Magnetic coordinates and lines drawing across the world map

[cmeeren]

I'm extending basemap to work with magnetic coordinates. Basically, latitude longitude are converted to/from magnetic coordinates during calls to Basemap.__call__(). However, I get results like this:

It seems that coastlines/polygons are not properly "broken" on the left/right plot edges. I suspect I have to change something in _readboundarydata() to fix this (since that's where the coastlines and polygons are initiated), and I have had a look at it, but I do not really see what I would need to change. I don't even see how this is handled in the normal (geographic coordinates) case. There are some references to the crossing of east/west datelines, are these "datelines" the left/right map boundaries in projections like the above? I don't really understand what this part of the code accomplishes.

Any tips?

[WeatherGod]

Just for completeness, which version of matplotlib and which version of basemap are you using?

The code dealing with wrappings is insane (I didn't write it), and I am having difficulty figuring it out. There is a lot of magic numbers involved. One thing I might speculate as the source of the problem is exactly what range of longitudes does your code produce? Is it -180 to 180? or is it 0 to 360? Also, is that range half-open or inclusive?

On Wed, Oct 28, 2015 at 9:53 AM, Christer van der Meeren < [email protected]> wrote:

I'm extending basemap to work with magnetic coordinates. Basically, latitude longitude are converted to/from magnetic coordinates during calls to Basemap.call(). However, I get results like this:

[image: figure_1] https://cloud.githubusercontent.com/assets/7766733/10790251/8446e6ee-7d82-11e5-847b-1838642da17a.png

It seems that coastlines/polygons are not properly "broken" on the left/right plot edges. I suspect I have to change something in _readboundarydata() to fix this (since that's where the coastlines and polygons are initiated), and I have had a look at it, but I do not really see what I would need to change. I don't even see how this is handled in the normal (geographic coordinates) case. There are some references to the crossing of east/west datelines, are these "datelines" the left/right map boundaries in projections like the above? I don't really understand what this part of the code accomplishes.

Any tips?

— Reply to this email directly or view it on GitHub https://github.com/matplotlib/basemap/issues/222.

[cmeeren]

matplotlib 1.4.2 basemap 1.0.7

I haven't really paid attention to what the longitude range is (I'm wrapping other code for the actual conversions), but I don't think that is relevant. Everything must be converted to map projection coordinates before plotting, and no matter if the longitude is -181 or 179 it should be converted to the exact same point. Right? Or is there some code that relies on the actual latitudes/longitudes? Seems strange if so, because whether or not a line is wrapped around depends only on the projection coordinates (alternatively on the longitude AND the specific projection of the map).

Sorry if I'm wrong here.

[WeatherGod]

... unless there is an impedance mismatch of some sort. Looking over the basemap code, I suspect that the code assumes [-180, 180), but is actually miscoded and mishandles the 180.0 edgecase.

On Wed, Oct 28, 2015 at 11:20 AM, Christer van der Meeren < [email protected]> wrote:

matplotlib 1.4.2 basemap 1.0.7

I haven't really paid attention to what the longitude range is, but I don't think that is relevant. Everything must be converted to map projection coordinates before plotting, and no matter if the longitude is -181 or 179 it should be converted to the exact same point. Right?

— Reply to this email directly or view it on GitHub https://github.com/matplotlib/basemap/issues/222#issuecomment-151879501.

[cmeeren]

I tried constraining the output longitude to [-180, 180) in the conversion functions but it had no effect.