Files
TransPyC/lib/core/SymbolExtractor.py
2026-07-18 19:25:40 +08:00

308 lines
15 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""Pass 1: ASTSymbolExtractor — 从 AST 提取原始符号信息"""
from __future__ import annotations
import ast
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from lib.core.SymbolTable import SymbolTable
from lib.core.SymbolData import (
ModuleSymbols, ClassSymbolData, TypedefSymbolData,
FuncSymbolData, DefineSymbolData, AnonymousTypeData
)
from lib.core.Handles.HandlesBase import CTypeInfo
from lib.core.SymbolUtils import AnnotationContainsTType, CheckAnnotationHasCInline, IsTModule, AnnotationContainsName
from lib.includes import t
class ASTSymbolExtractor:
"""从 Python 源文件的 AST 中提取所有符号信息"""
def __init__(self, symbol_table: SymbolTable) -> None:
self._symtab: SymbolTable = symbol_table
def extract(self, file_path: str) -> ModuleSymbols:
"""提取文件中的所有符号信息"""
with open(file_path, 'r', encoding='utf-8') as f:
code: str = f.read()
tree: ast.Module = ast.parse(code)
result: ModuleSymbols = ModuleSymbols(file_path=file_path)
# 第一步:收集所有类的成员和匿名类型
class_members: dict[str, dict[str, CTypeInfo]] = {}
for node in tree.body:
if isinstance(node, ast.ClassDef):
members: dict[str, CTypeInfo] = {}
self._collect_members(node, members, result.anonymous_types)
class_members[node.name] = members
# 第二步:从每个顶层节点提取符号信息
for node in tree.body:
if isinstance(node, ast.ClassDef):
members_info: dict[str, CTypeInfo] = class_members.get(node.name, {})
result.classes.append(self._extract_class(node, members_info))
elif isinstance(node, ast.AnnAssign) and isinstance(node.target, ast.Name):
self._extract_ann_assign(node, result)
elif isinstance(node, ast.FunctionDef):
result.functions.append(self._extract_function(node))
return result
# ------------------------------------------------------------------
# 成员收集(递归处理嵌套匿名类)
# ------------------------------------------------------------------
def _collect_members(self, class_node: ast.ClassDef, members: dict[str, CTypeInfo], anonymous_types: dict[str, AnonymousTypeData], prefix: str = '') -> None:
"""递归收集类成员,包括嵌套匿名类型"""
for item in class_node.body:
if isinstance(item, ast.AnnAssign) and isinstance(item.target, ast.Name):
member_name: str = item.target.id
full_name: str = f'{prefix}{member_name}' if prefix else member_name
member_type: CTypeInfo | None = CTypeInfo.FromNode(item.annotation, self._symtab)
if member_type is None:
member_type = CTypeInfo()
member_type.BaseType = t.CInt()
is_ptr: bool = member_type.IsPtr
if isinstance(item.annotation, ast.BinOp):
if AnnotationContainsName(item.annotation, 'CPtr'):
is_ptr = True
members[full_name] = member_type.Copy()
elif isinstance(item, ast.ClassDef):
is_anonymous: bool = any(
any(AnnotationContainsTType(base, 'Anonymous') for base in item.bases)
) if item.bases else False
if is_anonymous:
is_union: bool = any(
any(AnnotationContainsTType(base, 'CUnion') for base in item.bases)
)
# 也检查装饰器形式 @t.CUnion
if not is_union and hasattr(item, 'decorator_list') and item.decorator_list:
for deco in item.decorator_list:
if isinstance(deco, ast.Attribute) and hasattr(deco.value, 'id') and IsTModule(deco.value.id, self._symtab) and deco.attr == 'CUnion':
is_union = True
break
elif isinstance(deco, ast.Name):
cls2: type | None = getattr(t, deco.id, None)
if cls2 is None and hasattr(self._symtab, '_t_type_symbols'):
cls2 = self._symtab._t_type_symbols.get(deco.id)
if cls2 is not None and isinstance(cls2, type) and issubclass(cls2, t.CUnion):
is_union = True
break
nested_name: str = f'{prefix}{item.name}' if prefix else item.name
members[nested_name] = CTypeInfo()
members[nested_name].BaseType = f'union {item.name}' if is_union else f'struct {item.name}'
members[nested_name].PtrCount = 0
members[nested_name].ArrayDims = []
new_prefix: str = f'{prefix}{item.name}.' if prefix else f'{item.name}.'
self._collect_members(item, members, anonymous_types, new_prefix)
anonymous_members: dict[str, CTypeInfo] = {}
for full_name, member_info in members.items():
if full_name.startswith(f"{nested_name}."):
member_name: str = full_name[len(nested_name) + 1:]
anonymous_members[member_name] = member_info
if item.name not in anonymous_types:
anonymous_types[item.name] = AnonymousTypeData(
name=item.name,
is_union=is_union,
members=anonymous_members,
lineno=item.lineno
)
# ------------------------------------------------------------------
# 类符号提取
# ------------------------------------------------------------------
def _extract_class(self, node: ast.ClassDef, members_info: dict[str, CTypeInfo]) -> ClassSymbolData:
"""从 ClassDef 节点提取类符号数据"""
type_kind: str = 'struct'
is_cpython_object: bool = False
is_packed: bool = False
# 检查装饰器
if hasattr(node, 'decorator_list') and node.decorator_list:
for decorator in node.decorator_list:
if isinstance(decorator, ast.Attribute):
if (hasattr(decorator.value, 'id') and
IsTModule(decorator.value.id, self._symtab)):
if decorator.attr == 'Object':
is_cpython_object = True
elif decorator.attr == 'CUnion':
type_kind = 'union'
elif decorator.attr == 'CEnum':
type_kind = 'enum'
elif decorator.attr == 'REnum':
type_kind = 'renum'
elif decorator.attr == 'CStruct':
type_kind = 'struct'
elif isinstance(decorator, ast.Name):
# from t import CStruct / CUnion / CEnum / REnum
cls: type | None = getattr(t, decorator.id, None)
if cls is None and hasattr(self._symtab, '_t_type_symbols'):
cls = self._symtab._t_type_symbols.get(decorator.id)
if cls is not None and isinstance(cls, type) and issubclass(cls, t.CType):
if issubclass(cls, t.CUnion):
type_kind = 'union'
elif issubclass(cls, t.REnum):
type_kind = 'renum'
elif issubclass(cls, t.CEnum):
type_kind = 'enum'
elif issubclass(cls, t.CStruct):
type_kind = 'struct'
elif decorator.id == 'Object':
is_cpython_object = True
elif isinstance(decorator, ast.Call):
if isinstance(decorator.func, ast.Attribute):
if hasattr(decorator.func.value, 'id') and decorator.func.value.id == 'c' and decorator.func.attr == 'Attribute':
for arg in decorator.args:
if isinstance(arg, ast.Attribute):
if isinstance(arg.value, ast.Attribute):
if hasattr(arg.value.value, 'id') and IsTModule(arg.value.value.id, self._symtab) and arg.value.attr == 'attr' and arg.attr == 'packed':
is_packed = True
# 检查基类
for base in node.bases:
if AnnotationContainsTType(base, 'CUnion'):
type_kind = 'union'
break
elif AnnotationContainsTType(base, 'CEnum'):
type_kind = 'enum'
break
elif AnnotationContainsTType(base, 'REnum'):
type_kind = 'renum'
break
elif AnnotationContainsTType(base, 'CStruct'):
type_kind = 'struct'
break
elif AnnotationContainsTType(base, 'Object'):
is_cpython_object = True
type_kind = 'struct'
break
elif isinstance(base, ast.Name) and base.id == 'Exception':
type_kind = 'exception'
break
elif isinstance(base, ast.Name):
base_entry: CTypeInfo | None = self._symtab.get(base.id)
if base_entry and base_entry.IsExceptionClass:
type_kind = 'exception'
break
# 提取枚举成员
enum_members: list[tuple[str, int, int]] = []
if type_kind == 'enum':
next_enum_value: int = 0
for item in node.body:
if isinstance(item, ast.Assign):
if len(item.targets) == 1 and isinstance(item.targets[0], ast.Name):
member_name: str = item.targets[0].id
if isinstance(item.value, ast.Constant) and isinstance(item.value.value, int):
next_enum_value = item.value.value + 1
else:
next_enum_value += 1
enum_members.append((member_name, next_enum_value - 1, item.lineno))
elif isinstance(item, ast.AnnAssign):
if isinstance(item.target, ast.Name):
member_name: str = item.target.id
if item.value:
if isinstance(item.value, ast.Constant) and isinstance(item.value.value, int):
next_enum_value = item.value.value + 1
else:
next_enum_value += 1
else:
next_enum_value += 1
enum_members.append((member_name, next_enum_value - 1, item.lineno))
# REnum 变体提取(嵌套 ClassDef + Assign 显式值)
# 与 HandlesClassDef._EmitREnumLlvm 的 tag 分配逻辑对齐:
# - 先收集 Assign 显式值
# - ClassDef 变体从 max(explicit)+1 开始顺序编号
# - Assign 无显式值时用独立计数器 variant_index
if type_kind == 'renum':
explicit_values: dict[str, int] = {}
for item in node.body:
if isinstance(item, ast.Assign):
for target in item.targets:
if isinstance(target, ast.Name):
if item.value and isinstance(item.value, ast.Constant) and isinstance(item.value.value, int):
explicit_values[target.id] = item.value.value
next_class_tag: int = max(explicit_values.values()) + 1 if explicit_values else 0
variant_index: int = 0
for item in node.body:
if isinstance(item, ast.ClassDef):
variant_name_r: str = item.name
tag: int = next_class_tag
next_class_tag += 1
enum_members.append((variant_name_r, tag, item.lineno))
elif isinstance(item, ast.Assign):
for target in item.targets:
if isinstance(target, ast.Name):
var_name: str = target.id
if item.value and isinstance(item.value, ast.Constant) and isinstance(item.value.value, int):
value: int = item.value.value
else:
value = variant_index
variant_index += 1
enum_members.append((var_name, value, item.lineno))
return ClassSymbolData(
name=node.name,
type_kind=type_kind,
lineno=node.lineno,
members=members_info,
is_cpython_object=is_cpython_object,
is_packed=is_packed,
enum_members=enum_members,
)
# ------------------------------------------------------------------
# AnnAssign 提取typedef / define
# ------------------------------------------------------------------
def _extract_ann_assign(self, node: ast.AnnAssign, result: ModuleSymbols) -> None:
"""从 AnnAssign 节点提取 typedef 或 define 数据"""
var_name: str = node.target.id
has_cdefine: bool = AnnotationContainsTType(node.annotation, 'CDefine') if node.annotation else False
has_postdef: bool = AnnotationContainsTType(node.annotation, 'Postdefinition') if node.annotation else False
has_ctypedef: bool = AnnotationContainsTType(node.annotation, 'CTypedef') if node.annotation else False
if has_cdefine:
result.defines.append(DefineSymbolData(
name=var_name,
lineno=node.lineno,
value_node=node.value,
))
else:
result.typedefs.append(TypedefSymbolData(
name=var_name,
lineno=node.lineno,
annotation=node.annotation,
value=node.value,
has_cdefine=has_cdefine,
has_postdef=has_postdef,
has_ctypedef=has_ctypedef,
))
# ------------------------------------------------------------------
# 函数提取
# ------------------------------------------------------------------
def _extract_function(self, node: ast.FunctionDef) -> FuncSymbolData:
"""从 FunctionDef 节点提取函数符号数据"""
params: list[tuple[str, ast.AST | None]] = [(arg.arg, arg.annotation) for arg in node.args.args]
is_variadic: bool = node.args.vararg is not None
is_inline: bool = CheckAnnotationHasCInline(node.returns)
return FuncSymbolData(
name=node.name,
lineno=node.lineno,
returns_node=node.returns,
params=params,
is_variadic=is_variadic,
is_inline=is_inline,
)