Acs Bundle Laravel Package

Product Decisions This Supports

• Market entry / feature parity for Greece & Cyprus: The package's entire scope is fetching ACS pickup points (ParcelShopsService::getList(CountryIdEnum, ?int $kind) → ParcelShop[]) for GR and CY. That directly supports a "collect from an ACS point" delivery option at checkout — a localized expectation in those markets — or a store/pickup-point locator.
• Build vs. buy (and buy vs. wrap): Parcel-shop lookup is undifferentiated plumbing: one API call, a small DTO set, two error types. This package is a reference implementation and contract definition, not a drop-in — it is a Symfony bundle, and the target stack is Laravel/PHP. The real PM decision is not "adopt or build" but "wrap/port vs. reimplement the thin client," priced at a short spike rather than a multi-month build.
• Roadmap sequencing: Because coverage stops at discovery, it forces an explicit phased plan: Phase 1 = show pickup points (this package); Phase 2 = label creation, tracking, returns, webhooks (not covered — separate ACS integration work and separate buy/build decision).
• Checkout & UX decisions: The kind filter enables product choices — all points vs. lockers only — and the returned shop data drives selector UX, address/hours display, and default-point logic. The GR/CY-only constraint also bounds which markets the feature ships in.
• Reliability and fallback design: The typed ServiceUnavailable and MalformedResponse exceptions (plus the Guzzle timeout added in 2.1.1) define the failure surface. PM-visible decisions follow: retry policy, caching of shop lists, timeout budget, and whether checkout degrades gracefully to home delivery when ACS is down.
• Dependency risk budgeting: 0 stars, 0 dependents, single-vendor maintenance, MIT license. Correct framing is internal tooling we own and can fork — not a community-backed dependency to plan around.
• Documentation-quality signal: The README's config block is labeled answear_gls inside an ACS bundle, and installation points at a raw Git URL. Small, but it tells you onboarding requires engineering verification, and it should shape the estimate.

When to Consider This Package

Consider it when:

• You sell or ship into Greece and/or Cyprus, and pickup-point delivery is table stakes or a conversion lever.
• The needed scope is read-only parcel-shop discovery — no labels, no tracking, no returns yet.
• Your stack is PHP 8.2+ (8.4 for 4.0.0) on a modern Laravel; you're willing to spend a short spike wrapping or porting rather than installing the bundle directly.
• You want a free, MIT-licensed reference for ACS auth (apiKey, companyId, companyPassword, userId, userPassword, default language GR) and response shapes instead of reverse-engineering the carrier API.
• You can own a small internal surface: a wrapper service, config file, tests, and a maintainer.

Look elsewhere or deprioritize when:

• You need end-to-end shipping — label generation, tracking, returns, status webhooks. This package does none of it.
• You're Laravel-native with no Symfony components and low appetite for a wrapper: a ~200-line client using Laravel's Http facade plus simple DTOs may be cheaper and cleaner than adapting Symfony-bound code.
• You need multi-country or multi-carrier coverage, or a carrier-agnostic abstraction (EasyPost/Shippo-style) rather than a single-carrier bundle.
• You require published SLAs, support commitments, or a broad adopter base to clear procurement/security review.
• You have no PHP owner to fork, test, and patch — with zero dependents, the maintenance burden is entirely yours.

How to Pitch It (Stakeholders)

Executives: "ACS is a major Greek courier, and pickup-point delivery is expected in Greece and Cyprus. This MIT-licensed bundle gives us the ACS parcel-shop lookup for free — one call returning shops filtered by country and kind. It is not a Laravel package, so engineering wraps it; budget a one-to-two week spike, not a build cycle. It covers discovery only — labels and tracking remain separate roadmap decisions. Cash cost is near zero; the real risk is a single-maintainer, zero-adoption dependency, which we contain by isolating it behind our own service interface so it can be swapped without touching product code."

Engineering: "Symfony 6/7 bundle, PHP 8.2–8.4, Guzzle-based, config-driven, with a deliberately small public surface: ParcelShopsService::getList(CountryIdEnum, ?int $kind) returning ParcelShop DTOs, plus typed ServiceUnavailable and MalformedResponse exceptions. Recommendation for our Laravel stack: don't pull the bundle in for its container — reuse its ACS request/response contract and reimplement the client with Laravel's HTTP client. Move config to config/acs.php and env/vault, add timeout, retry, and caching for shop lists, map the two exception types, and fake responses in tests. Wrap it all in app/Services/AcsService so a carrier change is a one-file change. Note the README's config key is mislabeled (answear_gls), so validate against a live sandbox call before committing to the contract."

Proposed next step: approve a time-boxed spike for the GR/CY pickup-point option, gated on a documented fallback to home delivery when ACS is unavailable and a checkout-conversion read before full rollout.