.coveragerc

[run]
branch = true
parallel = true
data_file = ${COVERAGE_DIR}/coverage_data/coverage

[html]
ignore_errors = true

run-python-tests-with-coverage.sh

#!/usr/bin/env bash

# Note: case sensitivity can break things here. The path comparision in
# coverage is case-sensitive, so you need to be careful to use a properly
# normalized SPARK_HOME path if you're running on a case-insensitive filesystem
if [ -z "$SPARK_HOME" ]; then
    echo "ERROR: SPARK_HOME must point to a Spark source checkout."
    exit 1
fi

# Ensure that coverage is installed for all of the Python versions that we're
# going to test against:
#
if [ $(which python2.6) ]; then
    python2.6 -c "import coverage" || { echo 'must install coverage for python2.6' ; exit 1; }
fi

# TODO: include other python versions.  We should really refactor the run-tests
# script to be paramterized by the Python version in order to make it easier to
# run all of the tests against just one version.


export COVERAGE_DIR="$(cd "`dirname "$0"`"; pwd)"
mkdir -p "$COVERAGE_DIR/coverage_data"
# So our siteconfig.py is included:
export PYTHONPATH="$(cd "`dirname "$0"`"; pwd):$PYTHONPATH"
export COVERAGE_PROCESS_START="$(cd "`dirname "$0"`"; pwd)/.coveragerc"

$SPARK_HOME/python/run-tests

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

coverage combine
coverage report --include "$SPARK_HOME/python/pyspark/*"
coverage html -i --include "$SPARK_HOME/python/pyspark/*"

sitecustomize.py

import coverage
coverage.process_startup()