PSSParser

This project provides a ANTLR4-based parser for the Accellera PSS language. It also provides an AST (data model) for processing the result of the parser.

Supported standard

pssparser supports PSS 3.1 (the Public Review Draft of 2026-08-28). PSS 3.0 syntax that 3.1 removed is rejected; see docs/pss31_migration.rst.

The pssparser version is not the LRM revision. Major releases are for large changes, including support for a new LRM revision; minor releases for medium changes; patch releases for small changes and fixes. To check from a program, read pssparser.SUPPORTED_LRM; pssparser --version prints it on its second line.

Agent skills

The wheel ships agent skills that teach a coding agent to use pssparser: pssparser (check and lint PSS), pssparser-checkers (write lint rules: a rule file first, a Python checker when a rule file cannot express the check) and pssparser-api (build tools on the Python API). They are registered in the agent.skills entry-point group, so ivpm update installs them into .claude/skills/ and the other agent directories. The sources are in skills/, with pssparser-dev, a fourth skill for work on pssparser itself that is not shipped.

Checker Plug-ins

pssparser supports a plug-in system for custom Python-based checkers that run after a successful parse and link. Plug-ins can be contributed by any installed Python package via the pssparser.checkers entry_points group, or loaded on the fly with --load-checker MODULE:CLASS.

# List all registered checkers
pssparser --list-checkers

# List all marker IDs (built-in + plug-in)
pssparser --list-markers

# Describe a specific marker
pssparser --describe PSS001

# Run only a specific checker
pssparser --checker naming-convention model.pss

# Load and run a local checker (no install needed)
pssparser --load-checker myproject.rules:StyleChecker model.pss

See docs/checker_plugin_guide.rst for a full guide to writing and registering checker plug-ins.