spark-instrumented-optimizer/python/run-tests
Davies Liu 4f01265f7d [SPARK-3786] [PySpark] speedup tests
This patch try to speed up tests of PySpark, re-use the SparkContext in tests.py and mllib/tests.py to reduce the overhead of create SparkContext, remove some test cases, which did not make sense. It also improve the performance of some cases, such as MergerTests and SortTests.

before this patch:

real	21m27.320s
user	4m42.967s
sys	0m17.343s

after this patch:

real	9m47.541s
user	2m12.947s
sys	0m14.543s

It almost cut the time by half.

Author: Davies Liu <davies.liu@gmail.com>

Closes #2646 from davies/tests and squashes the following commits:

c54de60 [Davies Liu] revert change about memory limit
6a2a4b0 [Davies Liu] refactor of tests, speedup 100%
2014-10-06 14:07:53 -07:00

116 lines
3.3 KiB
Bash
Executable file

#!/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.
#
# Figure out where the Spark framework is installed
FWDIR="$(cd "`dirname "$0"`"; cd ../; pwd)"
# CD into the python directory to find things on the right path
cd "$FWDIR/python"
FAILED=0
rm -f unit-tests.log
# Remove the metastore and warehouse directory created by the HiveContext tests in Spark SQL
rm -rf metastore warehouse
function run_test() {
echo "Running test: $1"
SPARK_TESTING=1 time "$FWDIR"/bin/pyspark $1 2>&1 | tee -a unit-tests.log
FAILED=$((PIPESTATUS[0]||$FAILED))
# Fail and exit on the first test failure.
if [[ $FAILED != 0 ]]; then
cat unit-tests.log | grep -v "^[0-9][0-9]*" # filter all lines starting with a number.
echo -en "\033[31m" # Red
echo "Had test failures; see logs."
echo -en "\033[0m" # No color
exit -1
fi
}
function run_core_tests() {
echo "Run core tests ..."
run_test "pyspark/rdd.py"
run_test "pyspark/context.py"
run_test "pyspark/conf.py"
PYSPARK_DOC_TEST=1 run_test "pyspark/broadcast.py"
PYSPARK_DOC_TEST=1 run_test "pyspark/accumulators.py"
PYSPARK_DOC_TEST=1 run_test "pyspark/serializers.py"
run_test "pyspark/shuffle.py"
run_test "pyspark/tests.py"
}
function run_sql_tests() {
echo "Run sql tests ..."
run_test "pyspark/sql.py"
}
function run_mllib_tests() {
echo "Run mllib tests ..."
run_test "pyspark/mllib/classification.py"
run_test "pyspark/mllib/clustering.py"
run_test "pyspark/mllib/linalg.py"
run_test "pyspark/mllib/random.py"
run_test "pyspark/mllib/recommendation.py"
run_test "pyspark/mllib/regression.py"
run_test "pyspark/mllib/stat.py"
run_test "pyspark/mllib/tree.py"
run_test "pyspark/mllib/util.py"
run_test "pyspark/mllib/tests.py"
}
echo "Running PySpark tests. Output is in python/unit-tests.log."
export PYSPARK_PYTHON="python"
# Try to test with Python 2.6, since that's the minimum version that we support:
if [ $(which python2.6) ]; then
export PYSPARK_PYTHON="python2.6"
fi
echo "Testing with Python version:"
$PYSPARK_PYTHON --version
run_core_tests
run_sql_tests
run_mllib_tests
# Try to test with PyPy
if [ $(which pypy) ]; then
export PYSPARK_PYTHON="pypy"
echo "Testing with PyPy version:"
$PYSPARK_PYTHON --version
run_core_tests
run_sql_tests
fi
if [[ $FAILED == 0 ]]; then
echo -en "\033[32m" # Green
echo "Tests passed."
echo -en "\033[0m" # No color
fi
# TODO: in the long-run, it would be nice to use a test runner like `nose`.
# The doctest fixtures are the current barrier to doing this.