docs/STREAMLIT_APP.md

📄 Digi-Biz Streamlit App

Quick Start

1. Install Dependencies

pip install streamlit

2. Run the App

streamlit run app.py

The app will open in your browser at http://localhost:8501

---

Features

📤 Upload Tab

• Upload ZIP files containing business documents
• Supports PDF, DOCX, XLSX, images, videos
• Shows file size and job ID

⚙️ Processing Tab

• Real-time progress through 5 agents:
  1. File Discovery Agent
  2. Document Parsing Agent
  3. Table Extraction Agent
  4. Media Extraction Agent
  5. Vision Agent (Qwen3.5:0.8B)
• Live status updates
• Error handling with graceful degradation

📊 Results Tab

• File discovery summary (documents, images, videos)
• Document parsing results (pages, text preview)
• Table extraction results (count, types)
• Expandable details for each section

🖼️ Vision Analysis Tab

• Image gallery with analysis results
• Category classification (product, service, food, etc.)
• Confidence scores
• Tags and descriptions
• Product/service detection
• Association suggestions

---

Sidebar Features

• Model Status: Shows Ollama server and Qwen model availability
• Agent Cards: Quick reference for all 5 agents
• Reset Button: Clear all session data and start fresh

---

Requirements

System Requirements

• Python 3.10+
• Ollama installed and running
• Qwen3.5:0.8b model pulled

Python Packages

pip install -r requirements.txt

---

Usage Example

1. Prepare ZIP file with business documents:

   • Restaurant menu PDFs
   • Product catalogs
   • Service brochures
   • Business cards
   • Product photos

2. Upload the ZIP file in the "Upload" tab

3. Click "Start Processing" - watch real-time progress

4. View Results in "Results" and "Vision Analysis" tabs

---

Troubleshooting

Ollama Not Running

Error: Ollama Server Not Running

Solution: Start Ollama server

ollama serve

Qwen Model Not Found

Error: Qwen3.5:0.8B Not Available

Solution: Pull the model

ollama pull qwen3.5:0.8b

Processing Timeout

If processing takes too long:

• Reduce number of images in ZIP
• Vision analysis processes first 3 images by default
• Increase timeout in vision_agent.py

---

Screenshots

The app provides:

• ✅ Clean, modern UI with custom styling
• ✅ Progress bars and status indicators
• ✅ Interactive expandable sections
• ✅ Image gallery with analysis overlays
• ✅ Real-time agent status updates

---

Development

Run in Development Mode

streamlit run app.py --server.headless=true --server.port=8501

Enable Debug Logging

Add to app.py:

import logging
logging.basicConfig(level=logging.DEBUG)

---

Architecture

┌─────────────────────────────────────────┐
│         Streamlit Frontend              │
│  - Upload component                     │
│  - Progress tracking                    │
│  - Results display                      │
└─────────────────────────────────────────┘
                    ↓
┌─────────────────────────────────────────┐
│         Agent Pipeline                  │
│  1. File Discovery Agent                │
│  2. Document Parsing Agent              │
│  3. Table Extraction Agent              │
│  4. Media Extraction Agent              │
│  5. Vision Agent (Qwen3.5:0.8B)         │
└─────────────────────────────────────────┘
                    ↓
┌─────────────────────────────────────────┐
│         Ollama (Qwen Vision)            │
│  - Image analysis                       │
│  - Category classification              │
│  - Tag generation                       │
└─────────────────────────────────────────┘

---

Next Steps

After the demo:

1. Review extracted data
2. Export results (JSON export coming soon)
3. Edit/refine results (editing UI in development)
4. Integrate with downstream systems

---

Enjoy using Digi-Biz! 🚀