Use bioblend for local install? about bioblend

[nuwang]

Hi,
Could you describe your usage scenario a bit more? As things stand now, it would be pretty difficult, because it lies somewhat outside of bioblend's stated functionality. At the moment, bioblend doesn't deploy Galaxy. It launches a CloudMan image which contains a pre-installed Galaxy. Cloudman in turn connects up the required indices/data for Galaxy and starts up Galaxy. These Galaxy indices and data are fairly large (multi-gigabyte).

On the other hand, running a local Galaxy without indices/data would be very easy to do by hand, but I guess, of limited utility without the data and indices.

Perhaps the best way to install Galaxy the way you want, is to use CloudBioLinux (https://github.com/chapmanb/cloudbiolinux). Would that suit your requirement?

[nuwang]

I should add, the CloudMan image containing pre-installed Galaxy is also built using CloudBioLinux (CBL). However, CBL can also be used to deploy Galaxy etc. on a local machine.

[jjv5]

Hi,

Thanks for your quick reply. We find local installation of Galaxy to
be very painful. The base install itself is not so bad, but adding ftp
access, apache proxy and so forth leaves a lot of room for error and
takes a lot of time.

On the other hand, starting a cloud based galaxy instance is pure
bliss. Other than the fact that all the indices and data are included
by default (and thus large) it is ideal.

What we would really like is a way to install Galaxy locally in a
manner something like this:

from bioblend.cloudman import CloudManConfig
from bioblend.cloudman import CloudManInstance
from bioblend.cloudman import CloudManInstance
from bioblend.galaxy import GalaxyInstance

cfg = CloudManConfigLocal('myCloud')

gi = GalaxyInstance('', key='your API key')
gi.InstallGenomes('human')
gi.InstallGenomes('mouse')
gi.InstallData('blastnr')
gi.InstallTools( [blast,cufflinks,tophat] )

We have talked to many people who find it takes a very large amount of
IT effort to get a Galaxy instance running locally and to keep it
running properly with many users. We are in this same boat.

Jim

On Wed, Mar 20, 2013 at 3:50 AM, nuwang [email protected] wrote:

I should add, the CloudMan image containing pre-installed Galaxy is also built using CloudBioLinux (CBL). However, CBL can also be used to deploy on a local machine.

—
Reply to this email directly or view it on GitHub.Hi,

[nuwang]

Hi Jim,
Thanks for the details and I understand your usage scenario now. The good news is that Galaxy does expose a way to install tools via a Toolshed API, which is a brand spanking new REST API in Galaxy. This has not had time to make its way into BioBlend yet, but we hope to integrate this functionality into BioBlend as soon as we can spare the time, but unfortunately there's no timeline at the moment. Code contributions would also be welcome, since many of the contributors are tied up with Cloudman/GVL work.

As far as I know, the ability to install genomes into the Galaxy Index is not available, in the way you describe. Therefore, I think this would have to be a feature request in Galaxy first, after which it could be integrated into BioBlend. The idea sounds great to me though. Could you lodge an issue in Galaxy? The code sample you provided would also be useful.

[jjv5]

I have started to try a local install with biolinux. It's a little obtuse
at first, but I think it's all there.

The tool installation through an API will be super. I think that combined
with a fab file install of galaxy genomes would make half the pain go away.

On Wed, Mar 20, 2013 at 7:48 AM, nuwang [email protected] wrote:

Hi Jim,
Thanks for the details and I understand your usage scenario now. The good
news is that Galaxy does expose a way to install tools via a Toolshed API,
which is a brand spanking new REST API in Galaxy. This has not had time to
make its way into BioBlend yet, but we hope to integrate this functionality
into BioBlend as soon as we can spare the time, but unfortunately there's
no timeline at the moment. Code contributions would also be welcome, since
many of the contributors are tied up with Cloudman/GVL work.

As far as I know, the ability to install genomes into the Galaxy Index is
not available, in the way you describe. Therefore, I think this would have
to be a feature request in Galaxy first, after which it could be integrated
into BioBlend. The idea sounds great to me though. Could you lodge an issue
in Galaxy? The code sample you provided would also be useful.

—
Reply to this email directly or view it on GitHubhttps://github.com/afgane/bioblend/issues/16#issuecomment-15170767
.

[nuwang]

Apologies, it appears that there is indeed a new Genome API which allows the functionality you mentioned! (Thanks Clare). So bioblend will need to be updated with it. I'll leave this issue open.

[nuwang]

I've just added experimental support for the Genome API, but the install_genome method needs a lot more testing. Some of the params in the wrapped Galaxy REST API seem to have outdated docstrings, and I've logged an issue on trello.

However, the toolshed API may take a longer time. At first glance, I couldn't find a way to install tools, but it may well be there.

[afgane]

Galaxy now has an Ansible playbook for installing and configuring itself: https://github.com/galaxyproject/ansible-galaxy