Remove Uber Eats JSON-LD fast path — let Claude extract modifiers
The JSON-LD fast path only got items/categories/prices but no modifiers. Removing it lets Uber Eats pages fall through to Claude AI extraction which handles modifiers like every other platform. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
John Mizerek
parent
6cb0eefb5f
commit
bace74aa23
2 changed files with 2 additions and 161 deletions
|
|
@ -2190,166 +2190,7 @@
|
||||||
</cfif>
|
</cfif>
|
||||||
|
|
||||||
<!--- ============================================================ --->
|
<!--- ============================================================ --->
|
||||||
<!--- UBER EATS FAST PATH: Parse JSON-LD structured menu data --->
|
<!--- Uber Eats: No fast path — JSON-LD lacks modifiers. Falls through to Claude for full extraction. --->
|
||||||
<!--- ============================================================ --->
|
|
||||||
<cfif findNoCase("ubereats.com", pageHtml) OR findNoCase("uber.com/store", pageHtml)>
|
|
||||||
<cfset arrayAppend(response.steps, "Uber Eats page detected - looking for JSON-LD menu data")>
|
|
||||||
<cftry>
|
|
||||||
<!--- Extract all JSON-LD blocks --->
|
|
||||||
<cfset jsonLdBlocks = reMatchNoCase('<script[^>]*type\s*=\s*["'']application/ld\+json["''][^>]*>([\s\S]*?)</script>', pageHtml)>
|
|
||||||
<cfset arrayAppend(response.steps, "Found " & arrayLen(jsonLdBlocks) & " JSON-LD blocks")>
|
|
||||||
|
|
||||||
<cfset ueRestaurant = "">
|
|
||||||
<cfloop array="#jsonLdBlocks#" index="ldBlock">
|
|
||||||
<!--- Extract content between script tags --->
|
|
||||||
<cfset ldContent = reReplaceNoCase(ldBlock, '<script[^>]*>([\s\S]*?)</script>', '\1')>
|
|
||||||
<cfset ldContent = trim(ldContent)>
|
|
||||||
<cfif len(ldContent)>
|
|
||||||
<cftry>
|
|
||||||
<!--- Unescape unicode \u002F etc --->
|
|
||||||
<cfset ldContent = replace(ldContent, '\u002F', '/', 'all')>
|
|
||||||
<cfset ldContent = replace(ldContent, '\u0026', '&', 'all')>
|
|
||||||
<cfset ldContent = replace(ldContent, '\u0022', '"', 'all')>
|
|
||||||
<cfset ldContent = replace(ldContent, '\u0027', "'", 'all')>
|
|
||||||
<cfset ldParsed = deserializeJSON(ldContent)>
|
|
||||||
<cfif isStruct(ldParsed) AND structKeyExists(ldParsed, "@type") AND ldParsed["@type"] EQ "Restaurant" AND structKeyExists(ldParsed, "hasMenu")>
|
|
||||||
<cfset ueRestaurant = ldParsed>
|
|
||||||
</cfif>
|
|
||||||
<cfcatch><!--- skip unparseable blocks ---></cfcatch>
|
|
||||||
</cftry>
|
|
||||||
</cfif>
|
|
||||||
</cfloop>
|
|
||||||
|
|
||||||
<cfif isStruct(ueRestaurant) AND structKeyExists(ueRestaurant, "hasMenu")>
|
|
||||||
<cfset arrayAppend(response.steps, "Found Restaurant JSON-LD with menu data")>
|
|
||||||
|
|
||||||
<!--- Parse business info --->
|
|
||||||
<cfset ueBusiness = structNew()>
|
|
||||||
<cfset ueBusiness["name"] = structKeyExists(ueRestaurant, "name") ? ueRestaurant.name : "">
|
|
||||||
<!--- Unescape HTML entities in name --->
|
|
||||||
<cfset ueBusiness["name"] = replace(ueBusiness.name, "&", "&", "all")>
|
|
||||||
<cfset ueBusiness["name"] = replace(ueBusiness.name, "<", "<", "all")>
|
|
||||||
<cfset ueBusiness["name"] = replace(ueBusiness.name, ">", ">", "all")>
|
|
||||||
<cfset ueBusiness["name"] = replace(ueBusiness.name, "&##39;", "'", "all")>
|
|
||||||
<cfset ueBusiness["name"] = replace(ueBusiness.name, "'", "'", "all")>
|
|
||||||
|
|
||||||
<cfif structKeyExists(ueRestaurant, "address") AND isStruct(ueRestaurant.address)>
|
|
||||||
<cfset ueAddr = ueRestaurant.address>
|
|
||||||
<cfif structKeyExists(ueAddr, "streetAddress")><cfset ueBusiness["addressLine1"] = ueAddr.streetAddress></cfif>
|
|
||||||
<cfif structKeyExists(ueAddr, "addressLocality")><cfset ueBusiness["city"] = ueAddr.addressLocality></cfif>
|
|
||||||
<cfif structKeyExists(ueAddr, "addressRegion")><cfset ueBusiness["state"] = ueAddr.addressRegion></cfif>
|
|
||||||
<cfif structKeyExists(ueAddr, "postalCode")><cfset ueBusiness["zip"] = ueAddr.postalCode></cfif>
|
|
||||||
</cfif>
|
|
||||||
|
|
||||||
<!--- Parse menu sections --->
|
|
||||||
<cfset ueMenu = ueRestaurant.hasMenu>
|
|
||||||
<cfset ueCategories = []>
|
|
||||||
<cfset ueItems = []>
|
|
||||||
<cfset ueImageMappings = structNew()>
|
|
||||||
<cfset ueItemId = 1>
|
|
||||||
|
|
||||||
<cfif structKeyExists(ueMenu, "hasMenuSection") AND isArray(ueMenu.hasMenuSection)>
|
|
||||||
<cfloop array="#ueMenu.hasMenuSection#" index="ueSection">
|
|
||||||
<cfset ueCatName = structKeyExists(ueSection, "name") ? trim(ueSection.name) : "Menu">
|
|
||||||
<!--- Unescape HTML entities --->
|
|
||||||
<cfset ueCatName = replace(ueCatName, "&", "&", "all")>
|
|
||||||
<cfset ueCatName = replace(ueCatName, "&##39;", "'", "all")>
|
|
||||||
<cfset ueCatName = replace(ueCatName, "'", "'", "all")>
|
|
||||||
<cfset ueSectionItemCount = 0>
|
|
||||||
|
|
||||||
<cfif structKeyExists(ueSection, "hasMenuItem") AND isArray(ueSection.hasMenuItem)>
|
|
||||||
<cfloop array="#ueSection.hasMenuItem#" index="ueMenuItem">
|
|
||||||
<cfset ueItemName = structKeyExists(ueMenuItem, "name") ? trim(ueMenuItem.name) : "">
|
|
||||||
<cfif NOT len(ueItemName)><cfcontinue></cfif>
|
|
||||||
|
|
||||||
<!--- Unescape HTML entities in name and description --->
|
|
||||||
<cfset ueItemName = replace(ueItemName, "&", "&", "all")>
|
|
||||||
<cfset ueItemName = replace(ueItemName, "&##39;", "'", "all")>
|
|
||||||
<cfset ueItemName = replace(ueItemName, "'", "'", "all")>
|
|
||||||
<cfset ueItemName = replace(ueItemName, "<", "<", "all")>
|
|
||||||
<cfset ueItemName = replace(ueItemName, ">", ">", "all")>
|
|
||||||
|
|
||||||
<cfset ueItemDesc = structKeyExists(ueMenuItem, "description") ? trim(ueMenuItem.description) : "">
|
|
||||||
<cfset ueItemDesc = replace(ueItemDesc, "&", "&", "all")>
|
|
||||||
<cfset ueItemDesc = replace(ueItemDesc, "&##39;", "'", "all")>
|
|
||||||
<cfset ueItemDesc = replace(ueItemDesc, "'", "'", "all")>
|
|
||||||
|
|
||||||
<!--- Extract price from offers --->
|
|
||||||
<cfset uePrice = 0>
|
|
||||||
<cfif structKeyExists(ueMenuItem, "offers") AND isStruct(ueMenuItem.offers)>
|
|
||||||
<cfif structKeyExists(ueMenuItem.offers, "price")>
|
|
||||||
<cfset uePrice = val(ueMenuItem.offers.price)>
|
|
||||||
</cfif>
|
|
||||||
</cfif>
|
|
||||||
|
|
||||||
<!--- Extract image if available --->
|
|
||||||
<cfset ueItemImage = "">
|
|
||||||
<cfif structKeyExists(ueMenuItem, "image") AND len(trim(ueMenuItem.image))>
|
|
||||||
<cfset ueItemImage = trim(ueMenuItem.image)>
|
|
||||||
<cfset ueImageMappings[ueItemName] = ueItemImage>
|
|
||||||
</cfif>
|
|
||||||
|
|
||||||
<cfset arrayAppend(ueItems, {
|
|
||||||
"id": "item_" & ueItemId,
|
|
||||||
"name": ueItemName,
|
|
||||||
"price": uePrice,
|
|
||||||
"description": ueItemDesc,
|
|
||||||
"category": ueCatName,
|
|
||||||
"modifiers": [],
|
|
||||||
"imageUrl": ueItemImage
|
|
||||||
})>
|
|
||||||
<cfset ueItemId = ueItemId + 1>
|
|
||||||
<cfset ueSectionItemCount = ueSectionItemCount + 1>
|
|
||||||
</cfloop>
|
|
||||||
</cfif>
|
|
||||||
|
|
||||||
<cfif ueSectionItemCount GT 0>
|
|
||||||
<cfset arrayAppend(ueCategories, { "name": ueCatName, "itemCount": ueSectionItemCount })>
|
|
||||||
</cfif>
|
|
||||||
</cfloop>
|
|
||||||
</cfif>
|
|
||||||
|
|
||||||
<cfset arrayAppend(response.steps, "Parsed " & arrayLen(ueItems) & " items in " & arrayLen(ueCategories) & " categories from Uber Eats JSON-LD")>
|
|
||||||
|
|
||||||
<cfif arrayLen(ueItems) GT 0>
|
|
||||||
<!--- Try to get restaurant header image from JSON-LD images --->
|
|
||||||
<cfset ueHeaderImage = "">
|
|
||||||
<cfif structKeyExists(ueRestaurant, "image") AND isArray(ueRestaurant.image) AND arrayLen(ueRestaurant.image) GT 0>
|
|
||||||
<cfset ueHeaderImage = ueRestaurant.image[1]>
|
|
||||||
</cfif>
|
|
||||||
|
|
||||||
<cfset menuData = {
|
|
||||||
"business": ueBusiness,
|
|
||||||
"categories": ueCategories,
|
|
||||||
"items": ueItems,
|
|
||||||
"modifiers": [],
|
|
||||||
"imageUrls": [],
|
|
||||||
"imageMappings": ueImageMappings,
|
|
||||||
"headerCandidateIndices": [],
|
|
||||||
"headerImage": ueHeaderImage
|
|
||||||
}>
|
|
||||||
|
|
||||||
<cfset response["OK"] = true>
|
|
||||||
<cfset response["DATA"] = menuData>
|
|
||||||
<cfset response["sourceUrl"] = len(targetUrl) ? targetUrl : "ubereats-upload">
|
|
||||||
<cfset response["parsedVia"] = "ubereats_jsonld">
|
|
||||||
<cfset response["pagesProcessed"] = 1>
|
|
||||||
<cfset response["imagesFound"] = structCount(ueImageMappings)>
|
|
||||||
<cfcontent type="application/json" reset="true">
|
|
||||||
<cfoutput>#serializeJSON(response)#</cfoutput>
|
|
||||||
<cfabort>
|
|
||||||
</cfif>
|
|
||||||
<cfelse>
|
|
||||||
<cfset arrayAppend(response.steps, "No Restaurant JSON-LD with menu found - falling through")>
|
|
||||||
</cfif>
|
|
||||||
<cfcatch>
|
|
||||||
<cfset ueError = "Uber Eats JSON-LD parsing failed: " & cfcatch.message>
|
|
||||||
<cfif len(cfcatch.detail)><cfset ueError = ueError & " | Detail: " & cfcatch.detail></cfif>
|
|
||||||
<cfset arrayAppend(response.steps, ueError & " - falling back to Claude")>
|
|
||||||
</cfcatch>
|
|
||||||
</cftry>
|
|
||||||
</cfif>
|
|
||||||
<!--- ========== END UBER EATS FAST PATH ========== --->
|
|
||||||
|
|
||||||
<!--- Look for embedded JSON data (Next.js __NEXT_DATA__, Toast state, etc.) --->
|
<!--- Look for embedded JSON data (Next.js __NEXT_DATA__, Toast state, etc.) --->
|
||||||
<cfset embeddedJsonData = "">
|
<cfset embeddedJsonData = "">
|
||||||
|
|
|
||||||
|
|
@ -856,7 +856,7 @@
|
||||||
<p style="margin:0 0 2px;"><strong style="color:var(--gray-500);">Toast</strong> — full menu with modifiers & images (URL or saved page)</p>
|
<p style="margin:0 0 2px;"><strong style="color:var(--gray-500);">Toast</strong> — full menu with modifiers & images (URL or saved page)</p>
|
||||||
<p style="margin:0 0 2px;"><strong style="color:var(--gray-500);">Grubhub</strong> — full menu with modifiers (URL only)</p>
|
<p style="margin:0 0 2px;"><strong style="color:var(--gray-500);">Grubhub</strong> — full menu with modifiers (URL only)</p>
|
||||||
<p style="margin:0 0 2px;"><strong style="color:var(--gray-500);">DoorDash / order.online</strong> — items, categories, prices (URL or saved page)</p>
|
<p style="margin:0 0 2px;"><strong style="color:var(--gray-500);">DoorDash / order.online</strong> — items, categories, prices (URL or saved page)</p>
|
||||||
<p style="margin:0 0 2px;"><strong style="color:var(--gray-500);">Uber Eats</strong> — items, categories, prices (URL or saved page)</p>
|
<p style="margin:0 0 2px;"><strong style="color:var(--gray-500);">Uber Eats</strong> — full menu with modifiers (saved page)</p>
|
||||||
<p style="margin:0;"><strong style="color:var(--gray-500);">Other sites</strong> — upload saved page, images, or PDFs for AI extraction</p>
|
<p style="margin:0;"><strong style="color:var(--gray-500);">Other sites</strong> — upload saved page, images, or PDFs for AI extraction</p>
|
||||||
</div>
|
</div>
|
||||||
|
|
||||||
|
|
|
||||||
Reference in a new issue