Various fixes to the eclipse notebook

[ayshih]

• Refactored away the dependency on photutils
• Fixed the x/y confusion and polished some of the plotting
• Extracted the GPS altitude from the EXIF metadata
• Fixed the conflation of the lunar disk and the solar disk

The angular offset of Regulus matches better, but is still not great. However, we don't know how much lens distortion there is

[nabobalis]

Do we want to add a comment saying the offset left over is likely to be distortion from the lens?