google-maps-scraper 0.3.0

Apify-style Google Maps scraper for Rust — drive a real Chrome via CDP, search by query + scroll the results feed, click each place, extract name/address/phone/website. No API key needed.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101

# Changelog

All notable changes to this project are documented here. The format is based on
[Keep a Changelog](https://keepachangelog.com/en/1.1.0/).

## [Unreleased]

## [0.3.0] - 2026-07-17

First release published to **crates.io**: install with
`google-maps-scraper = "0.3"`. Earlier versions were git-only.

### Added
- `ScraperConfig::place_panel_jitter` — random extra delay (default `0..=750 ms`)
  added before each place visit so the enrich path no longer makes navigations
  at a fixed, easily-detected interval. **Breaking** for code that constructs
  `ScraperConfig` exhaustively (without `..Default::default()`) — hence the
  0.x minor bump.
- Release automation: pushing a `v*` tag now verifies the tag matches the crate
  version, runs the full check suite, publishes to crates.io, and creates a
  GitHub release (`.github/workflows/release.yml`).

### Changed
- `Cargo.toml`: the temporary `publish = false` guard (added while the crate was
  git-only) is removed for this deliberate crates.io release, and the
  `documentation` field now points at docs.rs.
- `search_many_on_page` now uses plain `Vec` / `HashSet` instead of
  `Arc<Mutex<…>>`; the function is single-task, so the async locks and the
  never-taken `Arc::try_unwrap` panic path were pure overhead.
- `MapsScraper::search` documents that each call opens/closes a tab and
  `search_many` should be preferred for multiple queries.
- `ScraperConfig::max_places` documents that `None` is truly unbounded (one
  navigation per place, no time bound).
- README installation instructions point at crates.io instead of the git
  dependency.

## [0.2.0] - 2026-06-15

### Added
- `Drop` implementation for `MapsScraper` so the CDP handler task is aborted
  even when `close()` is not called (panic / early return).
- `ScraperConfig::max_places` — cap the number of unique places returned per query.
- `ScraperConfig::nav_timeout` — bound every page navigation / feed-wait step.
- `ScraperConfig::proxy` (and `PROXY_URL` env fallback) — launch Chrome behind a proxy.
- `ScraperConfig::browserless_url` (and `BROWSERLESS_URL` env fallback) — connect to a
  remote Chrome over the DevTools WebSocket instead of launching a local browser.
- `Place::latitude` / `Place::longitude` — parsed from the `@lat,lng` segment of `maps_url`.
- `ScraperConfig::user_agent` — optional `User-Agent` override.
- GitHub Actions CI: build, test, and clippy on push / pull request.
- CI now runs `cargo audit` to scan dependencies for known security advisories.

### Changed
- The hardcoded (and stale, macOS-specific) Chrome user-agent is no longer set
  by default. Chrome now reports its own current UA unless `user_agent` is set,
  avoiding a UA/TLS-fingerprint and UA/host-OS mismatch.
- Headless runs now use Chrome's **new** headless mode (`--headless=new`)
  instead of the old `--headless`, so the reported user-agent no longer contains
  the `HeadlessChrome` bot-detection token.
- The proxy value (`proxy` / `PROXY_URL`) is now rejected at launch if it
  contains whitespace (a malformed value Chrome would silently ignore, falling
  back to a direct connection and leaking the real IP).
- Page navigations are wrapped in `tokio::time::timeout` and fail with a clear
  error instead of hanging indefinitely.
- Collected feed URLs are filtered to the `https://` scheme before navigation,
  preventing `javascript:` / `data:` URL execution.
- The German address regex is compiled once via `LazyLock` instead of on every call.
- Upgraded `chromiumoxide` 0.7 → 0.9 and `thiserror` 1 → 2.

### Fixed
- The working tab opened in `search_many` is now closed before returning,
  fixing a tab/memory leak when a scraper is reused for many searches.
- The working tab is now closed even when the run returns early with an error
  (e.g. the initial Maps navigation fails), not only on the success path.
- Each place detail now waits (bounded) for the panel `<h1>` to render before
  extraction, in addition to the fixed settle delay — reduces empty `Place`s on
  slow renders.
- The enrich path now skips a place URL it has already visited *before*
  navigating, and registers every visited URL (including ones whose extraction
  failed), avoiding wasted navigations for exact-duplicate URLs within and
  across queries. (Distinct URLs sharing one website domain are still
  deduplicated, but only after navigation, since the domain is read from the
  loaded panel.)

## [0.1.0] - 2026-05-02

### Added
- Initial release of `google-maps-scraper`.
- `MapsScraper::launch` — start a headless Chrome and return a scraper instance.
- `MapsScraper::search` — scrape Google Maps for a single query.
- `MapsScraper::search_many` — run multiple queries in one browser session with
  deduplication by website domain (or maps URL when no website).
- `ScraperConfig` — configure headless mode, scroll iterations, enrichment, and delays.
- `Place` — structured output with name, address, postcode, city, phone, website, maps_url.
- German postcode/city parsing via `parse_german_address`.
- Cookie consent auto-dismiss for German and English Google interfaces.

<!-- Version links resolve once the matching git tags are pushed. -->
[Unreleased]: https://github.com/Liohtml/google-maps-scraper-rs/compare/v0.3.0...HEAD
[0.3.0]: https://github.com/Liohtml/google-maps-scraper-rs/compare/v0.2.0...v0.3.0
[0.2.0]: https://github.com/Liohtml/google-maps-scraper-rs/compare/v0.1.0...v0.2.0
[0.1.0]: https://github.com/Liohtml/google-maps-scraper-rs/releases/tag/v0.1.0