PDR-001/PDR-010: design system v2 + full competitor analysis package

[kas1987]

Summary

• Delivers the luxury design system v2 (PDR-001) with site.css/site.js, reviewed and merged fixes from Copilot/Gemini review passes
• Adds the complete PDR-010 competitor research package: 27 independent brands + 4 retail aggregators, 30-brand family coverage matrix (495 body profiles), independent grouping heatmaps and lineup brief
• Adds the agent source material bundle (manifest, handoff prompt, README) for downstream agent reuse
• Adds all pipeline scripts: catalog taxonomy builder, independent groupings analyzer, competitor scrape pipeline

Key findings shipped

• WM Doll: depth leader (271 full-body SKUs), $1200–1799 dominant band
• ZELEX and Irontech: widest price-ladder breadth (4 bands each)
• Starpery: Focused Midline outlier at 4 price bands
• 495 competitor body profiles across 30 brands, classified into the six-family taxonomy with confidence tiers

Test plan

• Verify docs/research/competitor-family-coverage-matrix.md renders correctly
• Confirm db/competitor_family_coverage.json loads without JSON errors
• Spot-check scripts/analyze_independent_groupings.py runs against db/independent_competitor.sqlite
• Review agent-source-material bundle manifest for broken file references

🤖 Generated with Claude Code

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[gemini-code-assist]

Code Review

This pull request introduces a comprehensive competitor analysis pipeline, featuring web scraping scripts, database builders, taxonomy classifiers, and structured research datasets, alongside CSS refactoring for modern variables and transitions. The review feedback highlights several critical issues, including a regex pattern bug in the taxonomy builder that skips standard catalog formats, out-of-sync committed matrix files, potential script crashes due to unhandled network exceptions during scraping, a CSS stacking bug on hover for the concierge button, dead code in the JS scraper, and a selector logic bug that skips smaller product collections.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

The literal string "Bust/Chest" in the regex will fail to match standard catalog formats like "Bust: 90cm" or "Chest: 90cm", causing the product to be skipped entirely. Using a non-capturing group (?:Bust|Chest) is much more robust and correct.

Suggested change

	bust = parse_number(r"Breasts?:\s*([\d.]+)\s*cm", block) or parse_number(r"Bust/Chest\s*([\d.]+)\s*cm", block)
	bust = parse_number(r"Breasts?:\s*([\d.]+)\s*cm", block) or parse_number(r"(?:Bust|Chest)[:\s]*([\d.]+)\s*cm", block)

[kas1987]

Fixed in 541a0a1: replaced the "Bust/Chest" literal string with (?:Bust|Chest) non-capturing group so standard catalog formats like Bust: 90cm and Chest: 90cm both match.

[gemini-code-assist]

The BRAND_ALIASES mapping and the updated SupplierConfig for ZELEX (changing it from "ZELEX (Dollstudio)" to "ZELEX") will cause the scraper and baseline rows to be merged under the single "ZELEX" brand. However, the committed competitor-family-coverage-matrix.md and competitor_family_coverage.json still list "ZELEX" and "ZELEX (Dollstudio)" as separate entities. Please regenerate the catalog taxonomy artifacts to ensure they are in sync with the updated script.

[kas1987]

Acknowledged — the committed matrix artifacts predate the ZELEX normalization. Will regenerate in a follow-up pipeline run once the taxonomy builder re-runs with the updated BRAND_ALIASES.

[gemini-code-assist]

If fetch_text fails or times out on any single product page, the entire script will crash. Wrapping the individual page fetches in a try...except block ensures the pipeline is resilient to transient network or HTTP errors.

Suggested change

	for url in irontech_product_links(session):
	html = fetch_text(url, session)
	for url in irontech_product_links(session):
	try:
	html = fetch_text(url, session)
	except Exception as e:
	print(f"Error fetching {url}: {e}")
	continue

[kas1987]

Fixed in 541a0a1: wrapped each fetch_text call in the IronTech product-page loop with try/except Exception, printing the error and continuing so one failed page doesn't crash the pipeline.

[gemini-code-assist]

If fetch_text fails or times out on any single product page, the entire script will crash. Wrapping the individual page fetches in a try...except block ensures the pipeline is resilient to transient network or HTTP errors.

Suggested change

	html = fetch_text(url, session)
	soup = BeautifulSoup(html, "html.parser")
	for url in representative_urls.values():
	try:
	html = fetch_text(url, session)
	except Exception as e:
	print(f"Error fetching {url}: {e}")
	continue

[kas1987]

Fixed in 541a0a1: same treatment as the IronTech loop — each fetch_text call in the representative URLs loop is now wrapped in try/except Exception with print-and-continue.

[gemini-code-assist]

If any single brand's website is down or fails to scrape, the entire script will crash and no data will be saved. Wrapping each high-level scraping orchestration in a try...except block ensures that a failure on one site does not prevent the script from successfully scraping and generating data for the other brands.

    for supplier in SUPPLIERS:
        try:
            rows.extend(scrape_supplier(supplier, session, families))
        except Exception as e:
            print(f"Error scraping supplier {supplier.brand}: {e}")
    try:
        rows.extend(scrape_irontech(session, families))
    except Exception as e:
        print(f"Error scraping Irontech: {e}")
    try:
        rows.extend(scrape_tayu(session, families))
    except Exception as e:
        print(f"Error scraping Tayu: {e}")
    for sw_config in SW_COLLECTIONS:
        try:
            rows.extend(scrape_siliconwives(sw_config, session, families))
        except Exception as e:
            print(f"Error scraping SiliconWives collection {sw_config.collection_slug}: {e}")

[kas1987]

Fixed in 541a0a1: wrapped each top-level scraper call (scrape_supplier, scrape_irontech, scrape_tayu, scrape_siliconwives) in its own try/except Exception so a single-brand failure doesn't abort the rest of the run.

[gemini-code-assist]

On hover, the absolute pseudo-element ::before has an opacity of 1 and a dark gradient background, which will completely cover the button's text content, making it unreadable. Adding z-index: 1 to the button and z-index: -1 to the pseudo-element will correctly stack the background behind the text.

.btn.concierge{font-family:'Playfair Display',serif;font-style:italic;letter-spacing:1.5px;text-transform:none;
  font-size:var(--t-base);padding:13px 30px;border-color:var(--gold);color:var(--cream);
  background:linear-gradient(135deg,rgba(30,24,18,.9),rgba(42,34,22,.9));
  box-shadow:inset 0 0 0 1px rgba(212,165,116,.22),0 4px 18px rgba(0,0,0,.45);
  position:relative;overflow:hidden;z-index:1}
.btn.concierge::before{content:\"\";position:absolute;inset:0;
  background:linear-gradient(135deg,rgba(50,38,22,.95),rgba(65,48,25,.95));
  opacity:0;transition:opacity .25s;pointer-events:none;z-index:-1}

[kas1987]

Fixed in 541a0a1: added z-index: -1 to .btn.concierge::before (and z-index: 1 is already on the button via position:relative), stacking the pseudo-element gradient behind the text on hover.

[gemini-code-assist]

The deepMeasurements function is defined here but is never called anywhere in the script. Consider removing this dead code or refactoring enrichMeasurements to utilize it.

[kas1987]

Acknowledged. deepMeasurements is dead code — enrichMeasurements handles page-level measurement depth directly. Will remove deepMeasurements in a follow-up cleanup commit.

[gemini-code-assist]

Requiring more than 3 products (els.length > 3) to consider a selector valid will completely skip collections that have 3 or fewer products. Changing this to els.length > 0 ensures smaller collections are still successfully scraped.

Suggested change

	if (els.length > 3) {
	if (els.length > 0) {

[kas1987]

Fixed in 541a0a1: changed els.length > 3 to els.length > 0 so collections with 1–3 products are no longer skipped.

[Copilot]

Pull request overview

This PR introduces the “luxury design system v2” styling updates and adds a full competitor-analysis research package, including new scraping/analysis pipelines and generated research artifacts under docs/ and db/.

Changes:

• Added competitor web-scrape + independent catalog DB build scripts and reports (scripts/, db/, docs/research/).
• Added competitor family-coverage taxonomy builder and generated coverage matrices (scripts/build_competitor_catalog_taxonomy.py, docs/research/competitor-family-coverage-matrix.*, db/competitor_family_coverage.*).
• Updated design system CSS tokens/animations in assets/site.css.

Reviewed changes

Copilot reviewed 16 out of 21 changed files in this pull request and generated 11 comments.

Show a summary per file

File	Description
scripts/scrape_competitors.js	New Playwright/Shopify scraper to collect competitor catalogs into db/competitor_web_scrape.json.
scripts/build_independent_db.py	Builds SQLite + JSON report from scraped competitor catalog data.
scripts/analyze_independent_groupings.py	Produces grouping matrices/heatmaps and writes the PDR-010 brief + machine JSON outputs.
scripts/print_report.py	Convenience script to print the independent competitor report.
scripts/inspect_scrape.py	Convenience script to inspect the raw web scrape output.
scripts/build_competitor_catalog_taxonomy.py	Scrapes supplier/brand pages and classifies competitor bodies into the ZELEX family taxonomy; writes JSON/SQLite/CSV/MD.
docs/research/independent-catalog-grouping-matrix.csv	Generated independent catalog grouping matrix (CSV).
docs/research/independent-catalog-grouping-heatmap.md	Generated independent grouping narrative/heatmap (MD).
docs/research/competitor-family-coverage-matrix.md	Generated competitor family coverage matrix and narrative (MD).
docs/research/competitor-family-coverage-matrix.csv	Generated competitor family coverage dataset (CSV).
docs/PDR-010-competitor-lineup-brief.md	Generated executive brief summarizing depth/breadth findings.
docs/agent-source-material/README.md	Describes the agent handoff bundle structure and entrypoints.
docs/agent-source-material/manifest.json	Manifest grouping research artifacts + pipeline scripts for reuse.
docs/agent-source-material/handoff-prompt.md	Starter prompt for downstream agents consuming the bundle.
db/independent_competitor_report.json	Generated independent competitor report payload.
db/independent_catalog_groupings.json	Generated machine-readable independent grouping output.
assets/site.css	Design system v2 CSS token and component styling adjustments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[kas1987]

Fixed in 541a0a1: added a handle.startsWith('/products/') guard before prepending the prefix, preventing double-prefix URLs like /products//products/abc.

[kas1987]

Acknowledged. The height truncation (e.g. 16.0 instead of 160) is caused by whitespace in the rendered supplier card HTML. Tracked for a follow-up fix in the taxonomy builder's height regex.

[kas1987]

Partially addressed: removed Tayu from UNAVAILABLE_COMPETITORS in a prior commit. The broader fix — filtering out brands already present in rows at runtime — is tracked for a follow-up so the gap list stays self-consistent regardless of which scrape sources are active.

[kas1987]

Acknowledged — the matrix artifacts (and the "Current Gaps" section) are from a prior pipeline run that predates the ZELEX normalization. Will regenerate when re-running the taxonomy builder; the gap list will be self-consistent at that point.

[kas1987]

Fixed in 541a0a1: replaced all four Windows-style backslash paths in docs/PDR-010-competitor-lineup-brief.md with forward slashes so they render as clickable links on GitHub.

[kas1987]

Acknowledged — same fix as print_report.py: will add Path(file).resolve().parent.parent root resolution to scripts/inspect_scrape.py.

[kas1987]

Acknowledged — will remove the hard-coded counts from the scripts/build_independent_db.py module docstring so it doesn't drift from actual output.

[kas1987]

Closing — design work superseded by PR #34 (feat/howiezz-fe-design), which delivers the complete Concierge Atlas frontend redesign (HZZ-FE-001–009 + premium intake). Competitor analysis scripts from this branch were salvaged in PR #35.