Return active learning proposed placements with raw coordinate names

[tom-andersson]

Currently, the GreedyAlgorithm returns X_new_df with columns x1 and x2. See this part of the documentation, where the columns should be lat and lon:

The relevant part of the code is here:
https://github.com/tom-andersson/deepsensor/blob/e09f985a02ad79d1f5640173de6a9b2bc9587c27/deepsensor/active_learning/algorithms.py#L508

This should just require a one-line change to use self.x1_name and self.x2_name from here: https://github.com/tom-andersson/deepsensor/blob/e09f985a02ad79d1f5640173de6a9b2bc9587c27/deepsensor/active_learning/algorithms.py#L111-L112

I'll leave this open as a nice first issue for someone. Please add a unit test for this as well.

[tom-andersson]

Closed by #99