# Changelog of Zameen.com Scraper | Pakistan Real Estate & Property Data (`haketa/zameen-scraper`) Actor

- **URL**: https://apify.com/haketa/zameen-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/haketa/zameen-scraper.md

## Changelog

All notable changes to the **Zameen.com Scraper** are documented here.

### 1.1.0 — 2026-11-01

#### Fixed — bedrooms, bathrooms & location now populate reliably

- **`bedrooms` 11% → 100%** and **`bathrooms` 4% → 100%** (Homes). The parser now reads Zameen's dedicated `aria-label="Beds"` / `aria-label="Baths"` card nodes (the value is just the number, e.g. `5`), falling back to the previous `"N Bed"` / `"N Bath"` text regex only when those labels are absent.
- **`location` 0% → 100%.** Added `aria-label="Location"` as the primary neighborhood/sector selector (e.g. `DHA Phase 7 - Block T, DHA Phase 7`); the old class-based selectors remain as a fallback.
- **`society` detection improved** as a side effect — because `society` is derived from `location` first, listings whose title omits the scheme name (but whose location is `Bahria Town - Jinnah Block`, etc.) are now classified correctly.
- **`propertyType` 58% → 84%.** Expanded the type-keyword map so common synonyms normalize to a canonical type (`Bungalow`/`Villa` → `House`, `Apartment` → `Flat`, `Plaza` → `Building`, `Farmhouse` → `Farm House`) and the title is checked before the full card text.

#### Verified

- Fresh verification run of **400 listings** (Lahore for-sale Homes, residential proxy, list mode) in ~2.5 minutes: **0 duplicates** on `listingId`, **0 value bugs**, and **100% coverage** of all identity/core fields — `listingId`, `title`, `listingUrl`, `price`, `priceDisplay`, `city`, `area`, `areaUnit`, `areaInSqFt`, `images`. Pagination walked 16 pages (~25 listings/page) cleanly and stopped exactly at the `maxRecords` cap.
- Prices normalized correctly across the run (2.5M–480M PKR from `Crore`/`Lakh` strings); areas converted correctly from `Marla`/`Kanal` to `areaInSqFt`.

#### Docs

- README refreshed to the current output: corrected the field-coverage notes, updated the Output examples to match real records, and clarified which fields are list-mode vs `scrapeDetails` only.

### 1.0.4 — 2026-10

- `listingId` extraction hardened to match Zameen's current `/Property/...-<id>-<variant>-<page>.html` URL shape, so the numeric listing ID now populates reliably and is used as the primary dedupe key.

### 1.0.0 — 2026-05

- Initial public release — Cheerio + got-scraping, Crore/Lakh and Marla/Kanal normalization, society detection, JSON-LD fallback, optional detail-page mode, residential-proxy support.
