Benchmark script improvements

[domanchi]

Summary

Building off #181, this adds the following features:

• machine consumable output, by default. (--pretty preserves the human readable table)
• you can select which plugins you want to test
• timeout functionality, so scripts won't run forever
• customizing number of iterations, so you can get more accurate results

Example Output

$ python scripts/benchmark.py --plugin PrivateKeyDetector --plugin Base64HighEntropyString -n 5 | jq
Running performance tests on: PrivateKeyDetector, Base64HighEntropyString
for: ['~/pg/detect-secrets']
{
    "Base64HighEntropyString": 0.8232977999992727,
    "PrivateKeyDetector": 0.6896803999989061
}

$ python scripts/benchmark.py --plugin Base64HighEntropyString --harakiri 0.5 --pretty
Running performance tests on: Base64HighEntropyString
for: ['~/pg/detect-secrets']
------------------------------------------
plugin                              time
------------------------------------------
Base64HighEntropyString   Timeout exceeded!
------------------------------------------

[KevinHock]

lgtm 🚢 ⛵️
(when Travis passes)

[KevinHock]

lol

[domanchi]

FWIW, I copied uwsgi's syntax for this: https://uwsgi-docs.readthedocs.io/en/latest/Glossary.html

[KevinHock]

This seems like the way to do that

[domanchi]

Whoops. Forgot to remove this lol.

Yeah, we're not breaking new ground here. We've done this before: https://github.com/Yelp/detect-secrets/blob/master/detect_secrets/core/usage.py#L396

[KevinHock]

Super nit: variable names could maybe help readability here