Open to PRs?

[tarensanders]

Hi there,

Thanks for this package. Several of my colleagues have used it in the past and found it really useful, and I'm glad the code is open-sourced since the website had to be taken down.

I had a bit of trouble getting the code running locally (mostly due to dependency issues), but I got it to work in this fork. I just wanted to see if you were open to pull requests to save others this trouble.

I can give more details in a PR, but broadly the issues I fixed were:

• There are dependency mismatches between setup.py and the conda environment (see also Building wheel for scipy (setup.py) ... error while installing on Windows 10 #2 (comment))
• I rewrote the dockerfile to make it 1) faster to build and 2) use more modern versions
• Changed the README instructions to connect to localhost port 5000, not 5050.

[renespijker]

Hi taren. i;ve been trying to get it to work but encountered similar problems. would you be willing to share your solutions with me?

[tarensanders]

Hi @renespijker, try using the docker image in my fork. There were some mismatches between the conda environment and what was specified in the setup.py script, and I tidied those in my fork. See the instructions on running the docker image in the README.

Note that I was wrong about the port number (it should be 5050).

Hope that helps.

Edit: If you don't want to go the docker route, you could also try this.

[renespijker]

thanks, this worked. I've got it running on an intel mac for now. Next seeing if I can get it running on an ARM mac (need to get some things sorted for tensorflow for that