sjs/spirit-tracker
1
0
Fork 0
mirror of https://github.com/samsonjs/spirit-tracker.git synced 2026-04-27 15:07:43 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: Willow Park parsing issues

Browse source
This commit is contained in:
Brennan Wilkes (Text Groove) 2026-01-31 14:27:12 -08:00
parent f19c1404fa
commit 7df4e48b9f
1 changed files with 28 additions and 23 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

41
src/stores/willowpark.js
View file

@ -3,16 +3,29 @@
const { decodeHtml, stripTags, extractFirstImgUrl, cleanText } = require("../utils/html"); const { decodeHtml, stripTags, extractFirstImgUrl, cleanText } = require("../utils/html");
const { makePageUrlShopifyQueryPage } = require("../utils/url"); const { makePageUrlShopifyQueryPage } = require("../utils/url");
function extractSkuFromUrlOrHref(hrefOrUrl) { function extractSkuFromUrlOrHref(hrefOrUrl) {
const s = String(hrefOrUrl || ""); const s = String(hrefOrUrl || "");
// Common Willow patterns: // /products/<handle>-123456 or /collections/.../products/<handle>-123456
// /products/<handle>-123456
// /collections/rum/products/<handle>-123456
// Also sometimes querystring fragments etc.
const m = s.match(/-(\d{6})(?:\/)?(?:[?#].*)?$/); const m = s.match(/-(\d{6})(?:\/)?(?:[?#].*)?$/);
return m ? m[1] : ""; return m ? m[1] : "";
} }
function extractSkuFromWillowBlock(block) {
const b = String(block || "");
// Image filename pattern:
// /products/710296-Zaya-Gran-Reserva-16-Year_160x.png
const m1 = b.match(/\/products\/(\d{6})[-_]/i);
if (m1) return m1[1];
// Generic fallback
const m2 = b.match(/\b(\d{6})[-_][A-Za-z]/);
if (m2) return m2[1];
return "";
}
function canonicalizeWillowUrl(raw) { function canonicalizeWillowUrl(raw) {
try { try {
const u = new URL(String(raw)); const u = new URL(String(raw));
@ -52,14 +65,10 @@ function parseProductsWillowPark(html, ctx, finalUrl) {
const base = `https://${(ctx && ctx.store && ctx.store.host) || "www.willowpark.net"}/`; const base = `https://${(ctx && ctx.store && ctx.store.host) || "www.willowpark.net"}/`;
// Find start offsets of each product tile.
// This ignores <div class="grid-anchor" ...> nodes safely.
const starts = [...s.matchAll(/<div\b[^>]*class=["'][^"']*\bgrid-item\b[^"']*\bgrid-product\b[^"']*["'][^>]*>/gi)] const starts = [...s.matchAll(/<div\b[^>]*class=["'][^"']*\bgrid-item\b[^"']*\bgrid-product\b[^"']*["'][^>]*>/gi)]
.map((m) => m.index) .map(m => m.index)
.filter((i) => typeof i === "number"); .filter(i => typeof i === "number");
// Slice into blocks from each start to the next start.
// Robust to varying nesting/closing div counts.
const blocks = []; const blocks = [];
for (let i = 0; i < starts.length; i++) { for (let i = 0; i < starts.length; i++) {
const a = starts[i]; const a = starts[i];
@ -68,9 +77,6 @@ function parseProductsWillowPark(html, ctx, finalUrl) {
} }
for (const block of blocks) { for (const block of blocks) {
// Do NOT skip sold-out by badge; badge can exist but be display:none.
// Availability filtering should be done via URL query (?filter.v.availability=1).
const href = const href =
block.match(/<a\b[^>]*href=["']([^"']*\/collections\/[^"']*\/products\/[^"']+)["']/i)?.[1] || block.match(/<a\b[^>]*href=["']([^"']*\/collections\/[^"']*\/products\/[^"']+)["']/i)?.[1] ||
block.match(/<a\b[^>]*href=["']([^"']*\/products\/[^"']+)["']/i)?.[1]; block.match(/<a\b[^>]*href=["']([^"']*\/products\/[^"']+)["']/i)?.[1];
@ -91,22 +97,21 @@ function parseProductsWillowPark(html, ctx, finalUrl) {
const price = extractWillowCardPrice(block); const price = extractWillowCardPrice(block);
const img = extractFirstImgUrl(block, base); const img = extractFirstImgUrl(block, base);
// Some pages include data-product-id on the tile; useful but optional.
const pid = block.match(/\bdata-product-id=["'](\d+)["']/i)?.[1] || ""; const pid = block.match(/\bdata-product-id=["'](\d+)["']/i)?.[1] || "";
const sku = extractSkuFromUrlOrHref(href) || extractSkuFromUrlOrHref(url); const sku =
extractSkuFromUrlOrHref(href) ||
extractSkuFromUrlOrHref(url) ||
extractSkuFromWillowBlock(block);
items.push({ name, price, url, sku, img, pid }); items.push({ name, price, url, sku, img, pid });
} }
// De-dupe by canonical URL (same product can appear multiple times).
const uniq = new Map(); const uniq = new Map();
for (const it of items) uniq.set(it.url, it); for (const it of items) uniq.set(it.url, it);
return [...uniq.values()]; return [...uniq.values()];
} }
// Helps discovery + scanning stop when paging past inventory.
function willowIsEmptyListingPage(html) { function willowIsEmptyListingPage(html) {
const s = String(html || ""); const s = String(html || "");
if (/Sorry,\s+there are no products in this collection\./i.test(s)) return true; if (/Sorry,\s+there are no products in this collection\./i.test(s)) return true;