Skip to content

GUIDA_AVVIO.md

Latest commit

 

History

History
140 lines (92 loc) · 2.68 KB

GUIDA_AVVIO.md

File metadata and controls

140 lines (92 loc) · 2.68 KB

Quick Start Guide - MultimodalRAG

Setup

Minimum Requirements

Startup Steps

  1. Clone the repository

    git clone <repository-url>
cd multimodalrag

  2. Configure API Key

    cp .env.example .env
# Edit the .env file and replace "your_groq_api_key_here" with your actual key

  3. Launch the stack

    docker-compose up -d

  4. Open the application

Feature Testing

1. Upload a PDF

  • Use the left sidebar to upload a PDF
  • The system will automatically process text, images, and tables

2. Test Queries

Try the following prompts:

General Queries:

  • "What is this document about?"
  • "Summarize the key points"

Image-Based Queries:

  • "Show the images present in the document"
  • "What do the figures show?"

Table Queries:

  • "What data is in the tables?"
  • "Show me numerical results"

Multimodal Queries:

  • "Combine insights from text and images"

Key Features to Observe

Core Functionality

  • PDF Upload – Automatic processing pipeline
  • Multimodal Extraction – Text, images, tables
  • Semantic Retrieval – Context-aware answers
  • Source References – Transparent outputs
  • Modern UI – Clean and intuitive UX

Advanced Technical Features

  • AI Vision – Image captioning with BLIP
  • OCR Support – Text extraction from images
  • Object Detection – YOLO-based visual tagging
  • Vector Database – Semantic similarity via Qdrant
  • LLM Integration – Answer generation with Groq API

Troubleshooting

"Qdrant not reachable" Error

# Check if Qdrant is running
docker ps | grep qdrant

# If not running, restart services
docker-compose down
docker-compose up -d

"Invalid API Key" Error

# Check if the API key is correct in your .env file
cat .env | grep GROQ_API_KEY

# The key must start with "gsk_"

Port Already in Use

# Change port in docker-compose.yml
# Find the line "8501:8501" and update to "8502:8501"
# Then access the app via http://localhost:8502

Support

If you experience issues:

  1. Check logs:

    docker-compose logs -f multimodal-rag

  2. Full reset:

    docker-compose down -v
docker-compose up -d

  3. Alternative local setup:

    pip install -r requirements.txt
docker run -d -p 6333:6333 qdrant/qdrant
streamlit run streamlit_app/Home.py