Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[SPARK-7721][PYTHON][TESTS] Adds PySpark coverage generation script #20204

Closed
wants to merge 3 commits into from
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -62,6 +62,8 @@ project/plugins/src_managed/
project/plugins/target/
python/lib/pyspark.zip
python/deps
python/test_coverage/coverage_data
python/test_coverage/htmlcov
python/pyspark/python
reports/
scalastyle-on-compile.generated.xml
Expand Down
21 changes: 21 additions & 0 deletions python/.coveragerc
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
#
# 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.
#

[run]
branch = true
parallel = true
data_file = ${COVERAGE_DIR}/coverage_data/coverage
69 changes: 69 additions & 0 deletions python/run-tests-with-coverage
Original file line number Diff line number Diff line change
@@ -0,0 +1,69 @@
#!/usr/bin/env bash

#
# 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.
#

set -o pipefail
set -e

# This variable indicates which coverage executable to run to combine coverages
# and generate HTMLs, for example, 'coverage3' in Python 3.
COV_EXEC="${COV_EXEC:-coverage}"
FWDIR="$(cd "`dirname $0`"; pwd)"
pushd "$FWDIR" > /dev/null
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need to use pushd and its corresponding popd at the end of this file? I guess we can simply use cd here.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

my 2 c: I think it's ok, I'd prefer it; might be useful in the future when more cd are added

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I see, no problem at all. I just wanted to confirm. Thanks!


# Ensure that coverage executable is installed.
if ! hash $COV_EXEC 2>/dev/null; then
echo "Missing coverage executable in your path, skipping PySpark coverage"
exit 1
fi

# Set up the directories for coverage results.
export COVERAGE_DIR="$FWDIR/test_coverage"
rm -fr "$COVERAGE_DIR/coverage_data"
rm -fr "$COVERAGE_DIR/htmlcov"
mkdir -p "$COVERAGE_DIR/coverage_data"

# Current directory are added in the python path so that it doesn't refer our built
# pyspark zip library first.
export PYTHONPATH="$FWDIR:$PYTHONPATH"
# Also, our sitecustomize.py and coverage_daemon.py are included in the path.
export PYTHONPATH="$COVERAGE_DIR:$PYTHONPATH"

# We use 'spark.python.daemon.module' configuration to insert the coverage supported workers.
export SPARK_CONF_DIR="$COVERAGE_DIR/conf"

# This environment variable enables the coverage.
export COVERAGE_PROCESS_START="$FWDIR/.coveragerc"

# If you'd like to run a specific unittest class, you could do such as
# SPARK_TESTING=1 ../bin/pyspark pyspark.sql.tests VectorizedUDFTests
./run-tests "$@"
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Another tip is, if we use ../bin/pyspark here, do some simple tests and then exit, it looks still producing the coverage correctly.


# Don't run coverage for the coverage command itself
unset COVERAGE_PROCESS_START

# Coverage could generate empty coverage data files. Remove it to get rid of warnings when combining.
find $COVERAGE_DIR/coverage_data -size 0 -print0 | xargs -0 rm
echo "Combining collected coverage data under $COVERAGE_DIR/coverage_data"
$COV_EXEC combine
echo "Reporting the coverage data at $COVERAGE_DIR/coverage_data/coverage"
$COV_EXEC report --include "pyspark/*"
echo "Generating HTML files for PySpark coverage under $COVERAGE_DIR/htmlcov"
$COV_EXEC html --ignore-errors --include "pyspark/*" --directory "$COVERAGE_DIR/htmlcov"

popd
5 changes: 4 additions & 1 deletion python/run-tests.py
Original file line number Diff line number Diff line change
Expand Up @@ -38,7 +38,7 @@


from sparktestsupport import SPARK_HOME # noqa (suppress pep8 warnings)
from sparktestsupport.shellutils import which, subprocess_check_output # noqa
from sparktestsupport.shellutils import which, subprocess_check_output, run_cmd # noqa
from sparktestsupport.modules import all_modules # noqa


Expand Down Expand Up @@ -175,6 +175,9 @@ def main():

task_queue = Queue.PriorityQueue()
for python_exec in python_execs:
if "COVERAGE_PROCESS_START" in os.environ:
# Make sure if coverage is installed.
run_cmd([python_exec, "-c", "import coverage"])
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Manually tested:

./run-tests-with-coverage --python-executables=python3
Running PySpark tests. Output is in /.../spark/python/unit-tests.log
Will test against the following Python executables: ['python3']
Will test the following Python modules: ['pyspark-core', 'pyspark-ml', 'pyspark-mllib', 'pyspark-sql', 'pyspark-streaming']
Traceback (most recent call last):
  File "<string>", line 1, in <module>
ModuleNotFoundError: No module named 'foo'
[error] running python3 -c import foo ; received return code 1

python_implementation = subprocess_check_output(
[python_exec, "-c", "import platform; print(platform.python_implementation())"],
universal_newlines=True).strip()
Expand Down
21 changes: 21 additions & 0 deletions python/test_coverage/conf/spark-defaults.conf
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
#
# 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.
#

# This is used to generate PySpark coverage results. Seems there's no way to
# add a configuration when SPARK_TESTING environment variable is set because
# we will directly execute modules by python -m.
spark.python.daemon.module coverage_daemon
45 changes: 45 additions & 0 deletions python/test_coverage/coverage_daemon.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,45 @@
#
# 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 os
import imp


# This is a hack to always refer the main code rather than built zip.
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So, we will always track the coverage in the main code under python/pyspark, not the built zip Python library.

main_code_dir = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
daemon = imp.load_source("daemon", "%s/pyspark/daemon.py" % main_code_dir)

if "COVERAGE_PROCESS_START" in os.environ:
worker = imp.load_source("worker", "%s/pyspark/worker.py" % main_code_dir)

def _cov_wrapped(*args, **kwargs):
import coverage
cov = coverage.coverage(
config_file=os.environ["COVERAGE_PROCESS_START"])
cov.start()
try:
worker.main(*args, **kwargs)
finally:
cov.stop()
cov.save()
daemon.worker_main = _cov_wrapped
else:
raise RuntimeError("COVERAGE_PROCESS_START environment variable is not set, exiting.")


if __name__ == '__main__':
daemon.manager()
23 changes: 23 additions & 0 deletions python/test_coverage/sitecustomize.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
#
# 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.
#

# Note that this 'sitecustomize' module is a built-in feature in Python.
# If this module is defined, it's executed when the Python session begins.
# `coverage.process_startup()` seeks if COVERAGE_PROCESS_START environment
# variable is set or not. If set, it starts to run the coverage.
import coverage
coverage.process_startup()