Cian Parser — Showcase

Русская версия

This is a documentation-only showcase. The source code of Cian Parser is proprietary and is not included in this repository.

Cian Parser is a commercial automation tool for the Russian real-estate platform Cian.ru. It monitors listings on a schedule, applies user filters, and pushes new matches to Telegram. This repo demonstrates the product surface — feature scope, example configuration, screenshots — for portfolio and client-evaluation purposes.

Why this repo exists

Prospective clients ask "what does it actually do, and how is it configured?". This showcase answers that without exposing the working scraper, anti-detection logic, or proxy infrastructure — those remain proprietary.

How it works (high level)

1. Scheduler triggers periodic scrapes (APScheduler).
2. Fetcher pulls Cian search-result pages through a rotating proxy pool with user-agent randomization and request pacing.
3. Parser extracts listings into a normalized schema (price, area, rooms, location, posted-at, contact).
4. Filter engine applies user-defined rules (see examples/filters.example.json).
5. Deduper compares against a persistent listing store to skip already-seen ads.
6. Notifier pushes matches to Telegram with photos and a deep link.
7. Analytics keeps price history per listing for trend reporting.

Tech stack

Layer	Tools
Language	Python 3.9+
Scheduling	APScheduler
HTTP	requests / httpx with proxy rotation
Parsing	lxml / BeautifulSoup
Storage	SQLite / PostgreSQL
Notifications	Telegram Bot API
Deployment	systemd / Docker (client choice)

Features

• Scheduled scraping with configurable intervals (every N minutes, specific hours, or cron expressions)
• Rotating proxy pool with automatic health checks and failover
• User-agent randomization and request pacing to avoid detection
• Telegram notifications with listing photos, price, area, floor, and a direct link
• Price history tracking per listing for trend analysis
• Duplicate detection across runs — only new or updated listings trigger alerts
• Filter engine supporting nested boolean logic (AND/OR/NOT)
• Multi-user support — each user maintains independent filters and notification channels

Filter options

Filters are defined in JSON and support the following fields:

Field	Type	Example
price_min / price_max	integer	30000 / 80000
rooms	list	[1, 2]
area_min / area_max	float	35.0 / 90.0
floor_min / floor_max	integer	2 / 25
not_first_floor	boolean	true
not_last_floor	boolean	true
districts	list	["Presnensky", "Tverskoy"]
metro_stations	list	["Barrikadnaya"]
max_km_from_metro	float	1.5
keywords_exclude	list	["auction", "commercial"]

See examples/filters.example.json for a complete sample.

Alert configuration

Telegram alerts are configured per user:

{
  "telegram_chat_id": "123456789",
  "notify_on": "new_listing",
  "quiet_hours": { "from": "23:00", "to": "07:00" },
  "batch_interval_minutes": 15,
  "include_photos": true,
  "max_photos_per_listing": 3
}

Alerts can be batched to avoid notification spam during high-volume periods.

Architecture overview

APScheduler (cron)
      |
  Fetcher (httpx + proxy pool)
      |
  HTML Parser (lxml / BS4)
      |
  Filter Engine --> Deduper (SQLite/PG)
      |
  Notifier (Telegram Bot API)
      |
  Analytics (price history store)

Examples

• examples/filters.example.json — filter configuration shape

Limitations (showcase, not the product)

• This repository contains no runnable code. Cloning it will not give you a working scraper.
• Screenshots and demo media live in media/ (when present) and may lag behind the live product.
• Filter examples are illustrative — the production schema is broader and changes with client requirements.
• Cian's HTML/anti-bot surface changes frequently; the production tool is maintained against those changes, but no public guarantees are made here.

Inquiries

For commercial inquiries, similar projects, or licensing of the underlying tool: creatmanick@gmail.com · creatman.site.

EOF