fix: isolate path-backed python source modules

This commit is contained in:
lda
2026-06-15 01:45:47 +07:00 Verified
parent 93d3f06c0f
commit a74a58a829
2 changed files with 263 additions and 3 deletions
+117 -3
View File
@@ -3,8 +3,12 @@ from __future__ import annotations
import sys
from collections.abc import Mapping, Sequence
from dataclasses import dataclass
from hashlib import sha256
from importlib import import_module
from importlib import util as importlib_util
from importlib.machinery import ModuleSpec
from pathlib import Path
from types import ModuleType
from typing import Any, Protocol
from wf_authoring import NodeSpec, node
@@ -43,9 +47,11 @@ def load_python_source(
path: str | Path | None = None,
) -> CapabilitySource:
"""Load a trusted in-process Python source from a module registry object."""
if path is not None:
_ensure_import_path(Path(path))
module_obj = import_module(module)
module_obj = _import_source_module(
source_id=source_id,
module=module,
path=Path(path) if path is not None else None,
)
if not hasattr(module_obj, registry):
raise ValueError(f"missing registry object {registry!r} in module {module!r}")
raw_registry = getattr(module_obj, registry)
@@ -101,6 +107,114 @@ def _coerce_specs(raw_registry: object) -> list[NodeSpec[Any, Any]]:
return specs
def _import_source_module(
*,
source_id: str,
module: str,
path: Path | None,
) -> ModuleType:
"""Import a source module without letting same-name local modules collide.
Configured Python sources commonly use local names like `ops` under their
own source root. Importing those through `sys.path` and `import_module`
would share `sys.modules["ops"]` across unrelated examples/tests. When the
module file exists under the configured root, load it under a stable
synthetic name derived from the source id and root path.
"""
if path is None:
return import_module(module)
resolved_root = path.resolve()
module_file = _module_file_under_root(resolved_root, module)
if module_file is None:
_ensure_import_path(resolved_root)
return import_module(module)
synthetic_root = _synthetic_module_root(
source_id=source_id,
root=resolved_root,
)
_ensure_synthetic_packages(
synthetic_root=synthetic_root,
root=resolved_root,
module=module,
)
# Keep legacy absolute sibling imports (`import helper`) working where
# possible, but source isolation only applies to imports made through the
# synthetic package (`from . import helper`, `from ..shared import value`).
# Python's normal absolute import cache can still collide for local helper
# names; robust sources should use package-relative imports.
_ensure_import_path(resolved_root)
synthetic_name = f"{synthetic_root}.{module}"
cached = sys.modules.get(synthetic_name)
if cached is not None:
return cached
submodule_search_locations = (
[str(module_file.parent)] if module_file.name == "__init__.py" else None
)
spec = importlib_util.spec_from_file_location(
synthetic_name,
module_file,
submodule_search_locations=submodule_search_locations,
)
if spec is None or spec.loader is None:
raise ImportError(f"cannot load Python source module {module!r}")
module_obj = importlib_util.module_from_spec(spec)
sys.modules[synthetic_name] = module_obj
spec.loader.exec_module(module_obj)
return module_obj
def _module_file_under_root(root: Path, module: str) -> Path | None:
module_path = root.joinpath(*module.split("."))
file_path = module_path.with_suffix(".py")
if file_path.is_file():
return file_path
package_path = module_path / "__init__.py"
if package_path.is_file():
return package_path
return None
def _synthetic_module_root(*, source_id: str, root: Path) -> str:
digest = sha256(f"{source_id}\0{root}".encode("utf-8")).hexdigest()[:16]
safe_source = source_id.replace(".", "_").replace("-", "_")
return f"_wf_source_{safe_source}_{digest}"
def _ensure_synthetic_packages(
*,
synthetic_root: str,
root: Path,
module: str,
) -> None:
"""Create package shells so source modules can use relative imports."""
_ensure_package_module(synthetic_root, root)
parts = module.split(".")[:-1]
current_name = synthetic_root
current_path = root
for part in parts:
current_name = f"{current_name}.{part}"
current_path = current_path / part
_ensure_package_module(current_name, current_path)
def _ensure_package_module(name: str, path: Path) -> None:
existing = sys.modules.get(name)
if existing is not None:
existing.__path__ = [str(path)] # type: ignore[attr-defined]
return
package = ModuleType(name)
package.__file__ = str(path)
package.__package__ = name
package.__path__ = [str(path)] # type: ignore[attr-defined]
package.__spec__ = ModuleSpec(name, loader=None, is_package=True)
sys.modules[name] = package
def _ensure_import_path(path: Path) -> None:
"""Keep configured trusted source roots importable for deferred imports."""
resolved = str(path.resolve())
+146
View File
@@ -86,6 +86,152 @@ registry = [echo]
assert set(source.capabilities.node_specs) == {"local.ops.echo"}
def test_load_python_source_isolates_same_module_name_across_roots(
tmp_path: Path,
) -> None:
left_root = tmp_path / "left"
right_root = tmp_path / "right"
left_root.mkdir()
right_root.mkdir()
source_template = """
from __future__ import annotations
from pydantic import BaseModel
from wf_authoring import node
class Input(BaseModel):
text: str
class Output(BaseModel):
value: str
@node(name="{name}")
def operation(payload: Input) -> Output:
return Output(value=payload.text)
registry = [operation]
"""
(left_root / "ops.py").write_text(
source_template.format(name="left"), encoding="utf-8"
)
(right_root / "ops.py").write_text(
source_template.format(name="right"), encoding="utf-8"
)
left = load_python_source(
source_id="local.left",
path=left_root,
module="ops",
registry="registry",
)
right = load_python_source(
source_id="local.right",
path=right_root,
module="ops",
registry="registry",
)
assert set(left.capabilities.node_specs) == {"local.left.left"}
assert set(right.capabilities.node_specs) == {"local.right.right"}
def test_load_python_source_supports_relative_imports_under_source_root(
tmp_path: Path,
) -> None:
source_root = tmp_path / "relative_source"
package_root = source_root / "pkg"
package_root.mkdir(parents=True)
(package_root / "__init__.py").write_text("", encoding="utf-8")
(package_root / "labels.py").write_text(
'LABEL = "relative"\n',
encoding="utf-8",
)
(package_root / "ops.py").write_text(
"""
from __future__ import annotations
from pydantic import BaseModel
from wf_authoring import node
from .labels import LABEL
class Input(BaseModel):
text: str
class Output(BaseModel):
value: str
@node(name=LABEL)
def operation(payload: Input) -> Output:
return Output(value=payload.text)
registry = [operation]
""",
encoding="utf-8",
)
source = load_python_source(
source_id="local.relative",
path=source_root,
module="pkg.ops",
registry="registry",
)
assert set(source.capabilities.node_specs) == {"local.relative.relative"}
def test_load_python_source_supports_delayed_relative_imports(
tmp_path: Path,
) -> None:
source_root = tmp_path / "delayed_relative_source"
package_root = source_root / "pkg"
package_root.mkdir(parents=True)
(package_root / "__init__.py").write_text("", encoding="utf-8")
(package_root / "labels.py").write_text(
'LABEL = "delayed"\n',
encoding="utf-8",
)
(package_root / "registry.py").write_text(
"""
def build_registry():
from .ops import operation
return [operation]
""",
encoding="utf-8",
)
(package_root / "ops.py").write_text(
"""
from __future__ import annotations
from pydantic import BaseModel
from wf_authoring import node
from .labels import LABEL
class Input(BaseModel):
text: str
class Output(BaseModel):
value: str
@node(name=LABEL)
def operation(payload: Input) -> Output:
return Output(value=payload.text)
""",
encoding="utf-8",
)
source = load_python_source(
source_id="local.delayed",
path=source_root,
module="pkg.registry",
registry="build_registry",
)
assert set(source.capabilities.node_specs) == {"local.delayed.delayed"}
def test_load_python_source_propagates_enabled_flag() -> None:
source = load_python_source(
source_id="local.ops",