add script to parse the output from multiple runs

[btraven00]

• take backend and timestamp from output folder
• account for multiple seeds, optionally
• produce an unified dataframe

[btraven00]

done in here #6