kahst
Dockerfile.client and Dockerfile.server
Created new Dockerfiles for a server and a client to assist in having a local dockerized server and local dockerized client.
This would assist those hoping to test or run these without mucking up their Python environment and should give handy examples hoping to extend the environment.
BTW -- I love what y'all have already provided.
Great, thanks for this! Can you please check, there's a duplicate line in the server file ("Install required Python packages -- either librosa or numpy"). Could you also please add a note or quick "How-to" to the README explaining the two files? Thanks!
Also, can you please take a look at #16, seems like both PRs have the same goal but slightly different attempts.
I cleaned up that bad copy and paste then updated the README to include the client and server. I also reviewed #16 but I didn't integrate that change as it ends up making the dockerfile larger in an attempt to make it single purpose. The dockerfile running just the client.py doesn't need bottle or tensorflow.
I used Dockerfile.server and Dockerfile.client since docker expects a directory to only have one Dockerfile in it. After you build, the specified Dockerfile.X doesn't have any negative impact and the facility is included in docker for situations just like this.
Alright, I think that's ok considering your comments. Would you mind updating the README in a way that the command line arguments only appear once, instead of repeating them for each use case? Also, if you adjust headlines, can you please also adjust shortcut links in the "Contents" section? Thanks.
I like the exact versioning, but the split of the Dockerfile makes it overly complicated. We probably just add bottle to the Dockerfile to support server.py as well and take the puny overhead, which this change introduces to the analyzer use with Docker.
