Merge pull request #7 from cloud-py-api/feat/visionatrix-1.9

version 1.1

CHANGELOG.md

@@ -2,18 +2,39 @@
 
 All notable changes to this project will be documented in this file.
 
+## [1.1.0 - 2024-12-18]
+
+The [Visionatrix](https://github.com/Visionatrix/Visionatrix) service has been updated from version `1.4.1` to `1.9.0`.
+The main changes are summarized below:
+
+### Added
+
+- Support for flows that produce `video` files as output.
+- `OpenAPI` specs support, enabling easier development of Nextcloud Apps that can seamlessly use Visionatrix flows.
+- Many new `flows` with updated models.
+- Parallel installation of `flows` and `models`.
+
+### Changed
+
+- Online `flows` and `models` storage has been separated, ensuring the integration remains stable.
+- Added support for both `Ollama` and `Gemini` as providers for `Vision` and `Prompt translation`.
+
+### Fixed
+
+- Numerous bug fixes and significant performance improvements.
+
 ## [1.0.1 - 2024-10-18]
 
 ### Changed
 
-- Bumped bundled Visionatrix from `1.3.0` to `1.4.1` version, changelog can be found [here](https://github.com/Visionatrix/Visionatrix/releases/tag/v1.4.0)
+- Updated the bundled Visionatrix from version `1.3.0` to `1.4.1`. The changelog for this update can be found [here](https://github.com/Visionatrix/Visionatrix/releases/tag/v1.4.0).
 
 ### Fixed
 
-- Warning "sudo: unable to resolve host" during container startup.
+- Resolved the warning "sudo: unable to resolve host" during container startup.
 
 ## [1.0.0 - 2024-09-25]
 
 ### Added
 
-- First release with bundled Visionatrix version `1.3.0`
+- Initial release with the bundled Visionatrix version `1.3.0`.

Dockerfile

@@ -1,21 +1,26 @@
-FROM python:3.10-slim
+FROM python:3.12-slim
 
 ARG BUILD_TYPE
 
 # Visionatrix enviroment variables
 ENV VIX_HOST="127.0.0.1"
 ENV VIX_PORT=8288
 ENV USER_BACKENDS="vix_db;nextcloud"
-ENV FLOWS_DIR="/nc_app_visionatrix_data/vix_flows"
 ENV MODELS_DIR="/nc_app_visionatrix_data/vix_models"
 ENV TASKS_FILES_DIR="/nc_app_visionatrix_data/vix_tasks_files"
-ENV BACKEND_DIR="/Visionatrix/vix_backend"
 ENV VIX_SERVER_FULL_MODELS="1"
 
 RUN apt-get update && apt-get install -y git \
 	python3-dev python3-setuptools netcat-traditional \
 	libxml2-dev libxslt1-dev zlib1g-dev g++ \
-	ffmpeg libsm6 libxext6 lsb-release sudo wget procps nano xmlstarlet
+	ffmpeg libsm6 libxext6 lsb-release sudo wget procps nano xmlstarlet && \
+    apt-get autoclean
+
+ADD ex_app_scripts/common_pgsql.sh /ex_app_scripts/
+RUN chmod +x /ex_app_scripts/common_pgsql.sh
+
+ADD ex_app_scripts/install_pgsql.sh /ex_app_scripts/
+RUN chmod +x /ex_app_scripts/install_pgsql.sh && /ex_app_scripts/install_pgsql.sh && rm /ex_app_scripts/install_pgsql.sh
 
 COPY appinfo/info.xml /info.xml
 
@@ -44,7 +49,12 @@ RUN cd /Visionatrix && \
 RUN cd /Visionatrix && \
     venv/bin/python -m pip install "psycopg[binary]" greenlet && \
     venv/bin/python -m pip install . && \
+	rm -rf ~/.cache/pip
+
+RUN cd /Visionatrix && \
 	venv/bin/python -m visionatrix install && \
+    AUTO_INIT_CONFIG_MODELS_DIR=$MODELS_DIR venv/bin/python scripts/easy_install.py && \
+    rm visionatrix.db && \
 	rm -rf ~/.cache/pip
 
 # Setup nodejs and npm for building the front-end client

Makefile

@@ -4,7 +4,7 @@ APP_ID := visionatrix
 APP_NAME := Visionatrix
 APP_VERSION := $$(xmlstarlet sel -t -v "//version" appinfo/info.xml)
 VISIONATRIX_VERSION := $$(xmlstarlet sel -t -v "//image-tag" appinfo/info.xml)
-JSON_INFO := "{\"id\":\"$(APP_ID)\",\"name\":\"$(APP_NAME)\",\"daemon_config_name\":\"manual_install\",\"version\":\"$(APP_VERSION)\",\"secret\":\"12345\",\"port\":9100, \"routes\": [{\"url\":\".*\",\"verb\":\"GET, POST, PUT, DELETE\",\"access_level\":1,\"headers_to_exclude\":[]}], \"translations_folder\":\"\/tmp\/vix_l10n\"}"
+JSON_INFO := "{\"id\":\"$(APP_ID)\",\"name\":\"$(APP_NAME)\",\"daemon_config_name\":\"manual_install\",\"version\":\"$(APP_VERSION)\",\"secret\":\"12345\",\"port\":24000, \"routes\": [{\"url\":\".*\",\"verb\":\"GET, POST, PUT, DELETE\",\"access_level\":1,\"headers_to_exclude\":[]}], \"translations_folder\":\"\/tmp\/vix_l10n\"}"
 
 
 .PHONY: help

README.md

@@ -16,6 +16,9 @@ This is a standalone [Visionatrix](https://github.com/Visionatrix/Visionatrix) s
 
 Each user is authenticated via Nextcloud credentials and has a personal task history.
 
+> [!NOTE]
+> We recommend starting with the "light-weight" flows, such as SDXL Lighting, Remove background (BiRefNet), and Remove background (Bria), to test your hardware.
+
 ## TextToImage Provider
 
 You can use Visionatrix as the TextToImage provider via [Nextcloud Assistant](https://github.com/nextcloud/assistant).
@@ -48,12 +51,12 @@ Once the Visionatrix flow is installed and supported, it will be registered as a
 
 We offer two types of installation:
 
-1. **Standard Installation**: This includes both the Visionatrix server for task management and a built-in worker for task processing. 
+1. **Standard Installation**: Includes both the Visionatrix server for task management and a built-in worker for task processing.
 
-	> For this installation, you'll need either a powerful CPU with at least 32 GB of RAM or a GPU with at least 10 GB of VRAM. 
-	> _(Support for 8 GB GPUs is planned)._
+	> For this installation, you'll need either a powerful CPU with at least 32 GB of RAM or a GPU with at least 8 GB of VRAM.
+	> _(Some flows, like background removal, are fast even on CPU and don’t require high-end hardware.)_
 
-2. **Advanced Installation**: Starting from Nextcloud version `30.0.2`, you can install only the server component by setting `DISABLE_WORKER=1` in the environment variables during application installation. In this mode, the server will manage tasks, while external workers (installed elsewhere) will process them.
+2. **Advanced Installation**: Starting from Nextcloud version `31`, you can install only the server component by setting `DISABLE_WORKER=1` in the environment variables during application installation. In this mode, the server will manage tasks, while external workers (installed elsewhere) will process them.
 
 > [!NOTE]
 > Worker authentication uses Nextcloud credentials. If 2FA is enabled, you will need to use an App Password to connect external workers.

appinfo/info.xml

@@ -4,27 +4,38 @@
 	<name>Visionatrix</name>
 	<summary>Visionatrix - scalable AI Media processing</summary>
 	<description>
-	<![CDATA[Nextcloud Visionatrix is a standalone [Visionatrix](https://github.com/Visionatrix/Visionatrix) service that allows you to process your media files right within your Nextcloud.
+	<![CDATA[The best open-source solution for media processing, [ComfyUI](https://github.com/comfyanonymous/ComfyUI), extended with [Visionatrix](https://github.com/Visionatrix/Visionatrix), now available in your personal Nextcloud.
 
-*Each user gets their own Visionatrix account, authenticated using Nextcloud credentials, with a personalized task history.*
+✨ Process your media files directly in Nextcloud using powerful, customizable AI workflows.
 
 ---
 
-Experience the full potential of [ComfyUI](https://github.com/comfyanonymous/ComfyUI) integrated directly into Nextcloud Visionatrix, empowering you to explore a wide array of pre-built AI workflows or develop and integrate your own with ease.
+- 🌟 **Ready-to-use flows**: A library of workflows for various AI tasks.
+- 🛠️ **Custom workflows**: Design and adapt workflows to fit your needs.
+- 🤝 **Nextcloud integration**: Use Visionatrix with Nextcloud Assistant as an AI provider for seamless media processing.
+- 🚀 **Parallel task processing**: Scale up your workflows with support for parallel execution.
+- 🌍 **Multilingual prompts**: Create AI content in any language supported by your chosen language model (e.g., QWEN or Gemini).
+- 🤖 **LLM integration**: Use Ollama (LLM + Vision) and Google Gemini for enhanced functionality.
+- 🖥️ **Hardware support**: Works on both CPU and GPU, with flexibility for various configurations.
 
-- 🌐 **Ready-to-Use Flows**: Access an extensive collection of high-quality ComfyUI flows tailored for various AI tasks.
-- 🛠️ **Custom Flow Development**: Build, modify, and enhance your own workflows, adapting them to your unique requirements.
-- 🚀 **Seamless Integration**: Enjoy a unified experience where ComfyUI's power meets Nextcloud’s efficiency, offering an unparalleled AI workflow ecosystem.
+---
+
+⚠️ **This version is in beta** as we continue developing key features to deliver a complete and polished product:
+
+- 💬 Chat interface with an AI media assistant.
+- ✏️ Support for editing workflows directly within the Nextcloud interface using the ComfyUI interface.
+- 🔄 On-the-fly model support for [CivitAI](https://civitai.com).
+- ⚙️ Enhanced configuration for ultra low-VRAM GPUs (4 GB) and other performance improvements.
 
 ---
 
-Nextcloud Visionatrix can also be used from within the Nextcloud Assistant as a "txt2img" image provider, making it your go-to solution for all AI-powered media processing needs.
+💡 **We value your feedback and contributions!**
 
-Embrace the next level of AI flow management with the synergy of ComfyUI and Nextcloud today!
+Share your custom workflows, big or small. For guidance on creating AI workflows, join our [discussions](https://github.com/Visionatrix/Visionatrix/discussions). Contributions to the [documentation](https://visionatrix.github.io/VixFlowsDocs/) are always welcome.
 
-For more information, see the [repository](https://github.com/cloud-py-api/visionatrix) readme file.
+📚 For more details, visit the [repository](https://github.com/cloud-py-api/visionatrix) readme.
 	]]></description>
-	<version>1.0.1</version>
+	<version>1.1.0</version>
 	<licence>MIT</licence>
 	<author mail="[email protected]" homepage="https://github.com/andrey18106">Andrey Borysenko</author>
 	<author mail="[email protected]" homepage="https://github.com/bigcat88">Alexander Piskun</author>
@@ -44,7 +55,7 @@ For more information, see the [repository](https://github.com/cloud-py-api/visio
 		<docker-install>
 			<registry>ghcr.io</registry>
 			<image>cloud-py-api/visionatrix</image>
-			<image-tag>1.4.1</image-tag>
+			<image-tag>1.9.0</image-tag>
 		</docker-install>
 		<routes>
 			<route>