snapshot before regression test
This commit is contained in:
t
231
lib/Projectrans/Utils.py
Normal file
231
lib/Projectrans/Utils.py
Normal file
@@ -0,0 +1,231 @@
|
||||
from __future__ import annotations
|
||||
|
||||
import hashlib
|
||||
import os
|
||||
import ast
|
||||
|
||||
from lib.core.VLogger import get_logger as _vlog
|
||||
from lib.constants.config import mode as _ConfigMode
|
||||
from lib.core.SymbolUtils import AnnotationContainsName
|
||||
|
||||
|
||||
def compute_sha1(content: str) -> str:
|
||||
return hashlib.sha1(content.encode('utf-8')).hexdigest()[:16]
|
||||
|
||||
|
||||
# get_file_dependencies 缓存：键 (filepath, src_root)，值 set[str]
|
||||
# 单次编译中同一文件被 find_reachable_files 和 topological_sort 重复解析，
|
||||
# 缓存避免重复 open+read+ast.parse+ast.walk（~350 次调用→~175 次）
|
||||
_FILE_DEPS_CACHE: dict[tuple[str, str], set[str]] = {}
|
||||
|
||||
# Python 源文件 AST 缓存：键 filepath，值 (mtime, content, tree)
|
||||
# 多处解析同一 .py 文件（get_file_dependencies、used_includes 收集、
|
||||
# _collect_inline_symbols 等）共享 parse 结果，避免重复 ast.parse（~18s→~9s）
|
||||
_FILE_AST_CACHE: dict[str, tuple[float, str, ast.Module]] = {}
|
||||
|
||||
|
||||
def _check_annotation_for_export(annotation: ast.AST | None) -> bool:
|
||||
"""递归检查类型注解中是否包含 CExport/CExtern/State（t.State = t.CExtern | t.CExport）"""
|
||||
if not annotation:
|
||||
return False
|
||||
return (AnnotationContainsName(annotation, 'CExport') or
|
||||
AnnotationContainsName(annotation, 'CExtern') or
|
||||
AnnotationContainsName(annotation, 'State'))
|
||||
|
||||
|
||||
def sha1_file(filepath: str) -> str:
|
||||
with open(filepath, 'r', encoding='utf-8') as f:
|
||||
content: str = f.read()
|
||||
return compute_sha1(content)
|
||||
|
||||
|
||||
def parse_python_file(filepath: str) -> tuple[str, ast.Module] | tuple[str, None]:
|
||||
"""读取并解析 Python 文件，带 mtime 校验的 AST 缓存
|
||||
|
||||
多处调用共享同一文件的 parse 结果，避免重复 open+read+ast.parse。
|
||||
返回 (content, tree)；文件读取或解析失败返回 (content, None)。
|
||||
"""
|
||||
try:
|
||||
st: os.stat_result = os.stat(filepath)
|
||||
except OSError:
|
||||
return ('', None)
|
||||
mtime: float = st.st_mtime
|
||||
cached: tuple[float, str, ast.Module] | None = _FILE_AST_CACHE.get(filepath)
|
||||
if cached is not None and cached[0] == mtime:
|
||||
return (cached[1], cached[2])
|
||||
try:
|
||||
with open(filepath, 'r', encoding='utf-8') as f:
|
||||
content: str = f.read()
|
||||
except Exception:
|
||||
return ('', None)
|
||||
try:
|
||||
tree: ast.Module = ast.parse(content)
|
||||
except SyntaxError:
|
||||
return (content, None)
|
||||
_FILE_AST_CACHE[filepath] = (mtime, content, tree)
|
||||
return (content, tree)
|
||||
|
||||
|
||||
def get_file_dependencies(filepath: str, src_root: str) -> set[str]:
|
||||
"""解析 Python 文件的导入依赖（带缓存）
|
||||
|
||||
单次编译中 find_reachable_files_from_entries 和 topological_sort_files
|
||||
会对同一批文件重复调用，缓存避免重复 ast.parse（~8.6s→~4.3s）。
|
||||
"""
|
||||
cache_key: tuple[str, str] = (filepath, src_root)
|
||||
cached: set[str] | None = _FILE_DEPS_CACHE.get(cache_key)
|
||||
if cached is not None:
|
||||
return cached
|
||||
dependencies: set[str] = set()
|
||||
content, tree = parse_python_file(filepath)
|
||||
if tree is not None:
|
||||
try:
|
||||
for node in ast.walk(tree):
|
||||
if isinstance(node, ast.Import):
|
||||
for alias in node.names:
|
||||
module: str = alias.name
|
||||
dependencies.add(module)
|
||||
elif isinstance(node, ast.ImportFrom):
|
||||
if node.module:
|
||||
module: str = node.module
|
||||
if node.level > 0:
|
||||
# from .module import name → 依赖 pkg.module（模块本身）
|
||||
pkg_dir: str = os.path.dirname(filepath)
|
||||
pkg_rel: str = os.path.relpath(pkg_dir, src_root).replace(os.sep, '.').replace('/', '.')
|
||||
if pkg_rel == '.':
|
||||
pkg_rel = ''
|
||||
dep_module: str
|
||||
if pkg_rel:
|
||||
dep_module = f'{pkg_rel}.{module}'
|
||||
else:
|
||||
dep_module = module
|
||||
dependencies.add(dep_module)
|
||||
else:
|
||||
if not module.startswith('_'):
|
||||
dependencies.add(module)
|
||||
elif node.level > 0 and node.names:
|
||||
# from . import name → 依赖 pkg.name（子模块）
|
||||
for alias in node.names:
|
||||
pkg_dir: str = os.path.dirname(filepath)
|
||||
pkg_rel: str = os.path.relpath(pkg_dir, src_root).replace(os.sep, '.').replace('/', '.')
|
||||
if pkg_rel == '.':
|
||||
pkg_rel = ''
|
||||
dep_module: str = f'{pkg_rel}.{alias.name}' if pkg_rel else alias.name
|
||||
dependencies.add(dep_module)
|
||||
except Exception as _e:
|
||||
if _ConfigMode == "strict":
|
||||
raise
|
||||
_vlog().warning(f"解析文件导入依赖失败: {_e}", "Exception")
|
||||
_FILE_DEPS_CACHE[cache_key] = dependencies
|
||||
return dependencies
|
||||
|
||||
|
||||
def clear_file_deps_cache() -> None:
|
||||
"""清除 get_file_dependencies 和 AST 缓存（跨编译或文件变更时调用）"""
|
||||
_FILE_DEPS_CACHE.clear()
|
||||
_FILE_AST_CACHE.clear()
|
||||
|
||||
|
||||
def find_reachable_files_from_entries(src_root: str, entry_files: list[str]) -> set[str]:
|
||||
"""从入口文件出发，找出所有可达的 .py 文件（通过导入关系）"""
|
||||
all_files: dict[str, str] = {} # module_name -> filepath
|
||||
for root, dirs, files in os.walk(src_root):
|
||||
dirs[:] = [d for d in dirs if d not in ('__pycache__',)]
|
||||
for file in files:
|
||||
if file.endswith('.py'):
|
||||
filepath: str = os.path.join(root, file)
|
||||
rel: str = os.path.relpath(filepath, src_root)
|
||||
module: str = os.path.splitext(rel)[0].replace(os.sep, '.').replace('/', '.')
|
||||
if filepath.endswith('__init__.py'):
|
||||
all_files[module] = filepath
|
||||
parent_module: str = module.rsplit('.', 1)[0] if '.' in module else module
|
||||
if parent_module not in all_files or not all_files[parent_module].endswith('__init__.py'):
|
||||
all_files[parent_module] = filepath
|
||||
elif module not in all_files:
|
||||
all_files[module] = filepath
|
||||
top_module: str = module.split('.')[0]
|
||||
if top_module not in all_files:
|
||||
all_files[top_module] = filepath
|
||||
|
||||
reachable: set[str] = set()
|
||||
queue: list[str] = list(entry_files)
|
||||
|
||||
# BFS trace: record each file processed and unresolved deps
|
||||
_bfs_trace: list[str] = []
|
||||
_unresolved_deps: list[tuple[str, str]] = [] # (current_file, dep)
|
||||
|
||||
while queue:
|
||||
current: str = queue.pop(0)
|
||||
if current in reachable:
|
||||
continue
|
||||
reachable.add(current)
|
||||
deps: set[str] = get_file_dependencies(current, src_root)
|
||||
_bfs_trace.append(f" {os.path.basename(current)} deps={len(deps)}")
|
||||
for dep in deps:
|
||||
if dep in all_files and all_files[dep] not in reachable:
|
||||
queue.append(all_files[dep])
|
||||
else:
|
||||
_matched = False
|
||||
for key, val in all_files.items():
|
||||
if key.endswith('.' + dep) or key == dep:
|
||||
if val not in reachable:
|
||||
queue.append(val)
|
||||
_matched = True
|
||||
break
|
||||
if not _matched:
|
||||
_unresolved_deps.append((os.path.basename(current), dep))
|
||||
|
||||
return reachable
|
||||
|
||||
|
||||
def topological_sort_files(py_files: list[str], src_root: str) -> list[str]:
|
||||
"""对 Python 文件进行拓扑排序，确保依赖模块先被处理"""
|
||||
# 构建模块名到文件路径的映射
|
||||
module_to_file: dict[str, str] = {}
|
||||
for filepath in py_files:
|
||||
rel: str = os.path.relpath(filepath, src_root)
|
||||
module: str = os.path.splitext(rel)[0].replace(os.sep, '.').replace('/', '.')
|
||||
module_to_file[module] = filepath
|
||||
# 也添加顶层模块名
|
||||
top_module: str = module.split('.')[0]
|
||||
if top_module not in module_to_file:
|
||||
module_to_file[top_module] = filepath
|
||||
|
||||
# 构建依赖图
|
||||
graph: dict[str, set[str]] = {f: set() for f in py_files}
|
||||
for filepath in py_files:
|
||||
deps: set[str] = get_file_dependencies(filepath, src_root)
|
||||
for dep in deps:
|
||||
if dep in module_to_file:
|
||||
dep_file: str = module_to_file[dep]
|
||||
if dep_file != filepath: # 避免自依赖
|
||||
graph[filepath].add(dep_file)
|
||||
|
||||
# 拓扑排序 (Kahn's algorithm)
|
||||
in_degree: dict[str, int] = {f: 0 for f in py_files}
|
||||
for f, deps in graph.items():
|
||||
for dep in deps:
|
||||
if dep in in_degree:
|
||||
in_degree[f] += 1
|
||||
|
||||
queue: list[str] = [f for f, degree in in_degree.items() if degree == 0]
|
||||
sorted_files: list[str] = []
|
||||
|
||||
while queue:
|
||||
# 按字母顺序处理同级别的文件，确保确定性
|
||||
queue.sort()
|
||||
current: str = queue.pop(0)
|
||||
sorted_files.append(current)
|
||||
|
||||
for f, deps in graph.items():
|
||||
if current in deps:
|
||||
in_degree[f] -= 1
|
||||
if in_degree[f] == 0:
|
||||
queue.append(f)
|
||||
|
||||
# 如果有环，添加剩余文件
|
||||
if len(sorted_files) < len(py_files):
|
||||
remaining: list[str] = [f for f in py_files if f not in sorted_files]
|
||||
sorted_files.extend(remaining)
|
||||
|
||||
return sorted_files
|
||||
Reference in New Issue
Block a user