Extract gzip data from transfer-encoded WARC

[ianmilligan1]

Describe the bug
As reported in our Slack (and apologies in advance if I butcher the technical description here, this is at the edge of my knowledge).

Some WARC files keep compressed payload (as opposed to transfer-encoding), so we will have WARCs that are text/html but have compressed gzip data. Even though the content is gzipped within the WARC, as it is text/html it would be ideal to be able to provide the information.

Ilya Kreymer suggested on Slack:

probably the 'best' way to read this is, if a warc record has transfer-encoding: chunked, attempt to dechunk, if failed, assume not chunked. if a warc record has content-encoding: gzip, attempt to ungzip, otherwise assume not gzipped. the warcio library provides a method to read with both of these checks

To Reproduce
Load a WARC with this form of content, extract content, and see gzip'ed material instead of the text.

Expected behavior
If the webpage has text, it would be good to see the text not the compressed material.

[ruebot]

This will be resolved with #533. A build test was added to account for it, and it passes.