QUICKSTART.md

Quick Start Guide

Setup (One-time)

1. Install Python dependencies:

   pip install -r requirements.txt

2. Configure API key:

   # Copy the template
   cp .env.example .env
   
   # Edit .env and add your Google API key
   # GOOGLE_API_KEY=your_actual_api_key_here

   Get a free API key at: https://makersuite.google.com/app/apikey

Run Extraction

python src/extract_parameters.py

This will:

1. Read snippets from input/ directory
2. Process each snippet with Gemini 1.5 Pro
3. Extract architectural parameters
4. Save results to output/parameters.yaml

View Results

# Windows
type output\parameters.yaml

# Linux/Mac
cat output/parameters.yaml

Project Structure

.
├── README.md                    # Project overview
├── SUMMARY.md                   # Detailed summary report
├── QUICKSTART.md                # This file
├── requirements.txt             # Python dependencies
├── .env.example                 # API key template
│
├── input/                       # Input specification snippets
│   ├── snippet1_caches.txt
│   └── snippet2_csr.txt
│
├── src/                         # Source code
│   └── extract_parameters.py   # Main extraction script
│
├── output/                      # Generated results
│   ├── sample_parameters.yaml  # Example output
│   └── parameters.yaml         # Actual results (generated)
│
├── docs/                        # Documentation
│   ├── llm_details.md          # LLM specifications
│   └── methodology.md          # Technical approach
│
└── prompts/                     # Prompt engineering
    └── prompt_evolution.md     # Prompt development history

Troubleshooting

"No module named 'google.generativeai'"

pip install google-generativeai

"API key not found"

Make sure you:

1. Created .env file (copy from .env.example)
2. Added your actual API key to .env
3. Running from project root directory

"JSON parsing error"

The LLM occasionally returns malformed JSON. The script will print the error and continue. This is rare with temperature=0.1.

Next Steps

• Add more snippets: Place new .txt files in input/ directory
• Modify prompts: Edit the prompt in src/extract_parameters.py
• Validate results: Review output/parameters.yaml for accuracy
• Read documentation: See docs/ for detailed methodology

Support

For issues or questions:

1. Check SUMMARY.md for detailed documentation
2. Review prompts/prompt_evolution.md for prompt engineering insights
3. See docs/methodology.md for technical details