User Prasad G Thoppil reports this. He's drawing a shade plot where the length of one axis is way longer than the other, and gets a set of error messages. Doing this with half the number of points on the long axis works fine
The message is written by a routine in pipedviewerpq.py.
yes? let npts = 595390
yes? def axis/x=1:`npts`:1 xaxsst
!-> def axis/x=1:595390:1 xaxsst
yes? def axis/e=1:15:1 eaxsst
yes? let model_compress = 1 + 0*x[gx=xaxsst] + 0*_e[ge=eaxsst]
yes? save/clobber/file=xxx.nc model_compress
LISTing to file xxx.nc
yes? cancel var model_compress
yes? use xxx
yes? shade model_compress
**ERROR <type 'exceptions.ValueError'>: height of the rectangle (0.0) in not positive
**ERROR <type 'exceptions.ValueError'>: height of the rectangle (0.0) in not positive
**ERROR <type 'exceptions.ValueError'>: height of the rectangle (0.0) in not positive
**ERROR <type 'exceptions.ValueError'>: height of the rectangle (0.0) in not positive
...
(also typo in the message itself, should be "is not positive").
The color in the plot is also truncated. This is something that happens in Ferret too. The whole box should be filled in with color, but there's just a bar of it across the top. Maybe this is what's being reported in the python code - each cell in the shade plot is way too skinny to draw.
We could work on this in the common PuyFerret/Ferret code and have some kind of striding happen in an extreme case like this, or coaching to do striding. this plot all works fine.
User Prasad G Thoppil reports this. He's drawing a shade plot where the length of one axis is way longer than the other, and gets a set of error messages. Doing this with half the number of points on the long axis works fine
The message is written by a routine in pipedviewerpq.py.
(also typo in the message itself, should be "is not positive").
The color in the plot is also truncated. This is something that happens in Ferret too. The whole box should be filled in with color, but there's just a bar of it across the top. Maybe this is what's being reported in the python code - each cell in the shade plot is way too skinny to draw.
We could work on this in the common PuyFerret/Ferret code and have some kind of striding happen in an extreme case like this, or coaching to do striding. this plot all works fine.
yes? shade model_compress[x=1:595390:10]