Image generation auto-choose style

[Guz-Ali]

Each news image will be generated by Deep AI based on an image style. At first we will manually select one style, but as we progress through the project, choosing the style should be automated.

• [ ] Look at Deep AI documentation to find available image styles.
• [ ] Given the title and summary, ask AI Model (Gemini) which one would be more fitting.
• [ ] Set the image style of news card based on the answer. Requirements:
• [ ] This function should be robust and reliable, and it should not fail.
• [ ] The return answer of the AI Model should only be one image style that we can directly pass in to Deep AI API
• [ ] The return answer should not have any punctuation, or any other word/character other than the style name.

[Guz-Ali]

I'm currently taking a Google course on how to use Gemini in this way, but if you are interested in implementing this feel free to take it. This will be a work mainly on the Gemini API than the image generation API, so we can still implement it while deciding which API to use for image generation.