rose-ash/shared/sx/ref/boundary_parser.py

"""
Parse boundary.sx and primitives.sx to extract declared names.

Shared by both bootstrap_py.py and bootstrap_js.py, and used at runtime
by the validation module.
"""
from __future__ import annotations

import os
from typing import Any

# Allow standalone use (from bootstrappers) or in-project imports
try:
    from shared.sx.parser import parse_all
    from shared.sx.types import Symbol, Keyword, NIL as SX_NIL
except ImportError:
    import sys
    _HERE = os.path.dirname(os.path.abspath(__file__))
    _PROJECT = os.path.abspath(os.path.join(_HERE, "..", "..", ".."))
    sys.path.insert(0, _PROJECT)
    from shared.sx.parser import parse_all
    from shared.sx.types import Symbol, Keyword, NIL as SX_NIL


def _ref_dir() -> str:
    return os.path.dirname(os.path.abspath(__file__))


def _read_file(filename: str) -> str:
    filepath = os.path.join(_ref_dir(), filename)
    with open(filepath, encoding="utf-8") as f:
        return f.read()


def _extract_keyword_arg(expr: list, key: str) -> Any:
    """Extract :key value from a flat keyword-arg list."""
    for i, item in enumerate(expr):
        if isinstance(item, Keyword) and item.name == key and i + 1 < len(expr):
            return expr[i + 1]
    return None


def parse_primitives_sx() -> frozenset[str]:
    """Parse primitives.sx and return frozenset of declared pure primitive names."""
    by_module = parse_primitives_by_module()
    all_names: set[str] = set()
    for names in by_module.values():
        all_names.update(names)
    return frozenset(all_names)


def parse_primitives_by_module() -> dict[str, frozenset[str]]:
    """Parse primitives.sx and return primitives grouped by module.

    Returns:
        Dict mapping module name (e.g. "core.arithmetic") to frozenset of
        primitive names declared under that module.
    """
    source = _read_file("primitives.sx")
    exprs = parse_all(source)
    modules: dict[str, set[str]] = {}
    current_module = "_unscoped"

    for expr in exprs:
        if not isinstance(expr, list) or len(expr) < 2:
            continue
        if not isinstance(expr[0], Symbol):
            continue

        if expr[0].name == "define-module":
            mod_name = expr[1]
            if isinstance(mod_name, Keyword):
                current_module = mod_name.name
            elif isinstance(mod_name, str):
                current_module = mod_name

        elif expr[0].name == "define-primitive":
            name = expr[1]
            if isinstance(name, str):
                modules.setdefault(current_module, set()).add(name)

    return {mod: frozenset(names) for mod, names in modules.items()}


def parse_boundary_sx() -> tuple[frozenset[str], dict[str, frozenset[str]]]:
    """Parse boundary.sx and return (io_names, {service: helper_names}).

    Returns:
        io_names: frozenset of declared I/O primitive names
        helpers: dict mapping service name to frozenset of helper names
    """
    source = _read_file("boundary.sx")
    exprs = parse_all(source)
    io_names: set[str] = set()
    helpers: dict[str, set[str]] = {}

    for expr in exprs:
        if not isinstance(expr, list) or not expr:
            continue
        head = expr[0]
        if not isinstance(head, Symbol):
            continue

        if head.name == "define-io-primitive":
            name = expr[1]
            if isinstance(name, str):
                io_names.add(name)

        elif head.name == "define-page-helper":
            name = expr[1]
            service = _extract_keyword_arg(expr, "service")
            if isinstance(name, str) and isinstance(service, str):
                helpers.setdefault(service, set()).add(name)

    frozen_helpers = {svc: frozenset(names) for svc, names in helpers.items()}
    return frozenset(io_names), frozen_helpers


def parse_boundary_types() -> frozenset[str]:
    """Parse boundary.sx and return the declared boundary type names."""
    source = _read_file("boundary.sx")
    exprs = parse_all(source)
    for expr in exprs:
        if (isinstance(expr, list) and len(expr) >= 2
                and isinstance(expr[0], Symbol)
                and expr[0].name == "define-boundary-types"):
            type_list = expr[1]
            if isinstance(type_list, list):
                # (list "number" "string" ...)
                return frozenset(
                    item for item in type_list
                    if isinstance(item, str)
                )
    return frozenset()