pgames-manager/backend/app/converter/pyexpr.py

"""Translate the subset of Python used by visual novels into PGVN IR.

Expressions become JSON trees ``{"k": kind, ...}``; statements become
``{"op": ..., ...}`` nodes executed by the player's JS interpreter.
Anything outside the subset raises :class:`Unsupported` so the caller can
report it and degrade gracefully.
"""

import ast
from typing import Any

Expr = dict[str, Any]
Stmt = dict[str, Any]

BIN_OPS = {
    ast.Add: "+", ast.Sub: "-", ast.Mult: "*", ast.Div: "/", ast.FloorDiv: "//",
    ast.Mod: "%", ast.Pow: "**", ast.BitAnd: "&", ast.BitOr: "|", ast.BitXor: "^",
    ast.LShift: "<<", ast.RShift: ">>",
}
CMP_OPS = {
    ast.Eq: "==", ast.NotEq: "!=", ast.Lt: "<", ast.LtE: "<=", ast.Gt: ">", ast.GtE: ">=",
    ast.In: "in", ast.NotIn: "not in", ast.Is: "is", ast.IsNot: "is not",
}
UNARY_OPS = {ast.Not: "not", ast.USub: "-", ast.UAdd: "+", ast.Invert: "~"}

# Modules whose import is harmless: the player provides the functions it supports.
KNOWN_MODULES = {"random", "math", "renpy", "store", "time", "datetime"}


class Unsupported(Exception):
    pass


def const(v: Any) -> Expr:
    return {"k": "const", "v": v}


def parse_expr(source: str) -> Expr:
    try:
        tree = ast.parse(source.strip(), mode="eval")
    except SyntaxError as e:
        raise Unsupported(f"синтаксическая ошибка в выражении: {source!r}") from e
    return expr(tree.body)


def parse_stmts(source: str) -> list[Stmt]:
    try:
        tree = ast.parse(_dedent(source))
    except SyntaxError as e:
        raise Unsupported(f"синтаксическая ошибка в python: {e.msg} ({source[:60]!r})") from e
    return stmts(tree.body)


def _dedent(source: str) -> str:
    import textwrap
    return textwrap.dedent(source)


def literal(source: str) -> Any:
    """Evaluate a constant expression at conversion time, or raise Unsupported."""
    try:
        return ast.literal_eval(source.strip())
    except (ValueError, SyntaxError, TypeError, MemoryError, RecursionError) as e:
        raise Unsupported(f"не константа: {source!r}") from e


def expr(node: ast.AST) -> Expr:
    match node:
        case ast.Constant(value=v):
            if isinstance(v, bytes):
                v = v.decode("utf-8", "replace")
            if isinstance(v, complex) or v is Ellipsis:
                raise Unsupported("неподдерживаемая константа")
            return const(v)
        case ast.Name(id=name):
            return {"k": "name", "id": name}
        case ast.Attribute(value=ast.Name(id="store"), attr=attr):
            return {"k": "name", "id": attr}
        case ast.Attribute(value=value, attr=attr):
            return {"k": "attr", "o": expr(value), "a": attr}
        case ast.Subscript(value=value, slice=ast.Slice(lower=lo, upper=hi, step=st)):
            return {"k": "slice", "o": expr(value), "lo": _opt(lo), "hi": _opt(hi), "st": _opt(st)}
        case ast.Subscript(value=value, slice=index):
            return {"k": "sub", "o": expr(value), "i": expr(index)}
        case ast.BinOp(left=l, op=op, right=r):
            if type(op) not in BIN_OPS:
                raise Unsupported(f"оператор {type(op).__name__}")
            return {"k": "bin", "op": BIN_OPS[type(op)], "l": expr(l), "r": expr(r)}
        case ast.BoolOp(op=op, values=values):
            return {"k": "bool", "op": "and" if isinstance(op, ast.And) else "or",
                    "v": [expr(v) for v in values]}
        case ast.UnaryOp(op=op, operand=operand):
            return {"k": "unary", "op": UNARY_OPS[type(op)], "v": expr(operand)}
        case ast.Compare(left=l, ops=ops, comparators=rs):
            return {"k": "cmp", "l": expr(l), "ops": [CMP_OPS[type(o)] for o in ops],
                    "r": [expr(r) for r in rs]}
        case ast.IfExp(test=t, body=b, orelse=e):
            return {"k": "if", "c": expr(t), "t": expr(b), "e": expr(e)}
        case ast.List(elts=elts) | ast.Set(elts=elts):
            return {"k": "list", "v": [expr(e) for e in elts]}
        case ast.Tuple(elts=elts):
            return {"k": "tuple", "v": [expr(e) for e in elts]}
        case ast.Dict(keys=keys, values=values):
            if any(k is None for k in keys):
                raise Unsupported("распаковка ** в словаре")
            return {"k": "dict", "v": [[expr(k), expr(v)] for k, v in zip(keys, values)]}
        case ast.Call(func=func, args=args, keywords=kws):
            if any(isinstance(a, ast.Starred) for a in args) or any(k.arg is None for k in kws):
                raise Unsupported("распаковка аргументов *args/**kwargs")
            return {"k": "call", "f": expr(func), "a": [expr(a) for a in args],
                    "kw": {k.arg: expr(k.value) for k in kws}}
        case ast.JoinedStr(values=values):
            parts: list[Any] = []
            for v in values:
                if isinstance(v, ast.Constant):
                    parts.append(v.value)
                elif isinstance(v, ast.FormattedValue):
                    parts.append(expr(v.value))
            return {"k": "fmt", "v": parts}
        case ast.ListComp(elt=elt, generators=[ast.comprehension(target=t, iter=it, ifs=ifs,
                                                                     is_async=0)]):
            return {"k": "comp", "e": expr(elt), "t": target(t), "it": expr(it),
                    "if": [expr(i) for i in ifs]}
        case ast.Lambda():
            raise Unsupported("lambda")
    raise Unsupported(f"выражение {type(node).__name__}")


def _opt(node: ast.AST | None) -> Expr | None:
    return None if node is None else expr(node)


def target(node: ast.AST) -> Expr:
    match node:
        case ast.Name() | ast.Attribute() | ast.Subscript():
            return expr(node)
        case ast.Tuple(elts=elts) | ast.List(elts=elts):
            return {"k": "tuple", "v": [target(e) for e in elts]}
    raise Unsupported(f"присваивание в {type(node).__name__}")


def stmts(nodes: list[ast.stmt]) -> list[Stmt]:
    out: list[Stmt] = []
    for n in nodes:
        out.extend(stmt(n))
    return out


def stmt(node: ast.stmt) -> list[Stmt]:
    match node:
        case ast.Assign(targets=targets, value=value):
            v = expr(value)
            return [{"op": "set", "t": target(t), "v": v} for t in targets]
        case ast.AnnAssign(target=t, value=value) if value is not None:
            return [{"op": "set", "t": target(t), "v": expr(value)}]
        case ast.AugAssign(target=t, op=op, value=value):
            if type(op) not in BIN_OPS:
                raise Unsupported(f"оператор {type(op).__name__}")
            return [{"op": "aug", "t": target(t), "o": BIN_OPS[type(op)], "v": expr(value)}]
        case ast.Expr(value=ast.Constant()):
            return []  # docstring / bare literal
        case ast.Expr(value=value):
            return [{"op": "eval", "e": expr(value)}]
        case ast.If(test=test, body=body, orelse=orelse):
            return [{"op": "if", "c": expr(test), "then": stmts(body), "else": stmts(orelse)}]
        case ast.While(test=test, body=body, orelse=[]):
            return [{"op": "while", "c": expr(test), "body": stmts(body)}]
        case ast.For(target=t, iter=it, body=body, orelse=[]):
            return [{"op": "for", "t": target(t), "it": expr(it), "body": stmts(body)}]
        case ast.Delete(targets=targets):
            return [{"op": "del", "t": target(t)} for t in targets]
        case ast.Pass() | ast.Global() | ast.Nonlocal():
            return []
        case ast.Break():
            return [{"op": "break"}]
        case ast.Continue():
            return [{"op": "continue"}]
        case ast.Import(names=names) | ast.ImportFrom(names=names):
            mod = getattr(node, "module", None) or names[0].name
            if mod.split(".")[0] in KNOWN_MODULES:
                return []
            raise Unsupported(f"импорт модуля {mod}")
        case ast.FunctionDef(name=name) | ast.ClassDef(name=name):
            raise Unsupported(f"определение {name}()")
    raise Unsupported(f"оператор {type(node).__name__}")