Laraue.Apps.RealEstate

Application allows collecting real estate advertisements and ranking them with AI. Advertisements then can be requested via API.

System requirements

1. PostgreSQL database with version 15+. Settings are use default 'postgres:postgres' login/pass as default.
2. Local or remotely available ollama instance. The system is configured with default Ollama local address, so after ollama installing application should work without any setup. But the first run will load the model and make takes a time. The default model "qwen2.5vl:7b" uses near 8GB memory, and preferably it should be GPU memory. The model could be changed in settings.

Crawlers setup

As default, crawlers for Moscow, Saint Petersburg and Volgograd are enabled. It can be changed to the preferable cities.

Host information

Laraue.Apps.RealEstate.ApiHost: Provide API to the ready advertisements.
Laraue.Apps.RealEstate.CrawlingHost: Collect advertisements and put them into DB.
Laraue.Apps.RealEstate.GpuWorkerHost: Sequentially get predictions for each advertisement and update prediction in DB.
Laraue.Apps.RealEstate.TelegramHost: Host to set up telegram notifications about new advertisements.
Laraue.Apps.RealEstate.WorkerHost: The host marking advertisements as ready for API and also remove these flugs.