Optimize for alternate storage backends

[damianmoore]

At the digital agency I work for we use Amazon's S3 for our image storage in production. We found your optimization post-processor and it worked well in development. However, when we went live we noticed the images were no longer being optimised.

The Django package we use is called django-storages: http://django-storages.readthedocs.org/en/latest/index.html

After digging into the post_processor.py I found that thumbnail.path is not set for cloud-based storage providers and raises a NotImplementedError. In this case I detect the file-type from the url attribute and save the file referenced by the name attribute.

I confirm this is working with S3 but it should be generic enough to work with most other storage backends.

[SmileyChris]

Cheers for this, I think that we should add a test showing what's going on here.
Rather than just using os.path.splitext, how about using mimetypes.guess_type and splitting on /, taking the subtype. And we don't need to do it against .url -- .name should be enough