fix(dla): name Wix pages from nav label or slug when every page shares the site-wide SEO title - #4720
Open
ellenbauer wants to merge 138 commits into
Open
fix(dla): name Wix pages from nav label or slug when every page shares the site-wide SEO title#4720ellenbauer wants to merge 138 commits into
ellenbauer wants to merge 138 commits into
Conversation
Keep the site create change focused on generic static source imports while preserving the Figma implementation on its dedicated branch. AI assistance: OpenAI gpt-5.6-sol via OpenCode removed the isolated Figma handoff scope and verified the focused CLI tests, lint, typecheck, and build. Chris Huber reviewed and remains responsible for the change.
AI assistance: OpenAI gpt-5.6-sol via OpenCode reduced the fixture after reproducing and timing the test locally; Chris Huber reviewed and remains responsible for the change.
AI assistance: OpenAI gpt-5.6-sol via OpenCode verified the Homeboy release asset and updated Studio's default importer URL. Chris Huber remains responsible for the change.
The export selected portable media against its own 160 MiB cap without knowing how many bytes the routes, captured render dependencies, and reports already take. On a 28-page Wix capture the media filled that cap and the artifact writer then threw "Portable capture exceeds compiler limits" against the 192 MiB total, discarding a three-minute capture instead of degrading. Reserve the staged route HTML, the captured non-script resources, and the report files that already exist on disk, and give media only what is left under the artifact limit (still capped at 160 MiB). The reservation is an upper bound: media that does not fit is retained external, as before, rather than failing the export. The receipt reports the effective `max_bytes` and the `reserved_bytes` behind it, and the artifact header carries the effective total limit. `ExportCaptureOptions.limits` lets a caller (and the test) override the artifact and media limits without touching the compiler defaults. Re-exporting the same Wix capture selects the same 167 files (24.3 MB) with 145 MB reserved and a 56 MB media budget; the artifact is unchanged. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The media downloader fetched with no Accept header, so format-negotiating CDNs (Wix, Squarespace, Cloudinary f_auto, imgix auto=format, Photon) returned the legacy/original encoding. On a small Wix store the retina variants came back as 1-7 MB PNGs instead of the ~10 KB AVIFs a visitor's browser receives, the portable media set filled its 160 MiB budget, and the artifact export failed against the 192 MiB compiler limit. Send Chromium's image Accept header and name the downloaded file by the negotiated content type when it differs from the URL extension, so an AVIF body is not stored as `.png`. Same-format aliases (`.jpeg`/`.jpg`, `.tif`/`.tiff`) keep the URL extension. On the same site the export now completes at 128 MB, with 162 AVIF assets in the generated theme instead of 82 multi-megabyte PNGs. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The generated import script wrote Static Site Importer's full response into `.studio-import/result.json` as `import_receipt`. On a 28-page Wix artifact that response was 3.04 GB (`import_report` 2.72 GB, `materialization_receipt` 325 MB). macOS `write(2)` rejects a single write of 2 GB or more, so `file_put_contents` returned false and the import failed with "result receipt could not be saved" after every page and the theme had already been materialized. Store the existing bounded projection (`static_site_importer_studio_result_projection`) as the receipt on both the success and failure paths. The CLI only reads `continuation`, `canonicalization_pending`, `completed_routes`, and `total_routes` from the receipt, and the database option already stores the same projection. With this change the same import completes, canonicalizes 32 pages, and writes a 124 KB receipt. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…th-scope-nimbus # Conflicts: # packages/data-liberation-agent/dist/mcp-server.bundle.mjs # packages/data-liberation-agent/dist/scripts/carry-reconstruct-drive.mjs # packages/data-liberation-agent/dist/scripts/carry-replica-shots.mjs # packages/data-liberation-agent/dist/scripts/chunk-6UJ5H2IN.mjs # packages/data-liberation-agent/dist/scripts/chunk-Q5FWHB6M.mjs # packages/data-liberation-agent/dist/scripts/chunk-RH45CRZA.mjs # packages/data-liberation-agent/dist/scripts/enrich-product-marketing.mjs # packages/data-liberation-agent/src/adapters/wix/discover.ts
When desktop and mobile captures share a responsive body signature but their style blocks differ, the export returned the desktop document unchanged: mobile styles were dropped entirely and desktop styles were left unwrapped, so the page shipped with no mobile CSS (proven on a Wix homepage whose DOM matched across viewports while its CSS did not). Keep the single shared body but merge CSS responsively: wrap desktop styles in media="(min-width:769px)" and append the mobile document's style blocks as media="(max-width:768px)", without the .data-liberation-mobile-document scoping that only exists in two-body output. Identical bodies with identical styles keep the current single-document output, and differing bodies keep the existing two-document behavior. Known limitation (out of scope): mobile-only body attributes/classes are still lost when the shared desktop body is kept. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Direct-artifact imports wrote continuation receipts carrying only
{continuation: true}, so the CLI printed 'Static site import progress:
0/0 routes' for the entire run (observed for ~25 minutes on a real
studio create --from <url>): the fabricated ?? 0 fallbacks were the only
data on that path, since url_batch_run counts exist only for URL-batch
imports and the direct path returns before canonical documents exist.
Enrich the continuation receipts with what is actually measured: the
dependency-preparation lifecycle checkpoint reports its stage, and the
direct ingest continuation passes through the importer's status plus
url_batch_run route counts only when the importer provides them. The
CLI now prints real route counts when present, otherwise the reported
stage ('Static site import continuing: dependencies prepared'), and
never fabricates 0/0 route numbers.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Captured routes that resolve to the same website path now collapse onto the already-claimed route when the alternate address declares it as canonical (e.g. / and /index.html serving one document). The receipt records duplicateRoutes so aliases stay auditable. Non-canonical collisions (/about vs /about/) still fail with the existing error. Fixes Automattic/data-liberation-agent#107
…s the site-wide SEO title Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GpLtieepeCxY3hXzJfsfoV
chubes4
force-pushed
the
feat/site-artifact-import-cli
branch
from
August 30, 2026 05:03
43bc3f9 to
b7abf7d
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Related issues
feat/site-artifact-import-clibranch)How AI was used in this PR
Found, fixed, and tested with Claude Code while migrating a real Wix site (www.360chiro.co.uk). The helper and test were written test-first; I reviewed the diff and the evidence below comes from an actual import.
Proposed Changes
document.titlethen carries no page identity, and the existing" | Site Name"suffix-strip has nothing to strip — so every imported page lands in wp-admin with the same title. On www.360chiro.co.uk, all 14 pages imported as "Chiropractor Sheffield - Expert Care for Pain Relief", making the Pages list unusable.resolvePageTitle(inlib/html-extract, reusable by other adapters): after the suffix-strip, a title that equals the discovered site-wide title is demoted to the page's navigation label (what the Wix editor shows as the page name, e.g./treatments→ "Treatments",/new-patients→ "First Visit"), else the humanized slug ("neck-pain-stiffness" → "Neck Pain Stiffness"). The homepage keeps the site title unless the nav names it ("Home").seoTitle(_seo_titlemeta) — this only changes the WordPress admin title, not SEO data.siteMeta.title,navigation), so no new capture work is needed.Testing Instructions
cd packages/data-liberation-agent && npx vitest run test/html-extract.test.ts test/adapters/wix.test.ts(8 new + 8 existing tests).Pre-merge Checklist
test/html-extract.test.ts)tsc --noEmitclean for the packageNote: the full package suite has 9 pre-existing failures in
test/legacy-scripts.test.tson my machine only because the checkout path contains a space — the test shellsnode --check ${path}unquoted. Unrelated to this change; happy to fix that in a follow-up.🤖 Generated with Claude Code
https://claude.ai/code/session_01GpLtieepeCxY3hXzJfsfoV