Each guide explains one part of a reader spec with an input and the CLI’s actual output. Commands run from the repository root with iris on your PATH; see building the CLI.

• Patterns: match headers, capture typed fields and refer to a value already read.
• Sections: read keyword lines in different orders and require or allow headers and reject repeated ones.
• Records and checks: store repeated fields as columns and check each record or a whole collection.
• Several instances: list a catalog and select by ordinal or name, including files with rejected instances.
• Parameters: supply caller values, choose defaults and handle conversion failures.
• Errors: read the location, caret and execution context, and find an error code’s definition.
• Performance: understand input progress and work budgets.

For a walkthrough that builds up a reader spec, start with the tutorial. For exact rules, use the Specification.