Provide full support for unicode in filenames

This PR fixes support for unicode characters in submitted dataset filenames under the python2.7. This includes pulling in a fix to oar-metadata's jq/urldecode.jq script which establishes the names of dataset files. Supporting unicode on Python 2.7 is tricky since unicode is not the default string type. A key part for the fix is updating the pdrtest docker image to set python's default encoding to 'UTF-8'; this will also have to be applied at the oar-docker level. A number of tests (but not all) were updated to ensure support for unicode is working.

Also fixed in this PR:

updating jq to version 1.7 (instead of building from source)
fixing a typo in the python-source GitHub Action config file that prevented proper building

usnistgov / oar-pdr

Provide full support for unicode in filenames #340