Optimize create_timeseries for memory

[jdbcode]

Potentially helps issues with "User memory limit exceeded" errors like that described in this discussion post.

• Adds parallelScale argument to ee.ImageCollection.reduce
• Uses clipToFeatures when region is a feature collection
• reduces the list of images in a collection to 1 for check if size is 0 (size calls toList, which can be memory intensive with large collection)

I wasn't sure whether to use **kwargs for the new parallel_scale parameter for the function that call on create_timeseries - I went with named parameter to make it more obvious that it is an option.

[giswqs]

Great improvement! Adding the new parameter parallel_scale is more user-friendly than using **kwargs

[github-actions]

🚀 Deployed on https://652d8a4257f6b90d36ca6b6e--opengeos.netlify.app