fgmax data issues fg.x and fg.y are no longer increasing

[cjeffr]

I've updated my installation to the most recent version from 5.8.2 and with that some of my fgmax plotting routines failed. I found that fg.x and fg.y are no longer increasing but are the same value across the array. I'm wondering if I'm missing an input change somewhere? I did not remember reading changes to fgmax in the release notes but I did see completed pull requests where changes were made recently. My file moriches_fgmax.data was created using version 5.8.2. Thanks for your help.

Example arrays below with fg.x I can also paste or share screen shots of fg.y if needed. My fgmax declaration is below and the example I have as well.

INPUT:

rundata.fgmax_data.num_fgmax_val = 2 fgmax_grids = rundata.fgmax_data.fgmax_grids

fg = fgmax_tools.FGmaxGrid() fg.point_style = 4 fg.min_level_check = 6 fg.tstart_max = days2seconds(0.5) # just before wave arrives fg.tend_max = days2seconds(1.333) # when to stop monitoring max values fg.dt_check = 5.*3600 # how often to update max values fg.interp_method = 0
fg.xy_fname = '../../fgmax_data/moriches_fgmax.data' # file of 0/1 values in tt3 format fgmax_grids.append(fg) # written to fgmax_grids.data

EXAMPLE OUTPUT

fg.x == [-72.87995372 -72.87995372 -72.87995372 ... -72.87995372 -72.87995372 -72.87995372] fg.X == array([[-72.87995372, -72.87986113, -72.87976853, ..., -72.5802315 , -72.5801389 , -72.58004631], [-72.87995372, -72.87986113, -72.87976853, ..., -72.5802315 , -72.5801389 , -72.58004631], [-72.87995372, -72.87986113, -72.87976853, ..., -72.5802315 , -72.5801389 , -72.58004631], ..., [-72.87995372, -72.87986113, -72.87976853, ..., -72.5802315 , -72.5801389 , -72.58004631], [-72.87995372, -72.87986113, -72.87976853, ..., -72.5802315 , -72.5801389 , -72.58004631], [-72.87995372, -72.87986113, -72.87976853, ..., -72.5802315 , -72.5801389 , -72.58004631]])

Since fg.x = fg.X[:,0] I tried swapping the rows and columns fg.X[0,:] and I get my expected result. array([-72.87995372, -72.87986113, -72.87976853, ..., -72.5802315 , -72.5801389 , -72.58004631])

[rjleveque]

@cjeffr: This is probably a result of the changes in #561, which was included in v5.9.1.

I think this change should have fixed a bug in earlier versions where fg.x was improperly set to fg.X[0,:] rather than fg.X[:,0]. Is it possible that your Python script that calls fgmax_tools.read_output() was taking that bug into account and resetting fg.x after reading? I ask because I had some scripts like that, which I had to modify after fixing the bug.

I just tried running the example in $CLAW/geoclaw/examples/tsunami/chile2010_fgmax-fgout and it seems to run ok, and adding

    print(fg.X[:3, :3])
    print('x = ',fg.x[:3])

at the end of the function plot_fgmax_grid in plot_fgmax.py and then running this script prints out

[[-119.91666667 -119.91666667 -119.91666667]
 [-119.75       -119.75       -119.75      ]
 [-119.58333334 -119.58333334 -119.58333334]]
x =  [-119.91666667 -119.75       -119.58333334]

as expected.

Hope this helps you debug. Let us know if you continue to have problems.

[cjeffr]

Yes I think my plotting routine didn't take into account the indexing added into read_output() I will try that and if it goes well I'll close this issue.

Thanks!