>
+        <div ngbDropdownMenu aria-labelledby="suggestionsDropdown" class="shadow suggestions-dropdown">
+          <div class="list-group list-group-flush small pb-0">
+            @if (novelSuggestions === 0 && fieldSuggestions === 0) {
+              <div class="list-group-item text-muted fst-italic">
+                <small class="text-muted small fst-italic" i18n>No novel suggestions</small>
+              </div>
            }
-          }
-          @if (suggestions()?.suggested_document_types?.length > 0) {
-            <div class="list-group-item text-uppercase text-muted small"><i-bs class="me-2" name="hash"></i-bs><ng-container i18n>Document Types</ng-container></div>
-            @for (type of suggestions().suggested_document_types; track type) {
-              <button type="button" class="list-group-item list-group-item-action bg-light" (click)="addDocumentType.emit(type)">{{ type }}</button>
+            @if (suggestions()?.suggested_tags?.length > 0) {
+              <small class="list-group-item text-uppercase text-muted small"><i-bs class="me-2" name="tags"></i-bs><ng-container i18n>Tags</ng-container></small>
+              @for (tag of suggestions().suggested_tags; track tag) {
+                <button type="button" class="list-group-item list-group-item-action bg-light" (click)="addTag.emit(tag)">{{ tag }}</button>
+              }
            }
-          }
-          @if (suggestions()?.suggested_correspondents?.length > 0) {
-            <div class="list-group-item text-uppercase text-muted small"><i-bs class="me-2" name="person"></i-bs><ng-container i18n>Correspondents</ng-container></div>
-            @for (correspondent of suggestions().suggested_correspondents; track correspondent) {
-              <button type="button" class="list-group-item list-group-item-action bg-light" (click)="addCorrespondent.emit(correspondent)">{{ correspondent }}</button>
+            @if (suggestions()?.suggested_document_types?.length > 0) {
+              <div class="list-group-item text-uppercase text-muted small"><i-bs class="me-2" name="hash"></i-bs><ng-container i18n>Document Types</ng-container></div>
+              @for (type of suggestions().suggested_document_types; track type) {
+                <button type="button" class="list-group-item list-group-item-action bg-light" (click)="addDocumentType.emit(type)">{{ type }}</button>
+              }
            }
-          }
-          @if (fieldSuggestions > 0) {
-            <div class="list-group-item text-muted fst-italic">
-              <small class="text-muted small fst-italic" i18n>{fieldSuggestions, plural, =1 {1 suggestion available below} other {{{fieldSuggestions}} suggestions available below}}</small>
-            </div>
-          }
+            @if (suggestions()?.suggested_correspondents?.length > 0) {
+              <div class="list-group-item text-uppercase text-muted small"><i-bs class="me-2" name="person"></i-bs><ng-container i18n>Correspondents</ng-container></div>
+              @for (correspondent of suggestions().suggested_correspondents; track correspondent) {
+                <button type="button" class="list-group-item list-group-item-action bg-light" (click)="addCorrespondent.emit(correspondent)">{{ correspondent }}</button>
+              }
+            }
+            @if (fieldSuggestions > 0) {
+              <div class="list-group-item text-muted fst-italic">
+                <small class="text-muted small fst-italic" i18n>{fieldSuggestions, plural, =1 {1 suggestion available below} other {{{fieldSuggestions}} suggestions available below}}</small>
+              </div>
+            }
+          </div>
        </div>
      </div>
-    </div>
+    }
  </div>

  @if (aiEnabled()) {
@@ -251,37 +251,6 @@ describe('SuggestionsDropdownComponent', () => {
    expect(fixture.nativeElement.textContent).toContain('No novel suggestions')
  })

-  it('should open ML date suggestions with AI disabled', () => {
-    const getSuggestions = jest.spyOn(component.getSuggestions, 'emit')
-    fixture.componentRef.setInput('aiEnabled', false)
-    fixture.componentRef.setInput('fetchedSources', [SuggestionSource.ML])
-    fixture.componentRef.setInput('suggestions', {
-      dates: ['2026-01-04', '2026-02-01'],
-    })
-    fixture.componentRef.setInput('appliedCreated', '2026-01-04')
-    fixture.detectChanges()
-
-    const button = fixture.nativeElement.querySelector('button')
-    expect(button.disabled).toBeFalsy()
-    expect(button.querySelector('.badge').textContent.trim()).toBe('1')
-
-    button.click()
-    fixture.detectChanges()
-
-    expect(component.dropdown.isOpen()).toBeTruthy()
-    expect(fixture.nativeElement.textContent).toContain(
-      '1 suggestion available below'
-    )
-    expect(getSuggestions).not.toHaveBeenCalled()
-    expect(
-      fixture.nativeElement.querySelector('button[title="Suggestion options"]')
-    ).toBeNull()
-
-    button.click()
-    fixture.detectChanges()
-    expect(component.dropdown.isOpen()).toBeFalsy()
-  })
-
  it('should fetch unfetched sources and show existing suggestions', () => {
    jest.spyOn(component.getSuggestions, 'emit')
    fixture.componentRef.setInput('aiEnabled', true)
@@ -53,7 +53,11 @@ export class SuggestionsDropdownComponent {
  addCorrespondent: EventEmitter<string> = new EventEmitter()

  public clickSuggest(): void {
-    if (this.disabled() || this.loading()) {
+    if (
+      this.disabled() ||
+      this.loading() ||
+      (this.suggestions() && !this.aiEnabled())
+    ) {
      return
    }

@@ -25,8 +25,6 @@
            @if (displayFields().includes(field)) {
              <th
                scope="col"
-                class="text-truncate"
-                [title]="getColumnTitle(field)"
                [ngClass]="{
                  'd-none d-md-table-cell': i > 1,
                  'w-25': field === DisplayField.CREATED || field === DisplayField.ADDED
@@ -484,22 +484,6 @@ describe('FilterEditorComponent', () => {
    expect(component.textFilter).toBeNull()
  })

-  it('should ingest title filter rules alongside a fulltext query with date created', () => {
-    component.filterRules = [
-      {
-        rule_type: FILTER_SIMPLE_TITLE,
-        value: 'foo',
-      },
-      {
-        rule_type: FILTER_FULLTEXT_QUERY,
-        value: 'created:[-1 week to now]',
-      },
-    ]
-    expect(component.dateCreatedRelativeDate).toEqual(1) // RELATIVE_DATE_QUERYSTRINGS['-1 week to now']
-    expect(component.textFilter).toEqual('foo')
-    expect(component.textFilterTarget).toEqual('title') // TEXT_FILTER_TARGET_TITLE
-  })
-
  it('should ingest text filter rules for fulltext query that include date added', () => {
    expect(component.dateAddedRelativeDate).toBeNull()
    component.filterRules = [
@@ -1861,7 +1845,7 @@ describe('FilterEditorComponent', () => {
    ])
  })

-  it('should keep text filtering separate from date created relative date', () => {
+  it('should carry over text filtering on date created with relative date', () => {
    component.textFilter = 'foo'
    const dateCreatedDropdown = fixture.debugElement.queryAll(
      By.directive(DatesDropdownComponent)
@@ -1871,18 +1855,14 @@ describe('FilterEditorComponent', () => {
    fixture.detectChanges()
    tick(400)
    expect(component.filterRules).toEqual([
-      {
-        rule_type: FILTER_SIMPLE_TEXT,
-        value: 'foo',
-      },
      {
        rule_type: FILTER_FULLTEXT_QUERY,
-        value: 'created:[-1 week to now]',
+        value: 'foo,created:[-1 week to now]',
      },
    ])
  })

-  it('should keep text filtering separate from created and added relative dates', () => {
+  it('should carry over text filtering once with created and added relative dates', () => {
    component.textFilter = 'foo'
    const datesDropdown = fixture.debugElement.query(
      By.directive(DatesDropdownComponent)
@@ -1893,18 +1873,14 @@ describe('FilterEditorComponent', () => {
    fixture.detectChanges()
    tick(400)
    expect(component.filterRules).toEqual([
-      {
-        rule_type: FILTER_SIMPLE_TEXT,
-        value: 'foo',
-      },
      {
        rule_type: FILTER_FULLTEXT_QUERY,
-        value: 'created:[-1 week to now],added:[-1 month to now]',
+        value: 'foo,created:[-1 week to now],added:[-1 month to now]',
      },
    ])
  })

-  it('should keep legacy title filters as a title search when adding a created relative date', () => {
+  it('should convert legacy title filters into full text query when adding a created relative date', () => {
    component.filterRules = [
      {
        rule_type: FILTER_TITLE,
@@ -1919,18 +1895,14 @@ describe('FilterEditorComponent', () => {
    fixture.detectChanges()
    tick(400)
    expect(component.filterRules).toEqual([
-      {
-        rule_type: FILTER_SIMPLE_TITLE,
-        value: 'foo',
-      },
      {
        rule_type: FILTER_FULLTEXT_QUERY,
-        value: 'created:[-1 week to now]',
+        value: 'foo,created:[-1 week to now]',
      },
    ])
  })

-  it('should keep simple title filters as a title search when adding a created relative date', () => {
+  it('should convert simple title filters into full text query when adding a created relative date', () => {
    component.filterRules = [
      {
        rule_type: FILTER_SIMPLE_TITLE,
@@ -1945,13 +1917,9 @@ describe('FilterEditorComponent', () => {
    fixture.detectChanges()
    tick(400)
    expect(component.filterRules).toEqual([
-      {
-        rule_type: FILTER_SIMPLE_TITLE,
-        value: 'foo',
-      },
      {
        rule_type: FILTER_FULLTEXT_QUERY,
-        value: 'created:[-1 week to now]',
+        value: 'foo,created:[-1 week to now]',
      },
    ])
  })
@@ -2037,7 +2005,7 @@ describe('FilterEditorComponent', () => {
    ])
  })

-  it('should keep text filtering separate from date added relative date', () => {
+  it('should carry over text filtering on date added with relative date', () => {
    component.textFilter = 'foo'
    const datesDropdown = fixture.debugElement.query(
      By.directive(DatesDropdownComponent)
@@ -2047,13 +2015,9 @@ describe('FilterEditorComponent', () => {
    fixture.detectChanges()
    tick(400)
    expect(component.filterRules).toEqual([
-      {
-        rule_type: FILTER_SIMPLE_TEXT,
-        value: 'foo',
-      },
      {
        rule_type: FILTER_FULLTEXT_QUERY,
-        value: 'added:[-1 week to now]',
+        value: 'foo,added:[-1 week to now]',
      },
    ])
  })
@@ -1083,6 +1083,23 @@ export class FilterEditorComponent
        (fr) => fr.rule_type == FILTER_FULLTEXT_QUERY
      )

+      // if had a title / content search and added a relative date we need to carry it over...
+      if (
+        !existingRule &&
+        this._textFilter?.length > 0 &&
+        (this.textFilterTarget == TEXT_FILTER_TARGET_TITLE_CONTENT ||
+          this.textFilterTarget == TEXT_FILTER_TARGET_TITLE)
+      ) {
+        existingRule = filterRules.find(
+          (fr) =>
+            fr.rule_type == FILTER_TITLE_CONTENT ||
+            fr.rule_type == FILTER_SIMPLE_TEXT ||
+            fr.rule_type == FILTER_TITLE ||
+            fr.rule_type == FILTER_SIMPLE_TITLE
+        )
+        existingRule.rule_type = FILTER_FULLTEXT_QUERY
+      }
+
      let queryArgs = existingRule?.value.split(',') ?? []
      if (this.dateCreatedRelativeDate !== null) {
        const rd = RELATIVE_DATE_QUERYSTRINGS.find(
@@ -11,7 +11,7 @@
      <div class="input-group input-group-sm w-auto d-none d-md-flex">
        <label class="input-group-text bg-transparent border-0" for="managementPageSize" i18n>Show:</label>
      </div>
-      <div class="input-group input-group-sm w-auto">
+      <div class="input-group input-group-sm w-auto me-3">
        <select id="managementPageSize" class="form-select form-select-sm small" [(ngModel)]="pageSize">
          <option [ngValue]="25">25</option>
          <option [ngValue]="50">50</option>
@@ -19,9 +19,7 @@
        </select>
        <span class="input-group-text text-muted d-none d-md-flex" i18n>per page</span>
      </div>
-      @if (collectionSize() > pageSize) {
-        <ngb-pagination [pageSize]="pageSize" [collectionSize]="collectionSize()" [page]="page()" [maxSize]="5" (pageChange)="page.set($event); reloadData()" class="ms-3" size="sm" aria-label="Pagination"></ngb-pagination>
-      }
+      <ngb-pagination [pageSize]="pageSize" [collectionSize]="collectionSize()" [page]="page()" [maxSize]="5" (pageChange)="page.set($event); reloadData()" size="sm" aria-label="Pagination"></ngb-pagination>
    </div>
  </div>

@@ -72,7 +70,7 @@
        }
      </div>
    }
-    @if (collectionSize() > pageSize) {
+    @if (collectionSize() > 20) {
      <ngb-pagination class="ms-auto" [pageSize]="pageSize" [collectionSize]="collectionSize()" [page]="page()" [maxSize]="5" (pageChange)="page.set($event); reloadData()" size="sm" aria-label="Pagination"></ngb-pagination>
    }
  </div>
@@ -93,17 +93,15 @@
        [disabled]="selectedMailIds().size === 0"
        (confirm)="deleteSelected()">
      </pngx-confirm-button>
-      @if (processedMails().length > 50) {
-        <div class="ms-auto">
-          <ngb-pagination
-            [collectionSize]="processedMails().length"
-            [(page)]="page"
-            [pageSize]="50"
-            [maxSize]="5"
-            (pageChange)="loadProcessedMails()">
-          </ngb-pagination>
-        </div>
-      }
+      <div class="ms-auto">
+        <ngb-pagination
+          [collectionSize]="processedMails().length"
+          [(page)]="page"
+          [pageSize]="50"
+          [maxSize]="5"
+          (pageChange)="loadProcessedMails()">
+        </ngb-pagination>
+      </div>
    </div>
  }
 </div>
@@ -8,7 +8,7 @@ export const environment = {
  apiVersion: '10', // match src/paperless/settings.py
  appTitle: DEFAULT_APP_TITLE,
  tag: 'prod',
-  version: '3.3.3',
+  version: '3.2.1',
  webSocketHost: window.location.host,
  webSocketProtocol: window.location.protocol == 'https:' ? 'wss:' : 'ws:',
  webSocketBaseUrl: base_url.pathname + 'ws/',
@@ -41,7 +41,6 @@ import {
  bell,
  bodyText,
  bookmark,
-  boxArrowInRight,
  boxArrowUp,
  boxArrowUpRight,
  boxes,
@@ -294,7 +293,6 @@ const icons = {
  braces,
  bodyText,
  bookmark,
-  boxArrowInRight,
  boxArrowUp,
  boxArrowUpRight,
  boxes,
@@ -15,7 +15,7 @@
    "target": "ES2022",
    "module": "es2020",
    "lib": [
-      "es2022",
+      "es2020",
      "dom"
    ],
    "useDefineForClassFields": false
@@ -54,10 +54,6 @@ def _predict_with_threshold(classifier, X, threshold: float) -> int | None:
    cutoff so that uncertain predictions are discarded rather than assigned.
    """
    probas = classifier.predict_proba(X)[0]
-    if len(classifier.classes_) == 1:
-        # scikit-learn treats a single-class fit as binary, so predict_proba
-        # returns [1 - p, p] for one class. Column 0 is the real class.
-        probas = probas[:1]
    best_idx = int(probas.argmax())
    best_class = int(classifier.classes_[best_idx])

@@ -130,14 +130,12 @@ class DocumentMetadataOverrides:
            get_users_with_perms(
                doc,
                only_with_perms_in=["view_document"],
-                with_group_users=False,
            ).values_list("id", flat=True),
        )
        overrides.change_users = list(
            get_users_with_perms(
                doc,
                only_with_perms_in=["change_document"],
-                with_group_users=False,
            ).values_list("id", flat=True),
        )
        overrides.custom_fields = {
@@ -50,7 +50,6 @@ from documents.models import ShareLink
 from documents.models import ShareLinkBundle
 from documents.models import StoragePath
 from documents.models import Tag
-from documents.permissions import annotate_authorizing_fields
 from documents.permissions import permitted_document_ids
 from documents.permissions import permitted_object_ids
 from documents.versioning import annotate_effective_content
@@ -1076,9 +1075,6 @@ class PermittedObjectsFilter(BaseFilterBackend):

    include_granted: bool = True
    perm_codename: str | None = None
-    # A self-referencing foreign key whose target authorizes a row, so a row is
-    # judged by its parent's owner and grants (``Document.root_document``).
-    parent_field: str | None = None

    def filter_queryset(self, request, queryset, view):
        # Before the superuser and owner-only paths, neither of which consults
@@ -1090,23 +1086,11 @@ class PermittedObjectsFilter(BaseFilterBackend):
        if request.user.is_superuser:
            return queryset
        if not self.include_granted:
-            owner_field = "owner"
-            if self.parent_field is not None:
-                queryset = annotate_authorizing_fields(queryset, self.parent_field)
-                owner_field = "authorizing_owner"
-            return queryset.filter(
-                Q(**{owner_field: request.user.pk})
-                | Q(**{f"{owner_field}__isnull": True}),
-            )
+            return queryset.filter(Q(owner=request.user) | Q(owner__isnull=True))
        model = queryset.model
        perm = self.perm_codename or f"view_{model._meta.model_name}"
        return queryset.filter(
-            id__in=permitted_object_ids(
-                request.user,
-                model,
-                perm,
-                parent_field=self.parent_field,
-            ),
+            id__in=permitted_object_ids(request.user, model, perm),
        )


@@ -3,7 +3,6 @@ from itertools import islice
 from pathlib import Path
 from typing import TYPE_CHECKING
 from typing import Any
-from typing import Final

 from allauth.mfa.models import Authenticator
 from allauth.socialaccount.models import SocialAccount
@@ -71,15 +70,6 @@ from paperless.models import ApplicationConfiguration
 from paperless_mail.models import MailAccount
 from paperless_mail.models import MailRule

-# Derived names replace the original's extension with these suffixes
-THUMBNAIL_SUFFIX: Final[str] = "-thumbnail.webp"
-ARCHIVE_SUFFIX: Final[str] = "-archive.pdf"
-# Leave room for the longest derived suffix, which replaces the original's extension
-MAX_BASE_NAME_BYTES: Final[int] = Document.MAX_PUBLIC_FILENAME_BYTES - max(
-    len(THUMBNAIL_SUFFIX),
-    len(ARCHIVE_SUFFIX),
-)
-

 def serialize_queryset_batched(
    queryset: "QuerySet[Any]",
@@ -522,16 +512,10 @@ class Command(CryptMixin, PaperlessCommand):
                    counter=filename_counter,
                )
            else:
-                base_name = document.get_public_filename(
-                    counter=filename_counter,
-                    max_bytes=MAX_BASE_NAME_BYTES,
-                )
+                base_name = document.get_public_filename(counter=filename_counter)

-            # The thumbnail and archive names are derived from the stem, so
-            # uniqueness must hold without the original's extension
-            unique_key = str(Path(base_name).with_suffix(""))
-            if unique_key not in self.exported_files:
-                self.exported_files.add(unique_key)
+            if base_name not in self.exported_files:
+                self.exported_files.add(base_name)
                break
            else:
                filename_counter += 1
@@ -554,7 +538,7 @@ class Command(CryptMixin, PaperlessCommand):
        document_dict[EXPORTER_FILE_NAME] = original_arc

        if not self.no_thumbnail:
-            thumbnail_name = base_name.parent / (base_name.stem + THUMBNAIL_SUFFIX)
+            thumbnail_name = base_name.parent / (base_name.stem + "-thumbnail.webp")
            if self.use_folder_prefix:
                thumbnail_name = Path("thumbnails") / thumbnail_name
            thumbnail_arc = thumbnail_name.as_posix()
@@ -563,7 +547,7 @@ class Command(CryptMixin, PaperlessCommand):
            thumbnail_arc = None

        if not self.no_archive and document.has_archive_version:
-            archive_name = base_name.parent / (base_name.stem + ARCHIVE_SUFFIX)
+            archive_name = base_name.parent / (base_name.stem + "-archive.pdf")
            if self.use_folder_prefix:
                archive_name = Path("archive") / archive_name
            archive_arc = archive_name.as_posix()
@@ -66,7 +66,7 @@ def iter_manifest_records(path: Path) -> Generator[dict, None, None]:
    """Yield records one at a time from a manifest JSON array via ijson."""
    try:
        with path.open("rb") as f:
-            yield from ijson.items(f, "item", use_float=True)
+            yield from ijson.items(f, "item")
    except ijson.JSONError as e:
        raise CommandError(f"Failed to parse manifest file {path}: {e}") from e

@@ -157,8 +157,6 @@ class StoragePath(MatchingModel):

 class Document(SoftDeleteModel, ModelWithOwner):  # type: ignore[django-manager-missing]
    MAX_STORED_FILENAME_LENGTH: Final[int] = 1024
-    # Per-component limit of common filesystems (ext4, XFS, btrfs, ZFS, APFS)
-    MAX_PUBLIC_FILENAME_BYTES: Final[int] = 255

    correspondent = models.ForeignKey(
        Correspondent,
@@ -486,37 +484,28 @@ class Document(SoftDeleteModel, ModelWithOwner):  # type: ignore[django-manager-
    def archive_file(self):
        return Path(self.archive_path).open("rb")

-    def get_public_filename(
-        self,
-        *,
-        archive: bool = False,
-        counter: int = 0,
-        max_bytes: int = MAX_PUBLIC_FILENAME_BYTES,
-    ) -> str:
+    def get_public_filename(self, *, archive=False, counter=0, suffix=None) -> str:
        """
        Returns a sanitized filename for the document, not including any paths.
-
-        The name fits in max_bytes (UTF-8). A long title is shortened before
-        the counter and extension are appended, because sanitize_filename
-        truncates from the end. The tail is cut only when max_bytes is
-        smaller than the tail itself.
        """
        # Root owns metadata for all versions
        context_document = (
            self.root_document if self.root_document_id is not None else self
        )
+        result = str(context_document)

-        extension = ".pdf" if archive else self.file_type
-        tail = f"_{counter:02}{extension}" if counter else extension
+        if counter:
+            result += f"_{counter:02}"

-        head_bytes = max(0, max_bytes - len(tail.encode()))
-        head = str(context_document).encode()[:head_bytes].decode(errors="ignore")
+        if suffix:
+            result += suffix

-        return pathvalidate.sanitize_filename(
-            head + tail,
-            replacement_text="-",
-            max_len=max_bytes,
-        )
+        if archive:
+            result += ".pdf"
+        else:
+            result += self.file_type
+
+        return pathvalidate.sanitize_filename(result, replacement_text="-")

    @property
    def file_type(self):
@@ -8,9 +8,11 @@ from django.contrib.contenttypes.models import ContentType
 from django.db.models import Case
 from django.db.models import CharField
 from django.db.models import Count
+from django.db.models import Exists
 from django.db.models import F
 from django.db.models import IntegerField
 from django.db.models import Model
+from django.db.models import OuterRef
 from django.db.models import Q
 from django.db.models import QuerySet
 from django.db.models import Value
@@ -367,18 +369,14 @@ def permitted_object_ids(
    visible exactly when its parent is, judged by the parent's owner and
    grants, so the row's own owner and grants are ignored.

-    Guardian stores ``object_pk`` as a string, so the row key is cast to a
-    string and tested against the user's and groups' grants with a single
-    uncorrelated ``IN``. Postgres and SQLite build that set once. MariaDB
-    evaluates it as an index probe per row, which is cheap because the
-    lookups use guardian's unique indexes. Casting every ``object_pk`` to an
-    integer instead cannot use an index, and MariaDB cannot materialize it
-    inside the owner ``OR``, so it re-scans the user's grants for every row.
-    A correlated ``EXISTS`` per grant fixes MariaDB too, but Postgres and
-    SQLite re-run it for every row and end up slower than the original. The
-    user's groups are matched with an ``IN`` subquery rather than a join
-    through the membership table, which SQLite plans badly once the grant
-    tables grow.
+    Guardian stores ``object_pk`` as a string, so each grant is an ``EXISTS``
+    keyed on the row's id cast to a string, which can use guardian's
+    (user, permission, object_pk) unique index. Casting every ``object_pk`` to
+    an integer for an ``id IN (...)`` is not indexable, and MariaDB cannot
+    materialize it inside the owner ``OR``, so it re-scans the user's grants
+    for every row. The user's groups are matched with an ``IN`` subquery
+    rather than a join through the membership table, which SQLite plans badly
+    once the grant tables grow.
    """
    has_soft_delete = hasattr(model, "global_objects")
    manager = (
@@ -389,7 +387,14 @@ def permitted_object_ids(
    owner_field, key_field = "owner", "pk"
    if parent_field is not None:
        owner_field, key_field = "authorizing_owner", "authorizing_id"
-        base_qs = annotate_authorizing_fields(base_qs, parent_field)
+        base_qs = base_qs.annotate(
+            authorizing_id=Coalesce(f"{parent_field}_id", "id"),
+            authorizing_owner=Case(
+                When(**{f"{parent_field}_id__isnull": True}, then=F("owner_id")),
+                default=F(f"{parent_field}__owner_id"),
+                output_field=IntegerField(),
+            ),
+        )
    unowned = Q(**{f"{owner_field}__isnull": True})

    if user is None or not getattr(user, "is_authenticated", False):
@@ -417,51 +422,30 @@ def permitted_object_ids(
        "permission__content_type": content_type,
    }

-    # Both grant sets are compared to the row key as strings, exactly as
-    # guardian stores them, and are uncorrelated, so each engine can build the
-    # set once instead of probing per row.
-    user_keys = UserObjectPermission.objects.filter(
-        user=user,
-        **perm_filter,
-    ).values_list("object_pk", flat=True)
-    group_keys = GroupObjectPermission.objects.filter(
-        group_id__in=user.groups.values("id"),
-        **perm_filter,
-    ).values_list("object_pk", flat=True)
-    permitted_keys = user_keys.union(group_keys, all=True)
-
-    return (
-        base_qs.annotate(permitted_key=Cast(key_field, CharField(max_length=64)))
-        .filter(
-            Q(**{owner_field: user.pk}) | unowned | Q(permitted_key__in=permitted_keys),
-        )
-        .values_list("id", flat=True)
+    key_as_text = Cast(OuterRef(key_field), CharField(max_length=64))
+    granted_to_user = Exists(
+        UserObjectPermission.objects.filter(
+            user=user,
+            object_pk=key_as_text,
+            **perm_filter,
+        ),
    )
+    granted_to_group = Exists(
+        GroupObjectPermission.objects.filter(
+            group_id__in=user.groups.values("id"),
+            object_pk=key_as_text,
+            **perm_filter,
+        ),
+    )
+
+    return base_qs.filter(
+        Q(**{owner_field: user.pk}) | unowned | granted_to_user | granted_to_group,
+    ).values_list("id", flat=True)


 ModelT = TypeVar("ModelT", bound=Model)


-def annotate_authorizing_fields(
-    queryset: QuerySet[ModelT],
-    parent_field: str,
-) -> QuerySet[ModelT]:
-    """
-    Annotate each row with ``authorizing_id`` and ``authorizing_owner``: the id
-    and owner of the row that authorizes it. A row with a parent (the
-    self-referencing foreign key ``parent_field``) is authorized by its parent,
-    any other row by itself.
-    """
-    return queryset.annotate(
-        authorizing_id=Coalesce(f"{parent_field}_id", "id"),
-        authorizing_owner=Case(
-            When(**{f"{parent_field}_id__isnull": True}, then=F("owner_id")),
-            default=F(f"{parent_field}__owner_id"),
-            output_field=IntegerField(),
-        ),
-    )
-
-
 def user_is_unrestricted(user: User | None) -> bool:
    """
    True when ``user`` means "no restriction at all" (an absent user, or an
@@ -511,28 +495,6 @@ def restrict_queryset_to_visible(
    return queryset.filter(pk__in=permitted_object_ids(user, queryset.model, perm))


-def user_is_active_superuser(user: User | None) -> bool:
-    """
-    True only for an active, authenticated superuser.
-    """
-    return user is not None and user_is_unrestricted(user)
-
-
-def restrict_queryset_to_owner_visible(
-    queryset: QuerySet[ModelT],
-    user: User | None,
-    perm: str,
-) -> QuerySet[ModelT]:
-    """
-    Like ``restrict_queryset_to_visible``, but ``user=None`` sees only unowned
-    rows (``permitted_object_ids(None, ...)`` semantics) instead of every row.
-    See ``user_is_active_superuser``.
-    """
-    if user_is_active_superuser(user):
-        return queryset
-    return queryset.filter(pk__in=permitted_object_ids(user, queryset.model, perm))
-
-
 def permitted_document_ids(
    user: User | None,
    *,
@@ -926,7 +926,6 @@ class TantivyBackend:
        sort_field: str | None = None,
        sort_reverse: bool = False,
        search_mode: SearchMode = SearchMode.QUERY,
-        filter_query: str | None = None,
        limit: int | None = None,
    ) -> list[int]:
        """
@@ -942,7 +941,6 @@ class TantivyBackend:
            sort_field: Field to sort by (None for relevance ranking)
            sort_reverse: Whether to reverse the sort order
            search_mode: Query parsing mode (QUERY, TEXT, or TITLE)
-            filter_query: Optional advanced query results must also match
            limit: Maximum number of IDs to return (None = all matching docs)

        Returns:
@@ -950,13 +948,6 @@ class TantivyBackend:
        """
        self._ensure_open()
        user_query = self._parse_query(query, search_mode)
-        if filter_query is not None:
-            user_query = user_query.and_must_match(
-                tantivy.Query.const_score_query(
-                    self._parse_query(filter_query, SearchMode.QUERY),
-                    0.0,  # zero so doesn't affect relevance
-                ),
-            )
        final_query = self._apply_permission_filter(user_query, user)

        searcher = self._index.searcher()
@@ -1052,21 +1043,16 @@ class TantivyBackend:
        excluded from results.

        Args:
-            doc_id: Primary key of the reference document, or of one of its
-                versions
+            doc_id: Primary key of the reference document
            user: User for permission filtering (None for no filtering)
            limit: Maximum number of IDs to return (None = all matching docs)

        Returns:
            List of similar document IDs (excluding the original)
        """
-        from documents.versioning import root_document_ids
-
        self._ensure_open()
        searcher = self._index.searcher()

-        # Only root documents are indexed, so a version stands for its root
-        doc_id = next(iter(root_document_ids([doc_id])), doc_id)
        id_query = tantivy.Query.term_query(self._schema, "id", doc_id)
        results = searcher.search(id_query, limit=1)

@@ -28,9 +28,7 @@ logger = logging.getLogger("paperless.search")
 #      columns dropped. tantivy compares schemas by ordered field list, so an
 #      index built by v1 rejects every write against the v2 schema.
 # v3 - barcodes JSON field for stored barcode contents
-# v4 - root documents only. Earlier indexes may hold document versions under
-#      their own id and metadata, so they are rebuilt without them.
-SCHEMA_VERSION: Final[int] = 4
+SCHEMA_VERSION: Final[int] = 3

 # Present in the index directory from the moment a full rebuild starts until it
 # finishes. If a rebuild is interrupted it is left behind, so the half-built
Author	SHA1	Message	Date
stumpylog	642511c7b9	Fix: use a root document's effective content for the LLM index and prompts The search index and the document views use the newest version's content as a root document's content, but everything the LLM side reads used the root's own content: the text embedded in the LLM index, the content in the classification prompt, the context blocks from similar documents, and the text that is searched for similar documents. After a document was replaced by a new version the LLM kept answering from the old text. Read get_effective_content() in those four places. The index update annotates the content in the query like the search index does, and the similar-document lookup annotates it for the documents it fetches, so no query is made per document. Entries already in the LLM index keep the old text until the document next changes or the index is rebuilt.	2026-10-08 15:22:03 -07:00
stumpylog	210d97a522	Fix: index document versions as their root document The search index and the LLM index hold root documents only, a root being indexed with its newest version's content. Every write path therefore had to remember to hand them the root. Several did not: adding or deleting a note on a version, restoring a trashed document together with its versions, and reprocessing a version all wrote the version into the index under its own id, where it could be returned as a separate search result. The paths that walk the whole library did the same: the document_index reindex command put every version in the search index, a full LLM index rebuild embedded every version, and an incremental LLM update scoped to a version id indexed it under its own id. Asking for documents like a version looked the version's own id up in the index and silently found nothing. WriteBatch.add_or_update, WriteBatch.add_or_update_ids and llm_index_add_or_update_document now resolve a version to its root themselves. The reindex command and update_llm_index only walk root documents, and an incremental LLM update is scoped to the roots of the given ids through versioning.root_document_ids, which add_or_update_ids shares. more_like_id returns the root's id for a version. The reprocess task no longer picks the root for the indexes and only still clears the caches of both documents.	2026-10-08 15:22:02 -07:00
stumpylog	8a96190359	Fix: authorize document versions by their root in the single-object permission check has_perms_owner_aware judged a document by its own owner and grants, so each endpoint that fetches a document itself had to remember to map a version to its root document first, and one that forgot, like the more-like-this search filter, authorized by a stale version owner. The check now maps a Document to its root before looking at the owner and the guardian grants, matching what permitted_document_ids does for id sets. The eight call sites that mapped the document themselves pass it straight through. The DRF object permission class needs no change because the document viewset only ever serves root documents.	2026-10-08 15:21:45 -07:00
stumpylog	ece8769f59	Fix: authorize document versions by their root and speed up permission id sets permitted_document_ids judged a version by its own owner and grants, so a version whose owner had drifted from its root's was visible to the wrong people and hidden from the right ones. Callers patched this individually by mapping each document to its root first. The query itself was also slow on MariaDB: the guardian grants were a UNION cast to integers and tested with IN inside an OR with the owner checks, which MariaDB cannot materialize, so it re-scans the user's grants for every document. At 20k documents that took seconds for a user with a couple of hundred grants. permitted_object_ids now looks grants up as an EXISTS keyed on the row id cast to a string, which uses guardian's unique index, and matches the user's groups with an IN subquery. It takes an optional parent_field naming a self-referencing foreign key whose target authorizes the row, and permitted_document_ids passes root_document, so a version is visible exactly when its root is. The helper that mapped documents to their roots at the call sites is no longer needed, so the email, selection data, share link bundle, trash, bulk download and bulk edit checks use the id set directly.	2026-10-08 15:21:28 -07:00
				`@@ -1 +0,0 @@`
				<svg xmlns="http://www.w3.org/2000/svg" width="666.667" height="666.667" viewBox="0 0 666.667 666.667"><g stroke-width="1.333"><path fill="#fef400" d="m332.66668 418.66668h-84.66667v-85.33334-85.33333h84.66667 84.66667v-85.33334-85.333334h85.33333 85.33334v85.333334 85.33334h-85.33334-85.33333v85.33333 85.33334z"/><path fill="#fe6565" d="m248.00001 589.33335h-170.666674v-170.66667-170.66667h85.333334 85.33334v-85.33334-85.333334h84.66667 84.66667v85.333334 85.33334h85.33333 85.33334v85.33333 85.33334h-84.66667-84.66667v85.33334 85.33333zm84.66667-170.66667h84.66667v-85.33334-85.33333h-84.66667-84.66667v85.33333 85.33334z"/></g></svg>