Improve testing infrastructure to verify the output

[dandhlee]

Googlers see b/206698957.

Is your feature request related to a problem? Please describe.
Currently the testing infrastructure using tox -e docs does not verify the output produced.

Describe the solution you'd like
Whether using nox, tox or other forms of testing tools, have the test verify the output as well so we're not making changes that would be a breaking change.

Describe alternatives you've considered
For now, we're okay to move on as-is to ensure that the generation runs. We should revisit this ASAP to verify the content it outputs.

Additional context
N/A