pyTerra 0.9

History

This version, 0.9, brings pyTerra up-to-date with a number of TerraServer changes that have happened in the past couple of years that I have not kept up with. The API has moved around a little bit to avoid confusion with the TerraServer USA commercial offering, and now that quad tree-based image caches are all the rage, TerraServer’s SOAP API seems quite quaint. It does offer some unique things like metadata and a gazetteer which might make it useful in other contexts. I had used pyTerra as the basis for avTerra, which was an ArcView 3.x extension for fetching TerraServer imagery. I suspect that pyTerra’s lack of maintenance has meant avTerra is in disarray, but with this update, there’s a possibility of it being brought back to life.

The previous versions of this code were PSF-licensed, but that seems kind of silly. 0.9+ is now MIT, with the licensing text included in the source release as it should be.

Otherwise, the effort to bring this codebase to modernity was mostly one of vanity, though I still have one or two things that still could use it. Looking at the code, it’s hard to believe that it is nearly nine years old…

Changelog

0.9 brings a number of changes to pyTerra. First, SOAPpy has been removed in exchange for suds. Suds is much nicer than SOAPpy for simple tasks, and it handles the Microsoft WSDL like a champ. The movement to Suds simplified the internals quite a bit, and it should be much more straightforward to follow.

I have also modernized the setup.py and based it on distribute, with dependencies of PIL and suds declared. Additionally, the unit tests, which had to be run manually are now available via a simple python setup.py test invocation.

Finally, the external API has been updated in a number of ways. These include things like returning datetime instances where appropriate, PNG support, etc. Having not heard of anyone using pyTerra in a number of years, I doubt the API changes I have made will have much impact.

Usage

pyTerra now contains two modules – api and image. These were formerly a messy module structure of TerraImage.TerraImage and pyTerra.pyTerra names.

from pyTerra import image

class Object:
    pass

lg_ul = Object()
lg_ul.X = 433714.25
lg_ul.Y = 4661043.80
lg_ul.Zone = 15

lg_lr = Object()
lg_lr.X = 438603.35
lg_lr.Y = 4656591.96
lg_lr.Zone = 15

scale = 'Scale2m'
theme = 'Ortho'

img = image.TerraImage(ul, lr, scale, theme, lr.Zone, "/tmp")
t = img.download() # <-- PIL.Image instance you can do what you need with