Don't ignore rustdoc when generating comparison summaries

[jyn514]

Fixes #980.

[rylev]

This looks good to me, but I'd like @Mark-Simulacrum's perspective on why doc test cases were filtered out in the first place (as that decision predates my involvement in the project).

[Mark-Simulacrum]

Yeah, this is fine with me as well. I think when we just added the python triage script doc builds were still pretty recently added, and I recall numerous comments from @jyn514 on regressions that they should basically be ignored. I think those reasons no longer hold, though, and based on the perf runs I've seen doc builds don't behave particularly unusually. So I think we can proceed here.