You are generating a web-poet page object from an extraction spec. The spec contains a schema, saved HTML pages, and expected values. It may describe any data type — product details, navigation links, article content, etc. Codegen doesn't need to know the data type; it generates a PO that extracts according to the schema.

The spec was produced by /scrape-spec and the project by /scrape-ensure-project.

Read python-environments.md and docs-access.md from ${CLAUDE_SKILL_DIR}/../scrape/references.

Input

The raw argument string is $ARGUMENTS. Split it into up to 3 whitespace-separated positional arguments:

spec_path: path to spec folder, e.g. .scrape/books-toscrape
project_dir: path to the Scrapy project
fields: optional, comma-separated field names to generate (empty = all fields)

scrape-codegen

Input

Process