1. 📁 Upload Data

To upload multimedia data:

• Click "Upload" in the top-right menu.
• Enter the following details:
  • Collection Name: A logical grouping for related data (like a folder).
  • Description: Briefly describe the content (e.g., "A lovely dog").
  • Location: Provide either an address (e.g., "London Zoo") or coordinates.
• Drag and drop files or click to browse files.

Example:

• Uploaded: Image of "A lovely dog with sunglasses"
• Location: "University College London"
• Collection: "Demo"

(Uploaded files from different sources can be stored in the same collection to enrich your dataset.)

---

2. 🌍 Map View

To visualize your uploaded data geographically:

• Click "Map" in the top-right menu.
• Enter:
  • Collection Name
  • Central Location (address or coordinates)
  • Keywords (optional)
  • Search Radius

Example:

• Center: "King’s Cross", Radius: 3.6 km, Collection: "Demo"
• Result: Dog image at "University College London", 1.14 km away

---

3. ✅ Task Management (Kanban Interface)

Efficiently organize and prioritize your tasks:

• Click "Dashboard" from the top-right menu.
• Click "Add New Task".
• Enter task details:
  • Title
  • Description
  • Priority (High, Medium, Low)
  • Category
  • Due Date
  • Assignee
• Use Drag-and-Drop to move tasks across Todo, In Progress, and Done.

Example:

• Task: "Lost dog in Regent's Park"
• Description: "A white dog with a smile is lost"

---

4. 💬 Multimodal RAG

Engage in advanced retrieval-augmented conversations:

• Click "Chat" in the top-right menu.
• Check "Retrieve" to enable RAG-based search.
• Select the desired Collection from the dropdown.

Example (Text Retrieval):

"I am looking for a dog with sunglasses."

→ Bot retrieves image and notes: “pink and heart-shaped sunglasses.”

Example (Multimodal Retrieval):

• Click upload next to input box.
• Select an image and type a query.

"I am looking for this kind of dog."

→ Bot retrieves visually similar dog image + context.

---

5. 🎬 Chat with Video

Interact with video content using AI-powered Q&A:

• Click "Video" in the top-right menu.
• Click "Upload Video".
• Paste YouTube URL.
• Optional: Set Transcript Augmentation (n).
• Check No Language Sound for silent videos.
• Click Upload. Processing may take time.

n controls how many transcript segments are combined. Test to optimize results.

Example:

We upload the YouTube video "Welcome back to Planet Earth", documenting NASA astronauts Douglas Hurley and Robert Behnken’s return.

"What is the name of one of the astronauts?"

→ Bot identifies Robert Behnken and provides snippet.

---

🛠 Troubleshooting & FAQs

• Slow Upload/Processing: Use shorter videos with transcripts.
• Location Issues: Prefer precise addresses or coordinates.
• Docker Issues: Run docker compose logs to debug.