Skip to content

Hycarfilter Phase 4 Delivery Report: data.hycarfilter product-database sub-site

Produced: 2026-10-25 ｜ Author: data-product designer + export-filter industry editor (Xiaozhu ✍️) Repo: data.hycarfilter (MkDocs · product-database core) ｜ Basis: docs/hycarfilter-p1-inventory-architecture.md §2.2 + main-site src/data/products.json

1. Delivery overview

The MkDocs Material data sub-site is live, positioned as "the foundational data layer that is searched, machine-read and cited", with zero marketing and zero duplication of the main site's marketing copy. The canonical product dataset has been moved into this repo as the single source of truth.

2. MkDocs page count (37 content pages + 1 404)

Section Pages Notes
Home 1 Database positioning / coverage / data notes (no marketing copy)
SKU full index 1 22-model master table: model/category/dimensions/efficiency/filter media/HS Code
Category pages 7 Per-category full SKU spec table + category technical parameter definitions
SKU detail pages 22 Complete field spec (identification/dimensions/materials/performance/fitment/packing/export)
Cross-reference 3 Vehicle fitment mapping / OE cross-reference framework / export packing data
Data specification 3 Field schema / naming rules / changelog
Total 37

3. Build script notes

scripts/build-db-pages.py (Python 3, no third-party dependencies) reads data/products.json and generates all data pages in one pass:

  • docs/products/index.md — SKU full index
  • docs/products/{slug}.md ×7 — category pages (SKU table + category technical parameter definitions)
  • docs/products/{model}.md ×22 — SKU detail pages (full field spec)
  • docs/crossref/vehicle-fitment.md — model ↔ vehicle type (forward + reverse index)
  • docs/crossref/oe-cross-reference.md — model ↔ OE cross-reference framework (marked reference-only)
  • docs/crossref/export-packing.md — packing / carton quantity (authoritative) + pallet / 40HQ (script estimates)

Estimation logic: pallet 1200×1000 mm / load height 160 cm; 40HQ usable volume ≈ 65 m³; carton dimensions are per-category assumptions inside the script (some models override), and do not enter the authoritative dataset — packing / carton quantity / net weight are authoritative values, while pallet / 40HQ figures are estimates and are clearly marked with assumptions on the page.

4. Consistency mechanism

  1. Single source of truth: data/products.json is the single source of truth for the main site + the three sub-sites; every value is defined in exactly one place.
  2. Identical authoritative copy: this repo's data/products.json is byte-identical with the main site's src/data/products.json (verified by diff — IDENTICAL, 0 differences).
  3. Zero manual transcription: all data pages are rendered by the script; no value is transcribed by hand.
  4. Sync flow (written into README): main site changes data → copy products.json into this repo → python3 scripts/build-db-pages.py → mkdocs build → commit & push (p4-data-* marker).

5. build / commit / push

  • build: mkdocs build ✅ 37 content pages (+ 404), 1.12s, Material theme.
  • commit: 449a9ba p4-data: scaffold canonical product database (MkDocs Material).
  • push: origin/main (new branch main pushed).
  • sync: main site and data site products.json — 0 differences.

6. Compact summary

MkDocs pages: 37 content pages = home 1 · SKU index 1 · categories 7 · SKU detail pages 22 · cross-reference 3 (vehicle fitment / OE cross-reference / export packing) · data specification 3 (field schema / naming rules / changelog).

Build script: scripts/build-db-pages.py generates all data pages (index / category / SKU / cross-reference) from data/products.json in one pass, with zero manual transcription; pallet / 40HQ figures are script estimates, clearly marked with assumptions, and do not pollute the authoritative data.

Consistency mechanism: data/products.json is the single source of truth + byte-identical with the main site (diff 0 differences) + script-regenerated pages guarantee identical values across the two repos; README documents the sync flow (change data → copy → rebuild → commit).

build: mkdocs build, 37 pages all green. commit: p4-data: scaffold canonical product database. push: origin/main. sync: 0/0.

Report path: docs/p4-data-site-report.md (this document).