chore: baseline import

2026-04-23 01:08:18 +08:00
commit c8ddfdf660
260 changed files with 117643 additions and 0 deletions
@@ -0,0 +1,96 @@
+# Seed Data Policy
+
+## Purpose
+Define how example content (text, images, product specs) is sourced when populating the new site for design review, CMS onboarding, and staging environments.
+
+## Core Rule
+Seed data should be scraped from the existing Eversolo website and used as the default example content wherever possible.
+
+Legacy sources:
+- https://www.eversolo.com/Index/index/l/en-us.html
+- https://www.eversolo.com/Index/index/l/zh-cn.html
+
+This applies to product imagery, product copy, specifications, support content, news, awards, and dealer lists.
+
+Do not generate placeholder marketing copy, do not use stock photography, and do not invent product values when a legitimate source exists on the current site.
+
+## Scope
+Seed data covers everything used for pre-launch environments:
+- Homepage sample content
+- Featured products
+- Full product detail pages (all 14 current models)
+- Downloads (firmware, manuals)
+- FAQ entries
+- Warranty and Contact text
+- About page
+- News and reviews
+- Dealer list
+
+## Sourcing Priority
+1. Current Eversolo website (EN and ZH, per locale)
+2. Official product packaging / datasheets already published by the brand
+3. Previously approved press releases
+4. CMS-authored fresh content (only when the above are unavailable)
+
+## What to Import per Product
+For every product already present on the legacy site, import:
+- Model name
+- English and Chinese descriptive copy
+- Hero image and gallery images
+- Key feature list
+- Specifications table
+- Firmware downloads (if listed)
+- Manuals (if listed)
+
+## What to Import per Page
+- Support page copy
+- Warranty policy text
+- Contact information and regional entries
+- About / brand story
+- News articles
+- Awards and review quotes
+- Dealer information by region
+
+## Image Handling
+- Pull images at the highest resolution available on the legacy pages.
+- Preserve original aspect ratios; do not up-scale.
+- Convert to AVIF/WebP at ingest time.
+- Re-crop into the new design system aspect ratios on CMS upload, not during scraping.
+- Keep the original file name or model identifier in the asset metadata so provenance is traceable.
+
+## Copy Handling
+- Preserve the original English and Chinese versions separately. They are not translations of each other on the legacy site and should be stored as independent locale content.
+- Light copy-editing is allowed to match the new tone of voice, but the factual content must stay intact.
+- Do not fabricate specs, awards, or review quotes. If a value is missing, leave the field blank rather than guessing.
+
+## Storage
+- Imported assets go into the `media`, `videos`, and `files` upload collections.
+- Structured fields (specs, downloads, FAQ) are written into their respective collections.
+- Seed content is marked with an internal tag `source: legacy-site` for traceability.
+
+## Seed Workflow
+1. Crawl the legacy site with a scoped scraper (no PII, no admin pages).
+2. Normalize structured data (specs, downloads, FAQ) into the new schema shapes.
+3. Upload media into Payload, preserving origin metadata.
+4. Create CMS entries in both locales for products and globals.
+5. Mark all seed entries as `status: draft` until editorial review.
+6. Editorial team reviews, refines, and promotes entries to `published`.
+
+## Legal / IP
+- All imported assets are Eversolo's own property; use is permitted.
+- Third-party review quotes and media logos must be verified for reuse rights before going live.
+- Amazon, JD, Tmall, and shop logos are not imported into the site's own media library.
+
+## Staging vs Production
+- Staging: seed data is fine; mark clearly in the environment banner.
+- Production: no raw seed imports; every product and page must pass editorial review before going live.
+
+## Refresh
+- Seed data is a one-time import for bootstrap, not an ongoing sync.
+- Future updates to product content happen in CMS, not in the scraper.
+- The scraper can be re-run if the legacy site is still live during a window when content is being continuously updated.
+
+## Open Items
+- Whether the scraper should be committed to the repo or kept as a one-off tool.
+- Whether to snapshot the legacy site as archived HTML for reference during editorial review.
+- Whether to scrape dealer lists in full or only a representative sample for seed.