Add helpers to extract crawl metrics / data verification

[motin]

Currently after each crawl, we run data verification using a rather manual process, requiring quite a lot of notebook copying/cloning.

Ideally, it should be enough to run something like crawl_metrics(s3_bucket, crawl_directory) or similar to get relevant metrics, including those from https://github.com/citp/openwpm-data-release/blob/master/Crawl-Data-Metrics.ipynb and those in the notebook linked in openwpm/openwpm-crawler#30 (comment).

A companion crawl_metrics_summary(crawl_metrics) method could be included to print out the most relevant metrics in human-readable form.

Use cases:

• Include at the top of every crawl-analysis notebook to understand the nature of the gathered crawl dataset
• To easily set up notebooks that analyses notebook crawl datasets longitudinally and/or compares individual crawl datasets
• Include in OpenWPM CI to spot regressions in crawl performance/health (related: https://github.com/mozilla/OpenWPM/issues/479)