feat: Integrate Omost for Enhanced Image Prompt Generation

[bannert1337]

Is your feature request related to a problem? Please describe.
I'm always frustrated when generating image prompts because the current system lacks the advanced capabilities to leverage the coding potential of Large Language Models (LLMs) for more precise and creative image compositions.

Describe the solution you'd like
I would like to see the integration of Omost into open-webui. Omost utilizes the coding capabilities of LLMs to generate detailed and structured image prompts. This integration would allow users to create more complex and accurate image compositions by leveraging Omost's virtual Canvas agent, which can be rendered by specific implementations of image generators.

Describe alternatives you've considered

• Continuing with the current prompt generation system, which lacks the advanced features provided by Omost.
• Using other image generation tools that do not offer the same level of integration with LLMs for coding-based image composition.

Additional context
Omost is a project designed to convert LLM's coding capabilities into image generation capabilities. It provides pretrained LLM models that write code to compose image visual contents. The Canvas agent in Omost can be rendered by specific implementations of image generators to actually generate images. Integrating Omost would enhance the functionality of open-webui by providing more detailed and structured image prompts.

For more information, you can visit the official HuggingFace space or the Omost GitHub repository.