PROGRAM [edit section]

dataload.py

DESCRIPTION [edit section]

dataload.py is a python program that is used to put the data from benchmark test results into json files, for use by the Jenkins flot plugin module.

It is called by bench_processing during post_test for a benchmark test.

outline [edit section]

The program has the following rough outline:

• read the reference.log file to get the metric names
• parse plot.data into a python structure
• write data into the <testname>.<metric>.json file
• write data into the <testname>.info.json file

ENVIRONMENT and ARGUMENTS [edit section]

parser.py takes takes positional arguments on its command line.

• $1 - $JOB_NAME - the name of the test (ex: 'Benchmark.Dhrystone')
• $2 - $DATA_FILE - the name containing the plot data (<testname>/plot.data)
• $3 - $REF_FILE - the name of the reference log (<testname>/reference.log)

dataload.py is called with the following invocation, from bench_processing:

  run_python $PYTHON_ARGS $FUEGO_CORE/engine/scripts/parser/dataload.py $JOB_NAME $DATA_FILE $REF_FILE $TESTDIR

OUTPUT [edit section]

This program outputs the files:

• <testname>.<metric>.json (ex: Benchmark.Dhrystone.Dhrystone.json)
• <testname>.info.json (ex: Benchmark.Dhryston.info.json)

SOURCE [edit section]

Located in /home/jenkins/fuego/engine/scripts/parser/dataload.py.

SEE ALSO [edit section]

• parser.py, bench_processing, reference.log