Add IKEA WebHarbor mirror

[Lxr-max]

Site name: IKEA
Site slug: ikea
Upstream URLs:

• https://www.ikea.com/
  Contribution type: New Website
  Contributor: XuanRui LI / Lxr-max
  Contact email: mingduanmu845@gmail.com

Safety note

• Deterministic demo retail data only.
• No real payment, no real checkout, no real inventory reservation, no real delivery booking, and no live IKEA API calls.

Implementation summary

• Added a self-contained Flask + SQLAlchemy + SQLite IKEA retail mirror under sites/ikea/.
• Implemented product search, room/category browsing, filters, product detail pages, compare, wishlist, cart, room planner bundles, store lookup, support/help search, pickup/delivery mock checkout, account orders, rewards, and order lookup.
• Added deterministic local SVG product/store/category assets and idempotent seed flows.
• Registered the site in websyn_start.sh, control_server.py, and Dockerfile.

Seed row counts

• users: 4
• categories: 12
• products: 156
• stores: 15
• store_inventory: 2340
• reviews: 546
• orders: 60
• order_items: 120
• reward_activities: 20
• wishlist_items: 16
• compare_items: 12
• cart_items: 8
• deals: 18
• support_articles: 24
• support_tickets: 8
• protection_plans: 312
• pickup_slots: 45
• payment_mocks: 60

Benchmark users

• alice.j@test.com
• bob.c@test.com
• carol.d@test.com
• david.k@test.com
• Password: TestPass123!

Task count and categories

• 18 tasks in sites/ikea/tasks.jsonl
• Search
• Category / room browsing
• Compare flow
• Wishlist
• Cart
• Pickup / delivery checkout
• Order lookup
• Account orders / rewards
• Store amenities
• Support article lookup
• Room planner bundle flow

Selected WebSyn port

• 40015

Key routes

• /
• /home
• /categories
• /category/<category_slug>
• /products
• /product/
• /deals
• /stores
• /stores/<store_slug>
• /support
• /search?q=
• /cart
• /checkout
• /checkout/shipping
• /checkout/pickup
• /checkout/payment
• /checkout/review
• /checkout/confirmation
• /login
• /register
• /logout
• /account
• /account/orders
• /account/rewards
• /account/wishlist
• /order-lookup
• /order/<order_number>
• /room-planner
• /compare

Verification commands and results

• py -3 -m py_compile sites/ikea/app.py sites/ikea/seed_data.py sites/ikea/_health.py -> passed
• Tasks JSONL validation -> passed (18 tasks, required schema fields present, unique IDs)
• Additional Flask test-client smoke -> passed for auth, compare, cart, checkout, account, store, and support routes

Docker result

• docker build -t webharbor:ikea-pr . -> passed
• docker run -d --name webharbor-ikea-submit -p 8201:8101 -p 41000-41015:40000-40015 webharbor:ikea-pr -> passed
• GET http://127.0.0.1:8201/health -> all 16 registered sites alive; IKEA served on port 40015

Smoke test result

• Public routes /, /home, /categories, /products, /search?q=desk, /product/IK-10001, /deals, /stores, /stores/brooklyn-ny, /support, /login -> 200
• Real HTTP flow passed for login, wishlist, compare, add-to-cart, pickup checkout, confirmation, order lookup, account orders, rewards, wishlist, and support search
• Example confirmation order generated during smoke: IK-260061

Reset endpoint result

• POST http://127.0.0.1:8201/reset/ikea -> 200 / ready: true

MD5 result

• /opt/WebSyn/ikea/instance/ikea.db and /opt/WebSyn/ikea/instance_seed/ikea.db match after reset
• MD5: 2de50fad32407e799cb8c4659d0a2e19
• After container restart, homepage remained 200 and the MD5 still matched

Screenshot paths

• Local review directory prepared: sites/ikea/scraped_data/submission_review
• Browser capture is currently pending on this machine because npx playwright install chromium failed with ECONNRESET while downloading Chromium

HF asset status

• Local asset tarball: C:\Users\34475\Desktop\VScode\wh-static-pr-ikea\ikea.tar.gz
• Fork upload reference: https://huggingface.co/datasets/lxr-max/WebHarbor/commit/b8fb0114cec0b85fe57874f97a6bf808b46b5747
• Upstream asset reference from hf upload --create-pr: https://huggingface.co/datasets/ChilleD/WebHarbor/commit/3e755e70a7823d3c77648a67c3b80df338bec4ec
• .assets-revision was not modified; waiting for the HF-side asset review/merge outcome before pinning

Known limitations

• Product, category, and store visuals are deterministic local SVG assets rather than copied IKEA-owned imagery.
• No real payment, no real order placement, no real delivery booking, and no live external runtime calls.
• Playwright screenshots could not be captured on this machine because the Chromium download failed.

Manual review notes

• .gitignore runtime ignore lines were corrected so sites/*/instance/ and sites/*/scraped_data/ remain out of commits.
• The shell entrypoint was normalized to LF so the Docker image boots correctly on Linux.

[Susan571]

Review — Add IKEA WebHarbor mirror (#49)

Verdict: Request changes. Healthy in the container and idempotent, but the catalog ships
no product images, which fails the visual-fidelity bar for a furniture store. Plus the
shared asset-upload blocker.

Mechanical

• Registration synced in all 3 places; readiness counter correctly parametrized to 16.
  py_compile OK.
• ❌ Shared blocker: build.sh → check_assets.sh fails: missing
  sites/ikea/instance_seed; no ikea.tar.gz on HF; .assets-revision not bumped.
• ✅ After generating the seed and building: 16/16 → 200, /reset/ikea byte-identical,
  /reset-all ~1.3s. Seed idempotent (156 products / 15 stores / categories / rooms).

Visual fidelity: ❌ REQUEST CHANGES

Layout, nav, category/room structure, and store locator all match upstream IKEA IA (I
confirmed against live IKEA, which is reachable). But product <img> tags point at
images/products/… which don't exist, so every product renders as a flat colored-shape
placeholder (screenshots: ikea_products.png, ikea_product_hemlund.png). For a furniture
catalog the product imagery is core fidelity — please ship real product images in the asset
tarball (or reference scraped image URLs).

Functional depth: PASS

Search, category/room filtering, product detail, store locator, cart all work.

Task quality: PASS (with caveat)

Entities (products/stores/rooms) verified present and tasks are multi-step. Caveat: any task
that asks the agent to identify a product by its visual appearance is unsolvable while
images are placeholders — re-verify those after images land. Port :40015 correct.

Required before approval: (1) ship real product images; (2) upload ikea.tar.gz to HF +
bump .assets-revision.