[matplotlib-devel] [basemap] hammer projection and bug ?

Community matplotlib-users
1

Nathanaël:

Can't reproduce it (I'm using geos 3.2.2 as well).  Can you try

updating from svn and see if the problem persists?

You'll get hammer that way too.

-Jeff
···

On 1/18/11 6:43 AM, Nathanaël Schaeffer wrote:

  If it can help, here is a screenshot of the resulting

figure of the previous script :

  [http://www.zimagez.com/zimage/basemapbug.php](http://www.zimagez.com/zimage/basemapbug.php)

  Interestingly, when hoovering the mouse over the figure the

displayed x values start at 0 on the beginning of the coastlines,
not on the drawn boudary maps, and ends at 4.83 at the other end.

    On Tue, Jan 18, 2011 at 2:29 PM, > Nathanaël Schaeffer <nathanael.schaeffer@...287...> >         wrote:

2011/1/18 Jeff Whitaker <jswhit@…146…>

            On 1/18/11 3:33 AM, Nathanaël Schaeffer > > > wrote:

Dear matplotlib developpers,

              I use matplotlib for several years and I'm very

satisfied with the it.

              I started using the basemap package a few days ago,

and I noticed something that looks like a bug :

              With the Mollweide projection (and others too), when

specifying rsphere=1.0, the coastlines is not drawn in
the

              left part of the plot. With rsphere=2.0 the hidden

part is smaller, and with rsphere=10.0 it is not
visible.

              I'm using matplotlib 1.0.0 and basemap 1.0

              To reproduce the bug :



              m = Basemap(projection='moll',lon_0=180,rsphere=1.0)

              m.drawcoastlines(linewidth=0.5, color='grey')     #

draw discrete coastlines

              m.drawmapboundary()        # draw a line around the

map region

              show()

          Nathanaël:  Your test script works fine for me (no

coastlines are missing). What version of the geos library
did you link against?

          geos 3.2.2



          Concerning your previous message, sure r=1 is small, but I

would have thought that it is a simple scaling, and with
double precision floating point numbers, this should not
be a problem. In physics, it is common to set unit radius
for the sphere. Anyway, this is not a very important bug,
as setting a larger value works well.