Filter out sdist archives that are not of interest
Some PyPI origins declare sdist
archives in formats that are not supported by swh.core.tarball.uncompress,
for instance .deb, .egg, .rpm or .whl. This leads to errors when trying to load such origins and generates a lot of noise in
sentry reports.
As the content of those archives does not match standard sdist
format (there is no PKG-INFO
file for instance),
we should filter them out from the artifacts to process for a PyPI origin.
Migrated from T3575 (view on Phabricator)
Edited by Phabricator Migration user