E-auksion lot radari

Problem

The auction platform lists thousands of lots, and watching for the relevant ones by hand takes too long.

Solution

Twice a week the program collects lots, screens them by a criteria file, compares prices with other lots, scores them and sends the TOP-10 with analysis to Telegram. A lot that was already sent is not repeated unless its price drops.

Result

Runs on a server on a schedule. The code relies only on the Python standard library and is covered by 200+ automated tests.

Highlights

• Criteria in JSON: property group, category, starting price, payment terms
• Price comparison and a scoring system
• Duplicate-lot filtering (SQLite)
• Dry-run mode and logging; no external dependencies