cms benchmark analysis KEYWORD --files=FILES --tags=TAGS [--plot]

[laszewsk]

The by tests contain aconvenient mechanism to print csv data. To do a proper benchmark the pytests must be rerun multiple times. Each time we put the data into a new file.

Then we can just use this command to extract the data and generate a temporray pot to see how this data looks in a boxplot.

cms benchmark analysis KEYWORD --files=FILES --tags=TAGS [--plot]

could be the start for this command

however as the csv data also includes a tag we could even make multiple box plots at the sma time. The way this works is to just concatenate the file, fgrep through tem and looking for KEYWORD and TAGS.

cms benchmark analysis KEYWORD --file=name-[1-4].txt --tags=[aws,google,azure,chameleon] --plot

would then for the keyword put a boxplot for each cloud into the same diagram.

That is relatively easy to do due toe the "invention" of the #csv tag lines and using the Parameters.expand that we use a lot from cloudmesh

[laszewsk]

This can be developed without any cloud as this can be even used on the local benchmarks