Bazos Alert Bot

Automated web scraper for Bazos.sk/cz that monitors listings and sends Discord notifications for new matches.

Features

• Scrapes Bazos.sk and Bazos.cz for listings
• Filters by search terms, price range, and location
• Sends Discord notifications for new listings
• Saves listing data in a SQLite database (further analysis possible)
• Prevents duplicate notifications using database tracking
• Supports scheduled runs via GitHub Actions or Docker

Quick Start

1. Configuration

Create a config.json file in the project root based off the config.example.json provided. Example:

{
  "database_path": "./data/listings.db",
  "searches": [
    {
      "name": "Fiat Panda",
      "source": "bazos_sk",
      "search_term": "Panda",
      "price_min": 2000,
      "price_max": 15000,
      "max_pages": 3,
    }
  ]
}

or like this (preferred, because it only goes through the category in the URL):

{
  "database_path": "./data/listings.db",
  "searches": [
    {
      "name": "Macbook Air",
      "source": "bazos_sk",
      "url": "https://pc.bazos.sk/notebook/?hledat=MacBook+Air&rubriky=pc&hlokalita=81101&humkreis=45&cenaod=&cenado=&Submit=H%C4%BEada%C5%A5&order=&crp=&kitx=ano",
      "max_pages": 3
    }
  ]
}

Configuration Options

• database_path: Path to SQLite database file (default: ./data/listings.db)
• searches: Array of search configurations

Important: The Discord webhook URL is not stored in config.json. It must be provided via the DISCORD_WEBHOOK_URL environment variable (see .env.example).

Search Configuration

• name: Descriptive name for this search
• source: bazos_sk or bazos_cz
• url: (optional - no need to use other search parameters with this) Full URL to a specific search
• search_term: What to search for (e.g., "Panda", "A4")
• price_min: Minimum price filter (optional)
• price_max: Maximum price filter (optional)
• max_pages: Maximum number of pages to scrape (default: 3)
• location: Location to filter results (optional)
• radius: Radius in kilometers to filter results (optional)

2. Discord Webhook Setup

1. Go to your Discord server settings
2. Create a channel
3. Navigate to channel Edit Channel > Integrations > Webhooks
4. Click "New Webhook"
5. Copy the webhook URL

Deployment Options

Option 1: GitHub Actions (Recommended)

Setup:

1. Fork this repository

2. Add your Discord webhook as a repository secret:

   • Go to Settings > Secrets and variables > Actions
   • Click "New repository secret"
   • Name: DISCORD_WEBHOOK_URL
   • Value: Your Discord webhook URL

3. Edit config.json with your search preferences (commit and push)

4. The workflow will run automatically at 8:00, 14:00, and 20:00 CET

5. You can also trigger it manually:

   • Go to Actions > Scheduled Bazos Scraper
   • Click "Run workflow"

Notes:

• Database is stored as a GitHub Actions artifact (only 90-day retention!)
• First run will notify all matching listings
• Subsequent runs only notify new listings

Option 2: Docker

Setup:

1. Clone the repository:

git clone https://github.com/peter115342/auto_alert_bot.git
cd auto_alert_bot

2. Create a .env file:

echo "DISCORD_WEBHOOK_URL=your_webhook_url_here" > .env

DISCORD_WEBHOOK_URL=your_webhook_url_here

3. Edit config.json with your search preferences

4. Build and run:

docker compose -f docker-compose.scheduled.yaml up -d

Docker Configuration:

The docker-compose.scheduled.yaml file runs the scraper at 8:00, 14:00, and 20:00 daily using cron.

To modify the schedule, edit line 20:

echo "0 8,14,20 * * * cd /app && /root/.local/bin/uv run python -m src.main >> /app/data/cron.log 2>&1" > /etc/cron.d/bazos-scraper

Cron format: minute hour day month weekday

• 0 8,14,20 * * * = 8:00, 14:00, 20:00 daily
• 0 */6 * * * = Every 6 hours
• */30 * * * * = Every 30 minutes

Docker Commands:

# Start the bot
docker compose -f docker-compose.scheduled.yaml up -d

# View logs
docker compose -f docker-compose.scheduled.yaml logs -f

# Stop the bot
docker compose -f docker-compose.scheduled.yaml down

# Rebuild after code changes
docker compose -f docker-compose.scheduled.yaml build --no-cache
docker compose -f docker-compose.scheduled.yaml up -d

Database Location:

The database is stored in ./data/listings.db on your host machine and persists between container restarts.

Database Schema

The SQLite database tracks listings with the following fields:

• id: Listing ID
• source: bazos_sk or bazos_cz
• title: Listing title
• url: Full URL to listing
• price: Price string
• location: Location with postal code
• description: First 200 characters
• date_posted: Date posted
• view_count: Number of views
• first_seen: When first discovered
• last_checked: Last time seen in scrape
• notified: Whether Discord notification was sent