Re: Help rendering image with manifold + mapnik combo

Dane Springmeyer wrote:

Blank maps usually indicate a mis-match in coordinates used for the  
bounding box or the use of a coordinate system that does not work with  
the data being used.

I suspected as much.  I think the problem stems from the exported reprojection of the data out of manifold as I was trying to force a generic mercator projection (in manifold) into the google mercator (srid900913) on the postgres side.  ... I have had complete success since I started using the unprojected latlong export to postgres, and do the reprojection in mapnik.  Now that I know that works, I can try reprojecting the data on the server after export, but havent had time to get around to it yet.

You can download nik2img.py however and run that  
against a Mapnik XML file in verbose mode, which outputs information  
on your map bounding box and which data layers intersect it. Seeing  
that the bbox does not match the layer coordinates is often helpful  
for noticing problems with your usage of coordinate systems.

Excellent.  I will give nik2img a try.  I've also just build mapnik on the mac and am about to delve into that today.   Lots of good stuff to test today.

How did you come up with that? (referring to my environment latlong 0,0,.375.375)

Nothing special there...that's just where our environment was initially built.  We needed a generic fictional island near the equator, and the original designer started at 0,0 and built up and out.

Using which projection? If you are using mercator you will need to  
change your bounding box to fix mercator coordinates.

When I tried using the mercator projection natively from the database, I generated the bounding box coordinates by forwarding them thru provided latlong coordinates in the python script, but obviously something was not quite right in my source data to begin with.

And you'll need to make sure that your data is appropriate for mercator,
since that projection does not allow, for example, coordinates on the earths  
surface greater than 90 degrees north/south.

Since its a small island on the equator, I didnt think that would be an issue.  But, again...something must have gotten mucked up in the reprojection, so I'll try some other methods.

Thanks for your tips.  I'm halfway to my goal and have some good leads to move me closer.

-will