Port parts of do_extract_abs_sa1.R to Python

[truth-quark]

Requirements

• Allow a subset by study period/duration (a list of dates, or start/end dates)
• Extract NetCDF data from the NetCDF with different layers
• Generate graphs is optional, but useful for seeing the data
• Needs a "user friendly" interface - how?
• Are command line tools "user friendly" enough?

Idea is to demonstrate Python can interoperate with the 20 variable NetCDF file.

Solution Options:

• Avoid reinventing the wheel. Do cmd line tools like ncks have functionality to extract required data subsets?
• Try a simpler front end tool to generate command line commands? (QGIS?)

[ivanhanigan]

@truth-quark yes this is the best place to work on new requirements:

The user will want to specify:

1. place(s) e.g.. ACT, Tasmania etc
2. times e.g. 2019-2020
3. flags to use e.g. "smoke_2SD","active_fires_50000","dust_merra_2_p95"

Possibly you want to break this up to extraction step and then analysis step?

[truth-quark]

@ivanhanigan do you know if using command line tools (e.g. ncks) is considered user friendly enough for the target data users?

[truth-quark]

The user will want to specify:

1. place(s) e.g.. ACT, Tasmania etc
2. times e.g. 2019-2020
3. flags to use e.g. "smoke_2SD","active_fires_50000","dust_merra_2_p95"

The ncks tool contains a hyperslabbing cmd line option to subset multiple dimensions in a single command. From the user guide, subsets of cell coords & times can be specified, as well as selecting which variables to extract.

I didn't see an option to specify a bounding box to clip. If a user has some known bounds, these could be converted to pixel coords & hyperslab to the area of interest. It wasn't clear in the user guide, but it may be possible to specify coordinate system bounds with some cmd line options.