# -*- coding: utf-8 -*-
#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you 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 inspect
import os
import pickle
import subprocess
import sys
import types
from builtins import str
from textwrap import dedent
from typing import Optional, Iterable, Dict, Callable
import dill
from airflow.exceptions import AirflowException
from airflow.models import BaseOperator, SkipMixin
from airflow.utils.decorators import apply_defaults
from airflow.utils.file import TemporaryDirectory
from airflow.utils.operator_helpers import context_to_airflow_vars
[docs]class PythonOperator(BaseOperator):
"""
Executes a Python callable
.. seealso::
For more information on how to use this operator, take a look at the guide:
:ref:`howto/operator:PythonOperator`
:param python_callable: A reference to an object that is callable
:type python_callable: python callable
:param op_kwargs: a dictionary of keyword arguments that will get unpacked
in your function
:type op_kwargs: dict (templated)
:param op_args: a list of positional arguments that will get unpacked when
calling your callable
:type op_args: list (templated)
:param provide_context: if set to true, Airflow will pass a set of
keyword arguments that can be used in your function. This set of
kwargs correspond exactly to what you can use in your jinja
templates. For this to work, you need to define `**kwargs` in your
function header.
:type provide_context: bool
:param templates_dict: a dictionary where the values are templates that
will get templated by the Airflow engine sometime between
``__init__`` and ``execute`` takes place and are made available
in your callable's context after the template has been applied. (templated)
:type templates_dict: dict[str]
:param templates_exts: a list of file extensions to resolve while
processing templated fields, for examples ``['.sql', '.hql']``
:type templates_exts: list[str]
"""
[docs] template_fields = ('templates_dict', 'op_args', 'op_kwargs')
# since we won't mutate the arguments, we should just do the shallow copy
# there are some cases we can't deepcopy the objects(e.g protobuf).
[docs] shallow_copy_attrs = ('python_callable', 'op_kwargs',)
@apply_defaults
def __init__(
self,
python_callable, # type: Callable
op_args=None, # type: Optional[Iterable]
op_kwargs=None, # type: Optional[Dict]
provide_context=False, # type: bool
templates_dict=None, # type: Optional[Dict]
templates_exts=None, # type: Optional[Iterable[str]]
*args,
**kwargs
):
super(PythonOperator, self).__init__(*args, **kwargs)
if not callable(python_callable):
raise AirflowException('`python_callable` param must be callable')
self.python_callable = python_callable
self.op_args = op_args or []
self.op_kwargs = op_kwargs or {}
self.provide_context = provide_context
self.templates_dict = templates_dict
if templates_exts:
self.template_ext = templates_exts
[docs] def execute(self, context):
# Export context to make it available for callables to use.
airflow_context_vars = context_to_airflow_vars(context, in_env_var_format=True)
self.log.debug("Exporting the following env vars:\n%s",
'\n'.join(["{}={}".format(k, v)
for k, v in airflow_context_vars.items()]))
os.environ.update(airflow_context_vars)
if self.provide_context:
context.update(self.op_kwargs)
context['templates_dict'] = self.templates_dict
self.op_kwargs = context
return_value = self.execute_callable()
self.log.info("Done. Returned value was: %s", return_value)
return return_value
[docs] def execute_callable(self):
return self.python_callable(*self.op_args, **self.op_kwargs)
[docs]class BranchPythonOperator(PythonOperator, SkipMixin):
"""
Allows a workflow to "branch" or follow a path following the execution
of this task.
It derives the PythonOperator and expects a Python function that returns
a single task_id or list of task_ids to follow. The task_id(s) returned
should point to a task directly downstream from {self}. All other "branches"
or directly downstream tasks are marked with a state of ``skipped`` so that
these paths can't move forward. The ``skipped`` states are propagated
downstream to allow for the DAG state to fill up and the DAG run's state
to be inferred.
Note that using tasks with ``depends_on_past=True`` downstream from
``BranchPythonOperator`` is logically unsound as ``skipped`` status
will invariably lead to block tasks that depend on their past successes.
``skipped`` states propagates where all directly upstream tasks are
``skipped``.
"""
[docs] def execute(self, context):
branch = super(BranchPythonOperator, self).execute(context)
self.skip_all_except(context['ti'], branch)
[docs]class ShortCircuitOperator(PythonOperator, SkipMixin):
"""
Allows a workflow to continue only if a condition is met. Otherwise, the
workflow "short-circuits" and downstream tasks are skipped.
The ShortCircuitOperator is derived from the PythonOperator. It evaluates a
condition and short-circuits the workflow if the condition is False. Any
downstream tasks are marked with a state of "skipped". If the condition is
True, downstream tasks proceed as normal.
The condition is determined by the result of `python_callable`.
"""
[docs] def execute(self, context):
condition = super(ShortCircuitOperator, self).execute(context)
self.log.info("Condition result is %s", condition)
if condition:
self.log.info('Proceeding with downstream tasks...')
return
self.log.info('Skipping downstream tasks...')
downstream_tasks = context['task'].get_flat_relatives(upstream=False)
self.log.debug("Downstream task_ids %s", downstream_tasks)
if downstream_tasks:
self.skip(context['dag_run'], context['ti'].execution_date, downstream_tasks)
self.log.info("Done.")
[docs]class PythonVirtualenvOperator(PythonOperator):
"""
Allows one to run a function in a virtualenv that is created and destroyed
automatically (with certain caveats).
The function must be defined using def, and not be
part of a class. All imports must happen inside the function
and no variables outside of the scope may be referenced. A global scope
variable named virtualenv_string_args will be available (populated by
string_args). In addition, one can pass stuff through op_args and op_kwargs, and one
can use a return value.
Note that if your virtualenv runs in a different Python major version than Airflow,
you cannot use return values, op_args, or op_kwargs. You can use string_args though.
:param python_callable: A python function with no references to outside variables,
defined with def, which will be run in a virtualenv
:type python_callable: function
:param requirements: A list of requirements as specified in a pip install command
:type requirements: list[str]
:param python_version: The Python version to run the virtualenv with. Note that
both 2 and 2.7 are acceptable forms.
:type python_version: str
:param use_dill: Whether to use dill to serialize
the args and result (pickle is default). This allow more complex types
but requires you to include dill in your requirements.
:type use_dill: bool
:param system_site_packages: Whether to include
system_site_packages in your virtualenv.
See virtualenv documentation for more information.
:type system_site_packages: bool
:param op_args: A list of positional arguments to pass to python_callable.
:type op_kwargs: list
:param op_kwargs: A dict of keyword arguments to pass to python_callable.
:type op_kwargs: dict
:param provide_context: if set to true, Airflow will pass a set of
keyword arguments that can be used in your function. This set of
kwargs correspond exactly to what you can use in your jinja
templates. For this to work, you need to define `**kwargs` in your
function header.
:type provide_context: bool
:param string_args: Strings that are present in the global var virtualenv_string_args,
available to python_callable at runtime as a list[str]. Note that args are split
by newline.
:type string_args: list[str]
:param templates_dict: a dictionary where the values are templates that
will get templated by the Airflow engine sometime between
``__init__`` and ``execute`` takes place and are made available
in your callable's context after the template has been applied
:type templates_dict: dict of str
:param templates_exts: a list of file extensions to resolve while
processing templated fields, for examples ``['.sql', '.hql']``
:type templates_exts: list[str]
"""
@apply_defaults
def __init__(
self,
python_callable, # type: Callable
requirements=None, # type: Optional[Iterable[str]]
python_version=None, # type: Optional[str]
use_dill=False, # type: bool
system_site_packages=True, # type: bool
op_args=None, # type: Optional[Iterable]
op_kwargs=None, # type: Optional[Dict]
provide_context=False, # type: bool
string_args=None, # type: Optional[Iterable[str]]
templates_dict=None, # type: Optional[Dict]
templates_exts=None, # type: Optional[Iterable[str]]
*args,
**kwargs
):
super(PythonVirtualenvOperator, self).__init__(
python_callable=python_callable,
op_args=op_args,
op_kwargs=op_kwargs,
templates_dict=templates_dict,
templates_exts=templates_exts,
provide_context=provide_context,
*args,
**kwargs)
self.requirements = requirements or []
self.string_args = string_args or []
self.python_version = python_version
self.use_dill = use_dill
self.system_site_packages = system_site_packages
# check that dill is present if needed
dill_in_requirements = map(lambda x: x.lower().startswith('dill'),
self.requirements)
if (not system_site_packages) and use_dill and not any(dill_in_requirements):
raise AirflowException('If using dill, dill must be in the environment ' +
'either via system_site_packages or requirements')
# check that a function is passed, and that it is not a lambda
if (not isinstance(self.python_callable,
types.FunctionType) or (self.python_callable.__name__ ==
(lambda x: 0).__name__)):
raise AirflowException('{} only supports functions for python_callable arg',
self.__class__.__name__)
# check that args are passed iff python major version matches
if (python_version is not None and
str(python_version)[0] != str(sys.version_info[0]) and
self._pass_op_args()):
raise AirflowException("Passing op_args or op_kwargs is not supported across "
"different Python major versions "
"for PythonVirtualenvOperator. "
"Please use string_args.")
[docs] def execute_callable(self):
with TemporaryDirectory(prefix='venv') as tmp_dir:
if self.templates_dict:
self.op_kwargs['templates_dict'] = self.templates_dict
# generate filenames
input_filename = os.path.join(tmp_dir, 'script.in')
output_filename = os.path.join(tmp_dir, 'script.out')
string_args_filename = os.path.join(tmp_dir, 'string_args.txt')
script_filename = os.path.join(tmp_dir, 'script.py')
# set up virtualenv
self._execute_in_subprocess(self._generate_virtualenv_cmd(tmp_dir))
cmd = self._generate_pip_install_cmd(tmp_dir)
if cmd:
self._execute_in_subprocess(cmd)
self._write_args(input_filename)
self._write_script(script_filename)
self._write_string_args(string_args_filename)
# execute command in virtualenv
self._execute_in_subprocess(
self._generate_python_cmd(tmp_dir,
script_filename,
input_filename,
output_filename,
string_args_filename))
return self._read_result(output_filename)
[docs] def _pass_op_args(self):
# we should only pass op_args if any are given to us
return len(self.op_args) + len(self.op_kwargs) > 0
[docs] def _execute_in_subprocess(self, cmd):
try:
self.log.info("Executing cmd\n%s", cmd)
output = subprocess.check_output(cmd,
stderr=subprocess.STDOUT,
close_fds=True)
if output:
self.log.info("Got output\n%s", output)
except subprocess.CalledProcessError as e:
self.log.info("Got error output\n%s", e.output)
raise
[docs] def _write_string_args(self, filename):
# writes string_args to a file, which are read line by line
with open(filename, 'w') as f:
f.write('\n'.join(map(str, self.string_args)))
[docs] def _write_args(self, input_filename):
# serialize args to file
if self.use_dill:
serializer = dill
else:
serializer = pickle
# some items from context can't be loaded in virtual env
# see pr https://github.com/apache/airflow/pull/8256
not_serializable = {'dag', 'task', 'ti', 'macros', 'task_instance', 'var'}
if self._pass_op_args():
kwargs = {key: value for key, value in self.op_kwargs.items()
if key not in not_serializable}
with open(input_filename, 'wb') as f:
arg_dict = ({'args': self.op_args, 'kwargs': kwargs})
serializer.dump(arg_dict, f)
[docs] def _read_result(self, output_filename):
if os.stat(output_filename).st_size == 0:
return None
with open(output_filename, 'rb') as f:
try:
if self.use_dill:
return dill.load(f)
else:
return pickle.load(f)
except ValueError:
self.log.error("Error deserializing result. "
"Note that result deserialization "
"is not supported across major Python versions.")
raise
[docs] def _write_script(self, script_filename):
with open(script_filename, 'w') as f:
python_code = self._generate_python_code()
self.log.debug('Writing code to file\n{}'.format(python_code))
f.write(python_code)
[docs] def _generate_virtualenv_cmd(self, tmp_dir):
cmd = ['virtualenv', tmp_dir]
if self.system_site_packages:
cmd.append('--system-site-packages')
if self.python_version is not None:
cmd.append('--python=python{}'.format(self.python_version))
return cmd
[docs] def _generate_pip_install_cmd(self, tmp_dir):
if len(self.requirements) == 0:
return []
else:
# direct path alleviates need to activate
cmd = ['{}/bin/pip'.format(tmp_dir), 'install']
return cmd + self.requirements
@staticmethod
[docs] def _generate_python_cmd(tmp_dir, script_filename,
input_filename, output_filename, string_args_filename):
# direct path alleviates need to activate
return ['{}/bin/python'.format(tmp_dir), script_filename,
input_filename, output_filename, string_args_filename]
[docs] def _generate_python_code(self):
if self.use_dill:
pickling_library = 'dill'
else:
pickling_library = 'pickle'
fn = self.python_callable
# dont try to read pickle if we didnt pass anything
if self._pass_op_args():
load_args_line = 'with open(sys.argv[1], "rb") as f: arg_dict = {}.load(f)'\
.format(pickling_library)
else:
load_args_line = 'arg_dict = {"args": [], "kwargs": {}}'
# no indents in original code so we can accept
# any type of indents in the original function
# we deserialize args, call function, serialize result if necessary
return dedent("""\
import {pickling_library}
import sys
{load_args_code}
args = arg_dict["args"]
kwargs = arg_dict["kwargs"]
with open(sys.argv[3], 'r') as f:
virtualenv_string_args = list(map(lambda x: x.strip(), list(f)))
{python_callable_lines}
res = {python_callable_name}(*args, **kwargs)
with open(sys.argv[2], 'wb') as f:
res is not None and {pickling_library}.dump(res, f)
""").format(load_args_code=load_args_line,
python_callable_lines=dedent(inspect.getsource(fn)),
python_callable_name=fn.__name__,
pickling_library=pickling_library)