Added DOWNLOAD_CLOUD functionality, reworked the parameter file, smaller fixes

[tylerapritchard]

Two Major Features:

1. Reworked the configuration parameter to be fully astropy configuration namespace compatible. Added a convenience function to write a default config file. Updated the sample notebook with configuration file and cache directory examples.
2. Added functionality related to the DOWNLOAD_CLOUD configuration parameter. If False, this will return the location of MAST cloud bucket location for the file instead of downloading it directly. The intended use case for this is cloud-based compute platforms like TIKE where we would rather remote-read a file that is on the cloud than download it directly. Added tests for this as well.

This also includes a number of smaller fixes and updates, including:

• documentation changes - added citation and community contribution guidelines (just linking to the relevant Lightkurve documentation pages).
• fixed an issue in the getitem where trying to mask via a pandas series returned a datafram and not a lksearch object
• the MASTSearch.cloud_uris parameter was breaking for TESSSearch, since pandas.concat adds nans to missing columns when the TESScut results are added in. I've fixed the immediate case and added a regression test, but am concerned that these nans may show up again elsewhere. Is this brittle? Should we redo something? This is a separate issue.
• A bunch of the configuration parameters were being treated a little inconsistently, I streamlined these

[d-giles]

Minor: I believe this is redundant, as pd.Series has the attribute __iter__

[d-giles]

Variable changed to config.get_cache_dir()