@aaron-kaplan this is an FYI in case you want to check on some things.

@xchourio I know I made the case for small PRs, but when starting up, it's always a bit difficult.

This is about setting up a ccsr section of the python maprooms with for now one maproom (flex_fcst) that does IRI ELR flex fcst Maprooms. Namely exactly this:

https://iridl.ldeo.columbia.edu/maproom/Global/Forecasts/NMME_Seasonal_Forecasts/precip_full.html

It's using the work done in enacts/flex_fcst and preserves the case to use the app for a pycpt forecast. And it's using pepsico maproom/layout utilities.

So most of the work was to:
-- adapt enacts/flex_fcst to pepsico utils
-- adapt it to take either pycpt or ELR forecast

There's default and IRI config to run the ELR Maproom and a test-pycpt config to check that it still works with that. Then pycpt.py and iri_fd_elr.py contain the specific of reading and computing for these 2 cases. flex_fcast_calc.py is used by maproom.py to interface between the 2 cases. Everything else comes from either enacts/flex_fcst or pepsico, and adapted if needed.

There is no rush on this work so feel free to take your time to review.

Known problems / next steps to resolve in smaller incremental PRs:

• poe graph is actually pone graph
• there is a dry mask to apply to the map
• there is spatial Gaussian smoothing to apply to the map
• there is no option to overlay contours of historical percentile values (not sure pingrid can do contours, this may have to be dropped)
• Block controls that belong together with a button to fire them
• Didn't test mapping poe of a physical threshold
• While the maps I checked against Ingrid Maproom looks the same (except for smoothing and dry mask), physcial values of graphs seem to be smaller in python than ingrid
• There's an alarm (red contour) on the Y control
• Print e.g. 103W instead od -103E
• Review Text
• give units to the local graphs