HomeSnap 🏠 & EventSnap 📸

A secure, stateless Streamlit web application that extracts events and apartment viewing details from text or URLs using OpenAI's GPT-4o, adding them directly to your Google Calendar.

---

🌟 Key Features

• Multi-Mode Routing: Access the app in Event Mode (EventSnap) or Apartment Hunting Mode (HomeSnap) via query parameters (?mode=apartment).
• Smart LLM Extraction: Automatically parses titles, prices/rents, locations, and descriptions. Uses your local timezone context to correctly resolve relative descriptions like "Tomorrow at 18:30" or "Tonight".
• Unified Context Merging: Send a listing URL and a separate viewing date/time text block in the same paste; the app automatically scrapes the webpage and correlates it with your visit schedule into a single calendar entry.
• Persistent Login: A secure 24-hour cookie-based auto-login preserves your session across tab closures, silently refreshing expired tokens in the background.
• Stateless & Private: The app requests the least-privilege scope (calendar.app.created), meaning it cannot read or edit your primary personal calendar. No user credentials or event details are stored on any backend database.
• Production-Ready Security: Hardened against Cross-Site Request Forgery (CSRF) via cryptographic state signatures and protected from Server-Side Request Forgery (SSRF) during URL scraping.

---

🛠️ Local Setup & Configuration

Follow these steps to run the application locally:

1. Prerequisites

Ensure you have Python 3.9+ installed.

2. Clone and Install Dependencies

# Clone the repository
git clone https://github.com/your-username/EventSnap.git
cd EventSnap

# Create a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows use: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

3. Configure Google OAuth Credentials

1. Go to the Google Cloud Console API & Services.
2. Create a new project.
3. Configure the OAuth Consent Screen (User Type: External, Publishing Status: Testing).
   • Request scope: .../auth/calendar.app.created (minimum scope to write app-owned events).
4. Go to Credentials, click Create Credentials -> OAuth Client ID.
   • Application type: Web application.
   • Authorized redirect URIs: Add http://localhost:8501 (and your deployed Streamlit URL when hosting in production).

4. Create Streamlit Secrets

Create a file named .streamlit/secrets.toml in the project root:

OPENAI_API_KEY = "your-openai-api-key"

[google_oauth]
client_id = "your-google-oauth-client-id.apps.googleusercontent.com"
client_secret = "your-google-oauth-client-secret"
redirect_uri = "http://localhost:8501"

Warning

Never commit .streamlit/secrets.toml to GitHub! It is ignored by default in the project's .gitignore file.

5. Run the Application

streamlit run app.py

Open http://localhost:8501 in your browser. To access the Apartment Hunting mode, navigate to http://localhost:8501/?mode=apartment.

---

🔒 Security Measures

• Stateless CSRF Prevention: Integrates HMAC-SHA256 tokens using the client secret to sign redirect parameters, securing authentication statelessly without relying on fickle memory sessions.
• Scraper SSRF Filtering: Resolves target domain IPs and blocks requests attempting to access local networks (localhost) or cloud metadata endpoints (169.254.169.254).
• API Abuse Controls: Limits text inputs to 5,000 characters, scrapes a maximum of 3 URLs per request, and enforces a 5-second cooldown alongside session caps to prevent automated token depletion.

---

📄 License

This project is licensed under the MIT License.