Add IKEA furniture data research report

2026-02-07 12:50:33 +01:00
parent 53ee0fc1ec
commit cf0fe586eb
1 changed files with 352 additions and 0 deletions
--- a/RESEARCH-ikea-data.md
+++ b/RESEARCH-ikea-data.md
@@ -0,0 +1,352 @@
+# IKEA Furniture Data Import — Research Report
+
+**Task:** t-05dd7
+**Date:** 2026-02-07
+**Researcher:** researcher role
+
+## Executive Summary
+
+**There is no official public IKEA API for furniture data.** However, there are multiple viable paths to get IKEA product dimensions and 3D models into our room planner. The recommended approach is a **tiered strategy**: start with a curated hand-built catalog of common IKEA items using known dimensions, with optional GLB model loading for enhanced visuals later.
+
+---
+
+## 1. IKEA APIs (Official)
+
+### IKEA Has No Public Developer API
+
+IKEA does not offer a public REST/GraphQL API for developers. There is no developer portal, no API keys, no official documentation.
+
+### Undocumented Internal Endpoints
+
+IKEA's website uses internal APIs that have been partially reverse-engineered:
+
+| Endpoint | Purpose | Format |
+|----------|---------|--------|
+| `sik.search.blue.cdtapps.com/{country}/{lang}/search-result-page?q={query}&size=24&types=PRODUCT` | Product search | JSON |
+| `www.ikea.com/{country}/{lang}/products/{itemNo[5:]}/{itemNo}.json` | Product detail (PIP) | JSON |
+| `api.ingka.ikea.com/cia/availabilities/{country}/{lang}?itemNos={itemNo}` | Stock/availability | JSON |
+| `www.ikea.com/{region}/{country}/iows/catalog/availability/{itemNo}` | Legacy availability (deprecated Dec 2021) | XML |
+| `web-api.ikea.com/{country}/{lang}/rotera/data/exists/{itemNo}/` | 3D model existence check | JSON |
+| `web-api.ikea.com/{country}/{lang}/rotera/data/model/{itemNo}/` | 3D model metadata + CDN URL | JSON |
+
+The **Rotera API** is most relevant for us — it returns a `modelUrl` field pointing to the actual GLB file on IKEA's CDN. This is what the Blender IKEA Browser add-on uses internally.
+
+**IKEA Kreativ**: Their AI-powered room planner at `ikea.com/us/en/home-design/` uses AWS infrastructure (API Gateway at `ddh79d7xh5.execute-api.us-east-1.amazonaws.com`). No public API.
+
+**Status**: All endpoints are undocumented, unsupported, and may break at any time. Rate limiting and anti-bot measures exist. The IOWS endpoints are deprecated.
+
+**Sources**: https://sbmueller.github.io/posts/ikea/ | Postman workspace: https://www.postman.com/galactic-shuttle-566922/ikea/overview
+
+### Archived Python Client
+
+**vrslev/ikea-api-client** (GitHub, 1k+ stars) — Python client for IKEA internal APIs supporting cart, search, stock, item info, and 3D models. **Archived October 2024** — no longer maintained. MIT license.
+
+- Supported: stock lookup, product search, item specs, 3D model retrieval (`RoteraItem` endpoint)
+- Install: `pip install ikea-api[httpx]`
+- **Not recommended** for new projects since archived
+
+**Source**: https://github.com/vrslev/ikea-api-client
+
+### npm: ikea-availability-checker (Active)
+
+**Ephigenia/ikea-availability-checker** — Node.js package for checking IKEA stock across 40+ countries and 400+ stores. Actively maintained.
+
+```bash
+npm install ikea-availability-checker
+# CLI: ikea-availability-checker stock --country us 30457903
+```
+
+Not directly useful for dimensions/models but proves Node.js access to IKEA endpoints is feasible.
+
+**Source**: https://github.com/Ephigenia/ikea-availability-checker
+
+---
+
+## 2. Open Datasets with Dimensions
+
+### Best: Hugging Face — IKEA US CommerceTXT (30,511 products)
+
+- **Records**: 30,511 products across 632 categories
+- **Fields**: Name, SKU, Price, Dimensions (Width/Height/Depth), Materials, Category, Images, URLs
+- **Format**: CommerceTXT v1.0.1 (text-based, parseable) + Parquet (56.5 MB)
+- **License**: CC0 1.0 (Public Domain) — fully usable
+- **Date**: July 15, 2025
+- **Dimensions**: Yes, in the `@SPECS` section (e.g., `Width: 12", Height: 16"`)
+
+```python
+from datasets import load_dataset
+dataset = load_dataset("tsazan/ikea-us-commercetxt")
+```
+
+**Source**: https://huggingface.co/datasets/tsazan/ikea-us-commercetxt
+
+### Good: Kaggle — IKEA SA Furniture (2,962 items)
+
+- **Records**: 2,962 items
+- **Fields**: name, category, price, width (cm), height (cm), depth (cm), designer, description
+- **Format**: CSV
+- **License**: Kaggle dataset terms
+- **Date**: April 2020 (somewhat dated)
+- **Dimensions**: Yes — width, height, depth in centimeters
+
+**Source**: https://www.kaggle.com/datasets/ahmedkallam/ikea-sa-furniture-web-scraping
+
+### Good: GitHub — IKEA Dataset (12,600+ items)
+
+- **Records**: 12,600+ images with furniture dimensions
+- **Fields**: Name, dimensions (width/length/height cm), material
+- **Format**: ZIP archives + Python scraping script
+- **Organization**: By room (bedroom, bathroom, kitchen, living room, etc.)
+- **Date**: January 2020
+
+**Source**: https://github.com/valexande/IKEA-Dataset
+
+### Reference: Dimensions.com — IKEA Collection
+
+- **Content**: Dozens of IKEA products with precise dimensions + technical drawings
+- **Formats**: DWG, SVG, JPG (2D) + 3DM, OBJ, SKP (3D models)
+- **Coverage**: KALLAX, HEMNES, BESTÅ, BEKANT, and many more series
+- **Dual units**: Both inches and centimeters
+- **Includes**: Weight capacities
+- **Note**: Good for manual reference but not easily machine-parseable without scraping
+
+**Source**: https://www.dimensions.com/collection/ikea-furniture
+
+---
+
+## 3. IKEA 3D Models
+
+### Web-Sourced GLB Models (Best Path for Three.js)
+
+IKEA product pages with a "View in 3D" button serve simplified, Draco-compressed GLB (binary glTF) files. These are:
+- **Format**: GLB (binary glTF 2.0) — native Three.js format
+- **Quality**: "Lowest quality versions" — simplified from manufacturing models for web delivery
+- **Properties**: Scaled in metric, basic materials applied, KHR_Texture_Transform extension
+- **Compatibility**: Direct import into Three.js via `GLTFLoader` + `DRACOLoader`
+
+#### Tools to Download IKEA GLB Models
+
+| Tool | Type | Stars | Status | License |
+|------|------|-------|--------|---------|
+| [IKEA 3D Model Download Button](https://github.com/apinanaivot/IKEA-3D-Model-Download-Button) | Tampermonkey userscript | 941 | Active (Jan 2026) | — |
+| [IKEA 3D Model Batch Downloader](https://github.com/apinanaivot/IKEA-3d-model-batch-downloader) | Python (Selenium+BS4) | — | Active | GPL-3.0 |
+| [IKEA Browser for Blender](https://extensions.blender.org/add-ons/ikea-browser/) | Blender addon | 95k downloads | Active (v0.4.0, Oct 2025) | GPL-3.0 |
+| [Chrome Extension](https://chromewebstore.google.com/detail/ikea-3d-models-downloader/kihdhjecfjagoplfnnpdbbnhkjhnlfeg) | Chrome extension | — | Active | — |
+
+The **batch downloader** workflow:
+1. Give it an IKEA category URL (e.g., `https://www.ikea.com/fi/fi/cat/chairs-fu002/`)
+2. Scrapes product links → extracts 3D model URLs → downloads GLB files
+3. Stores metadata in SQLite database
+4. Deduplication built-in
+
+**Note**: Not all IKEA products have 3D models. Only items with the "View in 3D" button are available.
+
+### Official IKEA 3D Assembly Dataset (GitHub)
+
+- **Source**: https://github.com/IKEA/IKEA3DAssemblyDataset
+- **Items**: Only 5 products (LACK, EKET, BEKVÄM, DALFRED)
+- **Formats**: GLB/glTF + OBJ + PDF assembly instructions
+- **License**: CC BY-NC-SA 4.0 — **non-commercial only, research purposes**
+- **Verdict**: Too few items and restrictive license. Not suitable for our use.
+
+### Sweet Home 3D IKEA Libraries
+
+- **180 models**: https://3deshop.blogscopia.com/180-ikea-models-for-sweethome3d/
+- **342 models (bundle)**: https://3deshop.blogscopia.com/ikea-bundle-342-models/
+- **Format**: SH3F (Sweet Home 3D format) — would need conversion
+- **Quality**: Community-created, based on real IKEA products
+- **Note**: Format is proprietary to Sweet Home 3D, would require extraction/conversion
+
+### Third-Party Model Sources
+
+- **Sketchfab**: Search "IKEA" — many free community models, various licenses
+- **TurboSquid**: Free IKEA models available
+- **Clara.io**: Some IKEA models
+- **Trimble 3D Warehouse**: IKEA models available (SketchUp format)
+
+---
+
+## 4. Web Scraping Feasibility
+
+### Existing Scrapers
+
+| Project | Stack | Notes |
+|---------|-------|-------|
+| [IKEA Scraper](https://github.com/Abdelrahman-Hekal/IKEA_Scraper) | Python | Full scraper for ikea.com |
+| [ikea-webscraper](https://github.com/gamladz/ikea-webscraper) | Selenium | Product info extraction |
+| [IKEA-project](https://github.com/furkansenn/IKEA-project) | Selenium | Product scraping |
+| [ikea-scraper](https://github.com/bonzi/ikea-scraper) | — | IKEA data extraction |
+
+### Technical Challenges
+
+- IKEA uses heavy client-side rendering (React/Next.js) requiring browser automation (Selenium)
+- Anti-bot protections: CAPTCHAs, rate limiting, IP blocking
+- Dimension data is embedded in product pages in varying formats
+- Product IDs are 8-digit numbers but URL structure varies by locale
+
+### Legal Considerations
+
+- IKEA's Terms of Service prohibit automated scraping
+- IKEA actively protects trademarks and trade dress under US Trademark Act
+- For personal/non-commercial home planning use, tools like the Blender extension operate in a gray area
+- The HuggingFace CommerceTXT dataset (CC0) is the safest legal path for product data
+
+---
+
+## 5. Standard IKEA Dimensions Reference
+
+Manually verified dimensions for the most common IKEA product lines:
+
+### Storage
+
+| Product | Width (cm) | Depth (cm) | Height (cm) |
+|---------|-----------|-----------|------------|
+| KALLAX 1×4 | 42 | 39 | 147 |
+| KALLAX 2×2 | 77 | 39 | 77 |
+| KALLAX 2×4 | 77 | 39 | 147 |
+| KALLAX 4×4 | 147 | 39 | 147 |
+| BILLY bookcase (standard) | 80 | 28 | 202 |
+| BILLY bookcase (narrow) | 40 | 28 | 202 |
+| BILLY bookcase (short) | 80 | 28 | 106 |
+| HEMNES 6-drawer dresser | 108 | 50 | 131 |
+| HEMNES 3-drawer dresser | 108 | 50 | 96 |
+| HEMNES bookcase | 90 | 37 | 197 |
+| BESTÅ TV bench | 120/180 | 42 | 38 |
+| MALM 6-drawer dresser | 80 | 48 | 123 |
+| MALM 4-drawer dresser | 80 | 48 | 100 |
+| PAX wardrobe (standard) | 100/150/200 | 58 | 201/236 |
+
+### Tables
+
+| Product | Width (cm) | Depth (cm) | Height (cm) |
+|---------|-----------|-----------|------------|
+| LACK side table | 55 | 55 | 45 |
+| LACK coffee table | 90 | 55 | 45 |
+| LACK TV bench | 90 | 26 | 45 |
+| LISABO desk | 118 | 45 | 74 |
+| BEKANT desk (rect) | 120/140/160 | 80 | 65-85 |
+| LINNMON/ALEX desk | 150 | 75 | 73 |
+| MELLTORP dining table | 125 | 75 | 74 |
+| EKEDALEN ext. table | 120-180 | 80 | 75 |
+
+### Seating
+
+| Product | Width (cm) | Depth (cm) | Height (cm) |
+|---------|-----------|-----------|------------|
+| POÄNG armchair | 68 | 82 | 100 |
+| STRANDMON wing chair | 82 | 96 | 101 |
+| KLIPPAN 2-seat sofa | 180 | 88 | 66 |
+| EKTORP 3-seat sofa | 218 | 88 | 88 |
+| KIVIK 3-seat sofa | 228 | 95 | 83 |
+| MARKUS office chair | 62 | 60 | 129-140 |
+
+### Beds
+
+| Product | Width (cm) | Depth (cm) | Height (cm) |
+|---------|-----------|-----------|------------|
+| MALM bed (queen, 160) | 160 | 209 | 92 (headboard) |
+| MALM bed (king, 180) | 180 | 209 | 92 (headboard) |
+| MALM bed (single, 90) | 90 | 209 | 92 (headboard) |
+| HEMNES bed (queen) | 163 | 211 | 66 (headboard 112) |
+| KURA reversible bed | 99 | 209 | 116 |
+| SUNDVIK child bed | 80 | 167 | 83 |
+
+### Kitchen
+
+| Product | Standard Width (cm) | Depth (cm) | Height (cm) |
+|---------|-----------|-----------|------------|
+| METOD base cabinet | 60/80 | 60 | 80 |
+| METOD wall cabinet | 60/80 | 37 | 60/80/100 |
+| METOD tall cabinet | 60 | 60 | 200/220 |
+| KNOXHULT base cabinet | 120/180 | 61 | 85 |
+| VADHOLMA kitchen island | 126 | 79 | 90 |
+
+---
+
+## 6. Recommendations for Our Project
+
+### Recommended Approach: Tiered Strategy
+
+#### Tier 1: Curated IKEA Catalog (Immediate)
+
+Extend `data/furniture-catalog.json` with IKEA-specific items:
+- Add 30-50 most popular IKEA products with verified dimensions
+- Use our existing procedural mesh format (box geometry parts)
+- Add `ikeaId` field (8-digit product number) for future linking
+- Add `ikeaSeries` field (KALLAX, BILLY, MALM, etc.)
+- Add `ikeaUrl` field for reference
+
+**Schema extension**:
+```json
+{
+  "id": "ikea-kallax-2x4",
+  "name": "KALLAX Regal 2×4",
+  "ikeaId": "80275887",
+  "ikeaSeries": "KALLAX",
+  "ikeaUrl": "https://www.ikea.com/de/de/p/kallax-regal-weiss-80275887/",
+  "category": "storage",
+  "rooms": ["wohnzimmer", "arbeitszimmer", "kinderzimmer"],
+  "dimensions": { "width": 0.77, "depth": 0.39, "height": 1.47 },
+  "variants": [
+    { "color": "white", "hex": "#ffffff" },
+    { "color": "black-brown", "hex": "#3c3028" },
+    { "color": "white-stained-oak", "hex": "#d4c4a8" }
+  ],
+  "mesh": { ... }
+}
+```
+
+**Effort**: ~2-3 hours to add 30 items manually from IKEA website
+**Risk**: None — uses verified public dimensions
+
+#### Tier 2: GLB Model Import (Enhancement)
+
+Add optional GLB model loading to the renderer:
+- Use Three.js `GLTFLoader` + `DRACOLoader` to load IKEA GLB files
+- Users download GLB files themselves (personal use) via browser extension
+- Store in local `models/` directory
+- Fall back to procedural mesh if GLB not available
+
+```javascript
+// In catalog entry:
+{
+  "id": "ikea-kallax-2x4",
+  "model3d": "models/ikea/kallax-2x4.glb",  // optional
+  "mesh": { ... }  // fallback procedural geometry
+}
+```
+
+**Effort**: ~4-6 hours for GLTFLoader integration
+**Risk**: Low — GLB loading is standard Three.js functionality
+
+#### Tier 3: Dataset Import Tool (Future)
+
+Build a converter that parses the HuggingFace CommerceTXT dataset:
+- Extract product dimensions from `@SPECS` section
+- Map categories to our catalog format
+- Auto-generate procedural meshes from dimensions
+- Run offline as a data pipeline
+
+**Effort**: ~1-2 days
+**Risk**: Medium — dimension parsing from free-text specs varies in reliability
+
+### What NOT to Do
+
+1. **Don't scrape IKEA live** — Legal risk, fragile, unnecessary when datasets exist
+2. **Don't depend on undocumented APIs** — They change without notice
+3. **Don't bundle IKEA 3D models** — Trademark/IP issues. Let users provide their own
+4. **Don't use the IKEA 3D Assembly Dataset for the app** — License is NC-SA research only
+
+---
+
+## 7. Key Sources
+
+| Resource | URL | Best For |
+|----------|-----|----------|
+| HuggingFace IKEA US | https://huggingface.co/datasets/tsazan/ikea-us-commercetxt | Dimension data (CC0) |
+| Kaggle IKEA SA | https://www.kaggle.com/datasets/ahmedkallam/ikea-sa-furniture-web-scraping | Structured CSV with dims |
+| Dimensions.com IKEA | https://www.dimensions.com/collection/ikea-furniture | Reference dimensions + drawings |
+| 3D Batch Downloader | https://github.com/apinanaivot/IKEA-3d-model-batch-downloader | Downloading GLB models |
+| Blender IKEA Browser | https://extensions.blender.org/add-ons/ikea-browser/ | Model inspection/conversion |
+| IKEA API Blog | https://sbmueller.github.io/posts/ikea/ | Understanding IKEA endpoints |
+| Sweet Home 3D IKEA | https://3deshop.blogscopia.com/ikea-bundle-342-models/ | Pre-made 3D models (needs conversion) |