Updates/geometry fix

[abradley60]

• Correcting the STAC geometries by using the valid data within a geotiff. A polygon is created that represents the minimum rotated rectangle of the valid data. Note, more points are added along the rectangle edges to ensure reprojections to other coordinate systems work appropriately.
• Removing old geometry logic relying on the CDSE geometries, as these are also incorrect
• Adding a test that is specific to the metadata creation function
• Note, future work will clean the tests for consistency

[caitlinadams]

If my read of this logic is correct, the geometry will be created for the last tif that's found when iterating through polarisation files / incidence angle files in the directory. For dual pol, are we confident that the geocoding is always the same? I'm guessing we are, but just want to check that it's valid to get the geometry from either polarisaiton.