# Copyright 2019 Eli Lilly and Company
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import glob
import json
import os
from pathlib import Path
import re
import tempfile
from typing import Optional, Sequence, Union, cast
import subby
from pytest_wdl.executors import (
ExecutionFailedError,
JavaExecutor,
get_target_name,
read_write_inputs,
)
from pytest_wdl.utils import LOG, ensure_path
ENV_CROMWELL_JAR = "CROMWELL_JAR"
ENV_CROMWELL_CONFIG = "CROMWELL_CONFIG"
ENV_CROMWELL_ARGS = "CROMWELL_ARGS"
UNSAFE_RE = re.compile(r"[^\w.-]")
[docs]class Failures:
def __init__(
self,
num_failed: int,
failed_task: str,
failed_task_exit_status: Optional[str] = None,
failed_task_stdout: Optional[Union[Path, str]] = None,
failed_task_stderr: Optional[Union[Path, str]] = None
):
self.num_failed = num_failed
self.failed_task = failed_task
self.failed_task_exit_status = failed_task_exit_status
self._failed_task_stdout_path = None
self._failed_task_stdout = None
self._failed_task_stderr_path = None
self._failed_task_stderr = None
if isinstance(failed_task_stdout, Path):
self._failed_task_stdout_path = cast(Path, failed_task_stdout)
else:
self._failed_task_stdout = cast(str, failed_task_stdout)
if isinstance(failed_task_stderr, Path):
self._failed_task_stderr_path = cast(Path, failed_task_stderr)
else:
self._failed_task_stderr = cast(str, failed_task_stderr)
@property
def failed_task_stdout(self):
if self._failed_task_stdout is None and self._failed_task_stderr_path:
self._failed_task_stdout = Failures._read_task_std(
self._failed_task_stdout_path
)
return self._failed_task_stdout
@property
def failed_task_stderr(self):
if self._failed_task_stderr is None and self._failed_task_stderr_path:
self._failed_task_stderr = Failures._read_task_std(
self._failed_task_stderr_path
)
return self._failed_task_stderr
@staticmethod
def _read_task_std(path: Path) -> Optional[str]:
if path:
if not path.exists():
path = path.with_suffix(".background")
if path.exists():
with open(path, "rt") as inp:
return inp.read()
[docs]class CromwellExecutor(JavaExecutor):
"""
Manages the running of WDL workflows using Cromwell.
Args:
import_dirs: Relative or absolute paths to directories containing WDL
scripts that should be available as imports.
java_bin: Path to the java executable.
java_args: Default Java arguments to use; can be overidden by passing
`java_args=...` to `run_workflow`.
cromwell_jar_file: Path to the Cromwell JAR file.
cromwell_args: Default Cromwell arguments to use; can be overridden by
passing `cromwell_args=...` to `run_workflow`.
"""
def __init__(
self,
import_dirs: Optional[Sequence[Path]] = None,
java_bin: Optional[Union[str, Path]] = None,
java_args: Optional[str] = None,
cromwell_jar_file: Optional[Union[str, Path]] = None,
cromwell_config_file: Optional[Union[str, Path]] = None,
cromwell_args: Optional[str] = None
):
super().__init__(java_bin, java_args)
self._import_dirs = import_dirs
self._cromwell_jar_file = self.resolve_jar_file(
"cromwell*.jar", cromwell_jar_file, ENV_CROMWELL_JAR
)
if not cromwell_config_file:
config_file = os.environ.get(ENV_CROMWELL_CONFIG)
if config_file:
cromwell_config_file = ensure_path(config_file)
if cromwell_config_file:
self._cromwell_config_file = ensure_path(
cromwell_config_file, is_file=True, exists=True
)
else:
self._cromwell_config_file = None
if not self.java_args and self._cromwell_config_file:
self.java_args = f"-Dconfig.file={self._cromwell_config_file}"
self._cromwell_args = cromwell_args or os.environ.get(ENV_CROMWELL_ARGS)
[docs] def run_workflow(
self,
wdl_path: Path,
inputs: Optional[dict] = None,
expected: Optional[dict] = None,
**kwargs
) -> dict:
"""
Run a WDL workflow on given inputs, and check that the output matches
given expected values.
Args:
wdl_path: The WDL script to execute.
inputs: Object that will be serialized to JSON and provided to Cromwell
as the workflow inputs.
expected: Dict mapping output parameter names to expected values.
kwargs: Additional keyword arguments, mostly for debugging:
* workflow_name: The name of the workflow in the WDL script. If None,
the name of the WDL script is used (without the .wdl extension).
* inputs_file: Path to the Cromwell inputs file to use. Inputs are
written to this file only if it doesn't exist.
* imports_file: Path to the WDL imports file to use. Imports are
written to this file only if it doesn't exist.
* java_args: Additional arguments to pass to Java runtime.
* cromwell_args: Additional arguments to pass to `cromwell run`.
Returns:
Dict of outputs.
Raises:
ExecutionFailedError: if there was an error executing Cromwell
AssertionError: if the actual outputs don't match the expected outputs
"""
target, is_task = get_target_name(
wdl_path=wdl_path, import_dirs=self._import_dirs, **kwargs
)
if is_task:
raise ValueError(
"Cromwell cannot execute tasks independently of a workflow"
)
inputs_dict, inputs_file = read_write_inputs(
inputs_dict=inputs, namespace=target
)
imports_file = self._get_workflow_imports(kwargs.get("imports_file"))
inputs_arg = f"-i {inputs_file}" if inputs_file else ""
imports_zip_arg = f"-p {imports_file}" if imports_file else ""
java_args = kwargs.get("java_args", self.java_args) or ""
cromwell_args = kwargs.get("cromwell_args", self._cromwell_args) or ""
metadata_file = Path.cwd() / "metadata.json"
cmd = (
f"{self.java_bin} {java_args} -jar {self._cromwell_jar_file} run "
f"-m {metadata_file} {cromwell_args} {inputs_arg} {imports_zip_arg} "
f"{wdl_path}"
)
LOG.info(
f"Executing cromwell command '{cmd}' with inputs "
f"{json.dumps(inputs_dict, default=str)}"
)
exe = subby.run(cmd, raise_on_error=False)
metadata = None
if metadata_file.exists():
with open(metadata_file, "rt") as inp:
metadata = json.load(inp)
if exe.ok:
if metadata:
assert metadata["status"] == "Succeeded"
outputs = metadata["outputs"]
else:
LOG.warning(
f"Cromwell command completed successfully but did not generate "
f"a metadata file at {metadata_file}"
)
outputs = self._get_cromwell_outputs(exe.output)
else:
error_kwargs = {
"executor": "cromwell",
"target": target,
"status": "Failed",
"inputs": inputs_dict,
"executor_stdout": exe.output,
"executor_stderr": exe.error,
}
if metadata:
failures = self._get_failures(metadata)
if failures:
error_kwargs.update({
"failed_task": failures.failed_task,
"failed_task_exit_status": failures.failed_task_exit_status,
"failed_task_stdout": failures.failed_task_stdout,
"failed_task_stderr": failures.failed_task_stderr
})
if failures.num_failed > 1:
error_kwargs["msg"] = \
f"cromwell failed on {failures.num_failed} instances of " \
f"{failures.failed_task} of {target}; only " \
f"showing output from the first failed task"
else:
error_kwargs["msg"] = f"cromwell failed on workflow {target}"
else:
error_kwargs["msg"] = \
f"Cromwell command failed but did not generate a metadata " \
f"file at {metadata_file}"
raise ExecutionFailedError(**error_kwargs)
if expected:
self._validate_outputs(outputs, expected, target)
return outputs
def _get_workflow_imports(self, imports_file: Optional[Path] = None) -> Path:
"""
Creates a ZIP file with all WDL files to be imported.
Args:
imports_file: Text file naming import directories/files - one per line.
Returns:
Path to the ZIP file.
"""
write_imports = bool(self._import_dirs)
imports_path = None
if imports_file:
imports_path = ensure_path(imports_file)
if imports_path.exists():
write_imports = False
if write_imports and self._import_dirs:
imports = [
wdl
for path in self._import_dirs
for wdl in glob.glob(str(path / "*.wdl"))
]
if imports:
if imports_path:
ensure_path(imports_path, is_file=True, create=True)
else:
imports_path = Path(tempfile.mkstemp(suffix=".zip")[1])
imports_str = " ".join(imports)
LOG.info(f"Writing imports {imports_str} to zip file {imports_path}")
exe = subby.run(
f"zip -j - {imports_str}",
mode=bytes,
stdout=imports_path,
raise_on_error=False
)
if not exe.ok:
raise Exception(
f"Error creating imports zip file; stdout={exe.output}; "
f"stderr={exe.error}"
)
return imports_path
@classmethod
def _get_cromwell_outputs(cls, output) -> dict:
lines = output.splitlines(keepends=False)
if len(lines) < 2:
raise Exception(f"Invalid Cromwell output: {output}")
if lines[1].startswith("Usage"):
# If the cromwell command is not valid, usage is printed and the
# return code is 0 so it does not cause an exception above - we
# have to catch it here.
raise Exception("Invalid Cromwell command")
start = None
for i, line in enumerate(lines):
if line == "{" and lines[i+1].lstrip().startswith('"outputs":'):
start = i
elif line == "}" and start is not None:
end = i
break
else:
raise AssertionError("No outputs JSON found in Cromwell stdout")
return json.loads("\n".join(lines[start:(end + 1)]))["outputs"]
@classmethod
def _get_failures(cls, metadata: dict) -> Optional[Failures]:
for call_name, call_metadatas in metadata["calls"].items():
failed = list(filter(
lambda md: md["executionStatus"] == "Failed", call_metadatas
))
if failed:
failed_call = failed[0]
if "subWorkflowMetadata" in failed_call:
return cls._get_failures(
failed_call["subWorkflowMetadata"]
)
else:
if "returnCode" in failed_call:
rc = failed_call["returnCode"]
else:
rc = "Unknown"
stdout = stderr = None
if "stdout" in failed_call:
stdout = Path(failed_call["stdout"])
stderr = Path(failed_call["stderr"])
elif "failures" in failed_call:
failure = failed_call["failures"][0]
stderr = failure["message"]
if "causedBy" in failure:
stderr = "\n ".join(
[stderr] + [cb["message"] for cb in failure["causedBy"]]
)
return Failures(len(failed), call_name, rc, stdout, stderr)