Static Layers open items

[gmgunter]

For organizational purposes, here's a checklist of unresolved items identified in #144 that we may want to revisit in the future:

• [ ] Add a command line option to specify a log file. (link)
• [ ] Load only the subset of the input DEM raster that covers the output geo grid, instead of loading the entire raster into memory. (link)
• [ ] Add a runconfig option to allow users to specify parameters of the radar grid directly, instead of inferring them from the geo grid. (link)
• [ ] Add an option to decimate the radar grid for purposes of computing the layover/shadow mask. The full radar grid should still be used to compute each of the other layers. (link)
• [ ] If the output geo grid is in geodetic coordinates, and therefore posting.x and posting.y are in units of decimal degrees, consider approximately converting them to meters for purposes of forming the granule ID. Otherwise, they are very likely to be rounded to zero in the granule ID string. (link)
• [ ] Consider changing the way that processing_type is specified in the runconfig. The intent is to distinguish between production jobs and custom processing jobs. Currently, we use the mnemonic 'PR' for production jobs and 'OD' ('on-demand') for custom jobs. However, this is confusing because the NISAR On-Demand system will be used as the production system for processing Static Layers jobs. (link)
• [ ] Since the product writer copies the 'units' attributes from the product spec XML, the units of xCoordinates and yCoordinates are always 'meters', even if the output grid was in geodetic coordinates. We should make this logic smarter to correctly infer the units based on the output EPSG code. (link)
• [ ] Consider using set_get_geo_info() from h5_prep.py to populate the datasets in the output HDF5 file related to the geo grid (xCoordinates, yCoordinates, projection, etc) in order to avoid redundant logic in ISCE3. (link)
• [ ] Determine how best to populate zeroDopplerStartTime and zeroDopplerEndTime in the identification group of the output product. Currently, these are populated using the start & end time of the radar grid, but this might be impacted if we change the way the radar grid is defined (for example, if a new numberOfLooks layer is introduced). (link)