[Benchmarking] Vector tests: first oddity

Alan Boudreault aboudreault at mapgears.com
Tue Aug 24 16:35:02 EDT 2010


I've committed the changes in svn. Here's new feature of the script:

* -filter_within <filename> option: The random bbox has to be within the 
<filename> dataset. Note that I've committed the spain-boundary-simplified.shp 
in svn too.

* The script now writes file(s) in the current directory rather than using 
stdout. The filenames are the projection codes. Would you like an option like 
-name to specify the filename of the files? That way you could get something 
like  spain_4326.csv rather than 4326.csv. Of course, it is easy to simply 
rename the files, depends on how many time you use that script.

* -srs: You can specify the input projection of your parameters (ie. -region). 
Default is 4326.

* -srs2: The second projection you want the bbox. This implies a second file 
to be writed. Basically, we transform the initial bbox to get the bbox into 
the second projection. Then, we compute the new width/height or the map to 
preserve square pixels and the same number of pixels in both cases.

Since the gdal and oracle paths are not in the system ld paths, here's the 
command you'll have to use to run the script with the new options (use the 
python2.6 command rather than python):

$LD_LIBRARY_PATH python2.6 wms_request.py -count 200 -region -6 38 -1 43 -
minsize 640 480 -maxsize 1024 768 -minres 2.5e-06 -maxres 0.000755 -srs2 

Let me know if you have any questions or needs.


On August 24, 2010 03:17:34 pm Daniel Morissette wrote:
> Andrea Aime wrote:
> > Mind it would have to be a different script, one that takes a csv
> > and converts its contents to another projection.
> > Or the same script that writes down two files in two projections.
> >
> > Well, let me know what you plan to do :-)
> It will be a single script that writes to two files in a single pass.
> Alan is on it (almost done) and will let us know once the new script is
> available.
> Daniel

Alan Boudreault

More information about the Benchmarking mailing list