UNDO feature - easily track changes

Nicely track changes in the UI:

What's Changed

• Feature/undo by @ccrlawrence in #54

Full Changelog: v0.5.1...v0.6.0

v0.5.1

What's Changed

• bugfix/46-OCRonOpenAI by @ccrlawrence in #47

Full Changelog: v0.5.0...v0.5.1

v0.5.0

v0.5.0 Release Notes

We are thrilled to announce the release of Paperless GPT v0.5.0! This version brings a groundbreaking new feature—OCR powered by Large Language Models (LLMs)—along with several enhancements and bug fixes to improve your experience.

---

🚀 New Features

OCR via LLMs (Experimental)

• OCR via LLM by @icereed in #29

  Introducing an experimental feature that utilizes LLMs to perform Optical Character Recognition (OCR) on your documents. This advanced OCR method can provide more accurate text extraction, especially in cases where traditional OCR struggles.

  • Access the Feature: Enable the feature by specifying VISION_LLM_PROVIDER and VISION_LLM_MODEL as environment variables.
  • Processing Time: Please note that OCR via LLMs may take several minutes to process, depending on the document size and your hardware.
  • Feedback Welcome: As this is an early prototype, your feedback is invaluable for further improvements.

---

🔧 Improvements

• Improve Auto Throughput & Logging by @JonasHess in #40

  • Enhanced the automatic processing pipeline to increase throughput.
  • Improved logging for better monitoring and debugging.

---

🐛 Bug Fixes

• Tag Query Change by @ccrlawrence in #44

  • Improved tag querying to handle edge cases more effectively.
  • Enhanced document retrieval accuracy based on tags.

---

📢 Notes

• Experimental Feature: The OCR via LLMs is an experimental feature available to all users. Processing times may vary, and results might differ based on document quality and content.
• Feedback Encouraged: We invite all users to try out the new OCR feature and share their feedback to help us refine and improve it.
• Acknowledgments: A big thank you to all contributors for their hard work and dedication in bringing this release to life.

---

Upgrade to v0.5.0 today and experience the future of OCR with Paperless GPT!

New Contributors

• @ccrlawrence made their first contribution in #44

Full Changelog: v0.4.0...v0.5.0

Auto-Tagging and fix for custom prompts

What's Changed

• Fix GH Action for PRs from forks by @icereed in #23
• Specific error messages for missing environment variables by @JonasHess in #21
• Feature/exponential backoff for auto processing by @JonasHess in #26
• Fix prompts mount by @icereed in #30

New Contributors

• @JonasHess made their first contribution in #21

Full Changelog: v0.3.0...v0.4.0

Customizable Server-Side Prompts

📢 Release Notes: paperless-gpt v0.3.0

We are excited to announce the release of paperless-gpt v0.3.0! This version brings significant enhancements, including customizable server-side prompts and improvements to tag handling. These updates provide greater flexibility and control over how you generate document titles and tags using AI.

🚀 New Features and Improvements

1. Customizable Server-Side Prompts

• User-Defined Prompt Templates: You can now modify the prompt templates used for generating document titles and tags directly on the server side.

  • Prompt Templates: The application uses Go's text/template for the prompt templates.

  • Persistent Storage: Templates are stored in the prompts directory, allowing you to customize them by editing the files.

  • Template Variables: Use variables like {{.Language}}, {{.Content}}, {{.AvailableTags}}, and {{.Title}} within your templates to tailor the prompts.

  • Advanced Functions: Leverage functions from the Sprig library for more advanced templating capabilities.

• Easy Integration: Simply edit the template files to adjust how the AI generates suggestions, making the system more adaptable to your specific needs.

2. Improved Tag Handling

• API Pagination Fix: Resolved an issue where not all tags were fetched from paperless-ngx due to pagination limitations.

  • Comprehensive Tag Fetching: The application now correctly handles API pagination, ensuring that all available tags are retrieved and considered during tag suggestion generation.

• Enhanced Accuracy: With all tags available, the AI can suggest more accurate and relevant tags for your documents.

🛠 How to Update

To update to the latest version of paperless-gpt, please pull the latest Docker image:

docker pull icereed/paperless-gpt:latest

If you're using Docker Compose, update your docker-compose.yml file if necessary and redeploy the stack:

docker-compose pull
docker-compose up -d

🔧 Custom Prompt Templates Setup

Mounting the Prompts Directory

To enable custom prompt templates, you need to mount a local prompts directory into the container.

Docker Compose Example:

services:
  paperless-gpt:
    image: icereed/paperless-gpt:latest
    # ... (other configurations)
    volumes:
      - ./prompts:/app/prompts # Mount the prompts directory

Docker Run Command Example:

docker run -d \
  # ... (other configurations)
  -v $(pwd)/prompts:/app/prompts \
  paperless-gpt

Editing the Prompt Templates

1. Create the prompts Directory Locally:

   mkdir prompts

2. Start the Container:

   When you first start the container with the prompts directory mounted, it will automatically create the default template files in your local prompts directory if they do not exist.

3. Edit the Template Files:

   • Open prompts/title_prompt.tmpl and prompts/tag_prompt.tmpl with your favorite text editor.
   • Modify the templates using Go's text/template syntax.
   • Save the changes.

4. Restart the Container (if necessary):

   The application automatically reloads the templates when it starts. If the container is already running, you may need to restart it to apply the changes.

Template Syntax and Variables

• For title_prompt.tmpl:

  • {{.Language}}: The language specified in LLM_LANGUAGE (default is English).
  • {{.Content}}: The content of the document.

• For tag_prompt.tmpl:

  • {{.Language}}: The language specified in LLM_LANGUAGE.
  • {{.AvailableTags}}: A comma-separated list of available tags from paperless-ngx.
  • {{.Title}}: The suggested title for the document.
  • {{.Content}}: The content of the document.

Note: Advanced users can utilize additional functions from the Sprig template library, as it is included in the application.

📚 Documentation

For detailed instructions on installation, configuration, and customizing prompt templates, please refer to the README on our GitHub repository.

💡 Feedback and Contributions

We appreciate your feedback! If you encounter any issues or have suggestions for new features, please open an issue on our GitHub Issues page.

Contributions are welcome. Please read our contributing guidelines before submitting a pull request.

📦 What's Next

We are continuously working to improve paperless-gpt. Upcoming features and enhancements may include:

• UI Support for Prompt Editing: Providing an in-app interface to edit prompt templates without the need to access the file system.

• Manual Dark Mode Toggle: Allowing users to override the system's dark mode setting within the application.

• Additional LLM Support: Expanding compatibility with more Large Language Models.

• Performance Improvements: Optimizing the application for faster processing and better resource utilization.

---

Thank you for your continued support and for using paperless-gpt to enhance your paperless-ngx experience!

v0.2.0 – Major UI Redesign with Card-Based Layout and Dark Mode Support

📢 Release Notes: paperless-gpt v0.2.0

We are thrilled to announce the release of paperless-gpt v0.2.0! This version brings significant enhancements to the user interface and user experience, making it more intuitive and efficient to generate and manage document titles using AI.

🚀 New Features and Improvements

1. Redesigned User Interface with Card-Based Layout

• Card-Based Design: We've transitioned from a traditional table-based UI to a modern card-based layout. Each document is now presented as an individual card, providing a cleaner and more organized view.

• Enhanced Readability: The new design improves readability and allows for easier navigation through your documents.

• Responsive Layout: The card-based interface is fully responsive, ensuring a seamless experience across various devices and screen sizes.

2. Content Preview on Hover

• Hover Preview: You can now hover over document titles to get an instant preview of the content inside each document.

• Quick Access: This feature enables you to quickly glance at the document's content without opening it, saving you time and helping you make faster decisions.

• Interactive Experience: The hover effect adds an interactive layer to the UI, enhancing user engagement and efficiency.

3. Automatic Dark Mode Support

• System Dark Mode Integration: paperless-gpt now automatically adapts to your system's dark mode settings, providing a comfortable viewing experience in low-light environments.

  • macOS Compatibility: Fully compatible with macOS system settings for dark mode.

  • Windows and Linux Support: Works seamlessly with dark mode settings on Windows and compatible Linux distributions.

• No Manual Toggle Required: The application detects your system preference without the need for a manual switch.

  • Note: Currently, there is no in-app toggle to override the system's default dark mode setting. This feature may be considered for future releases based on user feedback.

🛠 How to Update

To update to the latest version of paperless-gpt, please pull the latest Docker image:

docker pull icereed/paperless-gpt:latest

If you're using Docker Compose, update your docker-compose.yml file if necessary and redeploy the stack:

docker-compose pull
docker-compose up -d

📚 Documentation

For detailed instructions on installation and configuration, please refer to the README on our GitHub repository.

💡 Feedback and Contributions

We appreciate your feedback! If you encounter any issues or have suggestions for new features, please open an issue on our GitHub Issues page.

Contributions are welcome. Please read our contributing guidelines before submitting a pull request.

📦 What's Next

We are continuously working to improve paperless-gpt. Upcoming features and enhancements may include:

• Manual Dark Mode Toggle: Providing an option within the application to override the system's dark mode setting.

• Additional LLM Support: Expanding compatibility with more Large Language Models.

• Performance Improvements: Optimizing the application for faster processing and better resource utilization.

---

Thank you for your continued support and for using paperless-gpt to enhance your paperless-ngx experience!

Tag generation support

Initial support for tags ✨

What's Changed

• Enhance UX for tag selection by @icereed in #5
• Add GitHub Actions workflow for building and pushing Docker images by @icereed in #6

Available with:

docker pull icereed/paperless-gpt:v0.1.0
# or
docker pull icereed/paperless-gpt:latest

Full Changelog: https://github.com/icereed/paperless-gpt/commits/v0.1.0