Follow up to https://bugzilla.mozilla.org/show_bug.cgi?id=1742674.

After we have an understanding of what PDF viewers are commonly set to default, we need to trim down to a set of observed viewers. This will largely follow what already exists for default browsers.

Romain: would you like to take a crack at preparing the top N list from the WDBA default PDF handler data (per looker)? It's clear some of these need to be grouped. Perhaps we keep groups that amount to >1% of the pings and see how many that is?

While we're here, it would be helpful to trim whitespace from the raw PDF handler string.

Also while we're here, perhaps we could track down why we're getting null at least 10% of the time (for Firefox 105, so we should be fishing the PDF default).

(In reply to Nick Alexander :nalexander [he/him] from comment #3)

Also while we're here, perhaps we could track down why we're getting null at least 10% of the time (for Firefox 105, so we should be fishing the PDF default).

I believe we chased this down to an error when inspecting the data - search included FF versions prior to the telemetry being added.

Worth noting that I did a lot of the relevant labour in this query; we could implement that normalize_pdf_default and be done here.

The "Other" category in that query is still fairly large, probably worth a round of finding trends in it.