Shopify JSON-LD Parser

Parse JSON-LD from Shopify product pages (or any schema.org/Product JSON-LD) to extract structured commerce data reliably.

Quick Start

1. Fetch HTML: web_fetch a product page URL.
2. Parse: Run scripts/parse_shopify_jsonld.py input.html (outputs JSON).
3. Query output for offers, availability, etc.

Example:

exec command="python3 skills/shopify-jsonld-parser/scripts/parse_shopify_jsonld.py page.html"

Workflow

1. Input: HTML file/path with JSON-LD (from web_fetch, read, or local).
2. Find JSON-LD: Script extracts all <script type="application/ld+json"> contents.
3. Parse Product: Traverse @graph or root for @type: Product, extract:
   • name, description, image, sku, brand
   • offers: array of Offer objects → price, priceCurrency, availability, url, itemCondition
   • availability (product-level or per-offer): http://schema.org/InStock, OutOfStock, PreOrder, etc.
   • Variants: Infer from multiple offers or hasVariant.
4. Output: Clean JSON: {product: {...}, offers: [...], inventory_status: "InStock|OutOfStock|Limited|Unknown"}

Inventory is typically availability enum, not exact quantity (use Shopify API for levels). Map:

• InStock → available
• OutOfStock → unavailable
• PreOrder/BackOrder → preorder

Usage Examples

Extract offers from page:

read path="shopify-page.html" → html
write path="tmp.html" content=html
exec command="python3 skills/shopify-jsonld-parser/scripts/parse_shopify_jsonld.py tmp.html" → parsed.json
read path="parsed.json"

Query CLI-style: Add --query price or --field offers[0].availability flags to script.

Script Reference

scripts/parse_shopify_jsonld.py [input.html] [--query FIELD] [--output json|yaml]

• Handles malformed JSON-LD gracefully.
• Supports multiple @graph items.
• Validates schema.org types.

Schema Reference

See references/schema.md for schema.org/Product + Offer properties.

Troubleshooting

• No JSON-LD: Returns {error: "No product JSON-LD found"}
• Multiple products: Takes first @type:Product.
• Dynamic Liquid: Static HTML only (no JS-rendered).