semantic-action-grammar 1.0.1

Semantic Action Grammar — a DSL for structured inter-agent communication, with parsing, validation, correlation, and fold-based context compression.

semantic-action-grammar

Semantic Action Grammar (SAG) — a DSL for structured inter-agent communication.

SAG provides parsing, schema validation, correlation, sanitization, and a grammar-level fold/unfold protocol for context compression. This package is the Python implementation; the canonical grammar lives in the main repository.

Install

pip install semantic-action-grammar

The Python import name is sag:

from sag import SAGMessageParser

Quick start

from sag import SAGMessageParser, MessageMinifier

parser = SAGMessageParser()
message = parser.parse('[id=1 src=A dst=B] DO deploy("app1", version=42) BECAUSE balance>1000;')

print(message.header.id)          # "1"
print(message.statements[0].verb) # "deploy"

# Re-serialize to the compact wire format
print(MessageMinifier().minify(message))

What's in the box

• Parser — SAGMessageParser, ANTLR4-backed
• Schema validation — SchemaRegistry, VerbSchema, SchemaValidator
• Guardrails — BECAUSE-clause expressions evaluated against a Context
• Sanitizer — four-layer firewall (parse → routing → schema → guardrail)
• Fold protocol — FoldEngine for lossless conversation compression
• Knowledge engine — versioned topic-based fact propagation
• Tree / Grove — multi-agent topology and bottom-up orchestration
• Accounting — token and cost tracking for grove executions

See the main repository for benchmarks, the chatbot demo, and the full specification.

License

MIT — see LICENSE.