L2 ENA Maps: Bin arrays based on matched indices

[nkerman]

Algorithm Description:

Once data value indices have been matched between two frames (e.g. from an L1C DPS frame to the HAE frame in which the L2 map will be projected), a function must "bin" from the first input frame to the second output frame. "Binning" here means taking values from the indices in frame 1, and assigning them to the matched index in frame 2.

Requirements:

• There may be a non- 1:1 correspondence between the indices in frame 1 and the indices in frame 2. That is, multiple indices in one frame may map onto the same index in another frame.
• Not all possible indices in each frame are guaranteed to be used. This can mean that not all input values are projected into the output frame. It can also mean that not all output frame indices receive any values from the input frame.

Algorithm Code Information:

Input data:

• input frame indices (ordered list)
• output frame indices (ordered list)
• input values to be binned
  • Allow multiple values to be binned on the same indices*
  • Must be at least as long on the axis to be indexed as input frame indices
  • May have an "extra axis" - e.g. a non-spatially-binned dimension like energy. Each bin of this axis will be treated independently.
• Shape of output frame grid, or the number of elements in it, or some other way of determining how large the output value array must be.
  • Must be at least as large as the output frame indices

* Suggest 2 functions: one to bin a single array, and another to loop through a dictionary of:
{"array_name": NDarray} and bin each.

Related Issues/PRs:

#1258
#1260