lacelang-validator (python)

Canonical Python validator for Lace — the reference implementation with 100% spec conformance.

This is the validator that the Lace specification is developed and tested against. All error codes, AST schemas, and validation rules are verified against this implementation before each spec release.

This package only parses and validates .lace source text. It has no HTTP client and does not execute probes. The runtime and its network surface live in the separate lacelang-executor package, per lace-spec.md §16.

Why split

Integrators that want syntax / semantic checking in a CI job, IDE extension, or block editor shouldn't have to pull in an HTTP stack, TLS, cookie state, or an extension dispatcher. Installing lacelang-validator alone gives a clean audit surface with zero network exposure.

Install

pip install lacelang-validator

Or from source:

pip install git+https://github.com/tracedown/lacelang-python-validator.git

Usage

# Parse → emit AST, or { errors: [...] } on syntax failure
lacelang-validate parse script.lace

# Validate → emit { errors: [...], warnings: [...] }
lacelang-validate validate script.lace \
    --vars-list vars.json \
    --context context.json

# With an extension's field/function registrations active
lacelang-validate validate script.lace \
    --enable-extension laceNotifications

# Pretty-print any subcommand
lacelang-validate parse script.lace --pretty

Exit code is always 0 on successful processing. Parse and validation errors are reported in the JSON body. Non-zero exit codes are reserved for tool failures (unreadable source file, malformed arguments).

Library usage

from lacelang_validator.parser import parse
from lacelang_validator.validator import validate

ast = parse(open("script.lace").read())
diagnostics = validate(
    ast,
    variables=["url"],
    context={"maxRedirects": 10, "maxTimeoutMs": 300_000},
)
for err in diagnostics.errors:
    print(err.code, err.call_index, err.field)

License

Apache License 2.0