Added ability for vision capable modal to talk to

[HaarisIqubal]

🚀 Feature: Vision Model Integration & Image Handling Enhancements

✨ Summary

This pull request introduces major enhancements to support interaction with vision-capable models. The following key features have been added:

✅ What's New

• Vision Model Support:
  Integrated support to chat with models capable of processing images (e.g., llava, "gemma:4b").

• Image Uploading:
  Users can now upload one or multiple images to send as part of their message payload.

• Image Deletion:
  Added the ability to remove selected images before sending a message to the model.

• Persistent Storage:
  Images are now stored persistently so that they're available across app sessions.

---

📦 Implementation Details

• Base64 encoding is used to serialize image data before sending it to the model.
• Uploaded images are resized and converted to JPEG format to reduce payload size.
• State management for image selection and deletion is handled using @State/@StateObject (if applicable).
• Persistent storage is implemented using FileManager or similar API.

---

📝 Notes

• Currently supports up to X images (if there's a limit we can put if needed need to test how much RAM needed when more photos are uploaded).
• Future enhancement: support drag & drop or camera capture (if desired).

---

I hope you consider integrating this into the main codebase. Let me know if any changes are needed or if you'd like additional enhancements! Below are some screenshots that are captured from application hope you get some idea how it works.