From fb92748784abf604f00188787b5403b9ef28327f Mon Sep 17 00:00:00 2001
From: John Mizerek <mizerek@gmail.com>
Date: Sat, 7 Mar 2026 14:24:22 -0800
Subject: [PATCH] Add WooCommerce fast-path with Playwright modifier extraction

Detects WooCommerce sites from Playwright HTML (woocommerce, wc-add-to-cart,
tm-extra-product-options). Runs woo-modifiers.js which navigates all product
pages, extracts items with categories, and scrapes TMEPO/variation modifiers.
Falls through to Claude if extraction fails.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---
 api/setup/analyzeMenuUrl.cfm | 80 ++++++++++++++++++++++++++++++++++++
 1 file changed, 80 insertions(+)

diff --git a/api/setup/analyzeMenuUrl.cfm b/api/setup/analyzeMenuUrl.cfm
index afc64c2..9ca6d21 100644
--- a/api/setup/analyzeMenuUrl.cfm
+++ b/api/setup/analyzeMenuUrl.cfm
@@ -1059,6 +1059,86 @@
             <cfset playwrightImages = structKeyExists(playwrightResult, "images") ? playwrightResult.images : arrayNew(1)>
             <cfset arrayAppend(response.steps, "Fetched " & len(pageHtml) & " bytes via Playwright, " & arrayLen(playwrightImages) & " images captured")>
 
+            <!--- ========== WOOCOMMERCE FAST PATH ========== --->
+            <cfif findNoCase("woocommerce", pageHtml) OR findNoCase("wc-add-to-cart", pageHtml) OR findNoCase("tm-extra-product-options", pageHtml)>
+                <cfset arrayAppend(response.steps, "WooCommerce site detected - running modifier extraction")>
+
+                <!--- Use the shop root URL --->
+                <cfset wooUrl = reReplace(targetUrl, "(https?://[^/]+).*", "\1")>
+
+                <cftry>
+                    <cfset wooOutput = "">
+                    <cfexecute name="/opt/playwright/run-woo-modifiers.sh" arguments="'#wooUrl#'" timeout="300" variable="wooOutput" />
+
+                    <cfif len(trim(wooOutput))>
+                        <cfset wooResult = deserializeJSON(wooOutput)>
+
+                        <cfif structKeyExists(wooResult, "items") AND isArray(wooResult.items) AND arrayLen(wooResult.items) GT 0>
+                            <cfset arrayAppend(response.steps, "WooCommerce extraction: " & arrayLen(wooResult.items) & " items, " & arrayLen(wooResult.modifiers) & " modifier groups")>
+
+                            <!--- Build categories from item category fields --->
+                            <cfset wooCats = {}>
+                            <cfset wooItems = []>
+                            <cfset wooImageMappings = {}>
+                            <cfloop from="1" to="#arrayLen(wooResult.items)#" index="wi">
+                                <cfset wItem = wooResult.items[wi]>
+                                <cfset catName = structKeyExists(wItem, "category") AND len(trim(wItem.category)) ? trim(wItem.category) : "Menu">
+                                <cfif NOT structKeyExists(wooCats, catName)>
+                                    <cfset wooCats[catName] = 0>
+                                </cfif>
+                                <cfset wooCats[catName] = wooCats[catName] + 1>
+
+                                <cfset itemId = "item_" & wi>
+                                <cfset itemMods = structKeyExists(wooResult, "itemModifierMap") AND structKeyExists(wooResult.itemModifierMap, wItem.name) ? wooResult.itemModifierMap[wItem.name] : []>
+                                <cfset arrayAppend(wooItems, {
+                                    "id": itemId,
+                                    "name": wItem.name,
+                                    "price": structKeyExists(wItem, "price") ? val(wItem.price) : 0,
+                                    "description": structKeyExists(wItem, "description") ? wItem.description : "",
+                                    "category": catName,
+                                    "modifiers": itemMods,
+                                    "hasModifiers": arrayLen(itemMods) GT 0
+                                })>
+
+                                <cfif structKeyExists(wItem, "imageUrl") AND len(trim(wItem.imageUrl))>
+                                    <cfset wooImageMappings[wItem.name] = wItem.imageUrl>
+                                </cfif>
+                            </cfloop>
+
+                            <cfset wooCategories = []>
+                            <cfloop collection="#wooCats#" item="wcName">
+                                <cfset arrayAppend(wooCategories, { "name": wcName, "itemCount": wooCats[wcName] })>
+                            </cfloop>
+
+                            <cfset wooModifiers = structKeyExists(wooResult, "modifiers") ? wooResult.modifiers : []>
+
+                            <cfset menuData = {
+                                "business": { "name": "", "address": "", "phone": "", "hours": "" },
+                                "categories": wooCategories,
+                                "items": wooItems,
+                                "modifiers": wooModifiers,
+                                "imageUrls": [],
+                                "imageMappings": wooImageMappings,
+                                "headerCandidateIndices": []
+                            }>
+
+                            <cfset response["OK"] = true>
+                            <cfset response["DATA"] = menuData>
+                            <cfset response["sourceUrl"] = targetUrl>
+                            <cfset response["parsedVia"] = "woocommerce_playwright">
+                            <cfcontent type="application/json" reset="true">
+                            <cfoutput>#serializeJSON(response)#</cfoutput>
+                            <cfabort>
+                        </cfif>
+                    </cfif>
+                    <cfset arrayAppend(response.steps, "WooCommerce extraction returned no items - falling through to Claude")>
+                <cfcatch>
+                    <cfset arrayAppend(response.steps, "WooCommerce extraction failed: " & cfcatch.message & " - falling through to Claude")>
+                </cfcatch>
+                </cftry>
+            </cfif>
+            <!--- ========== END WOOCOMMERCE FAST PATH ========== --->
+
             <!--- Extract base URL for resolving relative links --->
             <cfset baseUrl = reReplace(targetUrl, "(https?://[^/]+).*", "\1")>
             <cfset basePath = reReplace(targetUrl, "(https?://[^/]+/[^?]*/?).*", "\1")>