And.. set-up.py for hcluster is hanging on Ubuntu 13.10

[dartdog]

It is being called from your install script... Guess I'll try an independent install Not sure what the issue is
and for the record my Ipython notebook is clobbered with can't import zmq.ipkernel (sigh)

[dlovell]

Hi dartdog,

Similar to issue #7, I'm going to point you to the VM as the only solution we have on hand. If you'd like to submit a pull request for the solution you find for an Ubuntu 13.10 install, I'll be happy to pull that into a branch for when we migrate over to 13.10 (from 12.04).

As a side note, we wrote the install_hcluster.sh script because the standard pip install would freeze.

Best
Dan

[mszep]

Hi, sorry to comment on a closed issue, but I wanted to report that I had basically the same experience as @dartdog above.

For me (on ubuntu 13.10) simply changing the libboost version to install to 1.49 worked, and for hcluster I was able to download and install it manually. I thought it might be relevant though that hcluster seems to have been incorporated in scipy (as scipy.cluster), so that might be a dependency that you can eliminate?

Perhaps a bit more off topic, the reason I opted to install manually instead of using the VM is that I wanted to play with BayesDB, which also mentions a VM install option, but there's no link to an image in the Downloads section. If that's fixed, perhaps people would be less likely to do as I did and say "screw it, I'll just do the manual install of both"...?

[jbaxter]

Hi @mszep,

Thanks! Just fixed the hcluster issue you ran into (and mentioned in the other issue).

Regarding the BayesDB VM, we removed it because many people were trying to use fairly large datasets that caused the standard VM setup to run out of memory. You'll run into the same issue with CrossCat though, unfortunately, if you try using larger datasets without much memory.