Put project and spider in joblogs path

[wvengen]

In #45 joblogs could be stored on container/object storage. This is currently configured with a bucket name, and the jobid is used for the filename. scrapyd uses project and spider name in the logs path (e.g. see here), let's do that here as well. Is also useful in practice.

As joblogs is still in beta phase, we can change the path without the need for backward compatibility.