bigdataprocessor2 plugin on conda #188
Comments
|
Dear @lldelisle, Thank you very much for your interest! Let me first point out that there is code inside BDP2 for submitting processing jobs from ImageJ to a Slurm cluster: https://github.com/bigdataprocessor/bigdataprocessor2/blob/master/src/main/java/de/embl/cba/bdp2/cluster/BDP2MacroClusterSubmitterCommand.java The good thing about this is that it will process all the time points of a data set in parallel, thus potentially speeding up processing massively. We could discuss whether this also may be an option for you. Let me know. This solution relies on my image-cluster code, which comes with BDP2. However, while this is convenient I am actually not sure whether submitting and managing jobs from ImageJ is the way to go. I think it could be better to use standard tools (e.g. nextflow) to orchestrate and submit BPP2 jobs to an HPC. Thus, I would be super interested to be able to run BDP2 via conda. It would be amazing if you could help there! |
|
I am quite new to the imaging field so I may say things totally stupids. |
|
I think we should not "manually" download the jars. One can install update sites via command line. This looks interesting (we only need one update site): https://forum.image.sc/t/add-more-than-one-update-site-via-commandline/6481 Could you try? We only need the BDP2 update site.
|
|
For what it's worth, here's the full documentation of updater commands available from the command line: |
|
Thanks, I will need to give a try and come back to you. |
|
I don't know whether this helps you but within a Dockerfile this seems to work: |
|
Thanks. In fact Then to install it: |
The ... it should install all up-to-date files from all active update sites, if I understand correctly. |
|
Yes, @lldelisle, after adding all the "websites" one has to do a final "update", see last line in below example: |
|
Definitely, for a personnal usage this is perfect. Do you know where the OBSOLETE_MODIFIED come from? |
|
What I meant is that you only do this:
You got this doing what? Doing this? |
It will update all the plugins and fiji which have been updated since the installation of Fiji in the conda recipe. Then If I add BigDataProcessor with: Then I get: The goal of a conda recipe is to be able to easily reinstall a given version of each tool and plugin so I cannot do this in a bioconda recipe. That's why I need to add only the files related to BigDataProcessor.
|
I see, so you need to somehow download a fixed version of both Fiji and of the BigDataProcessor update site, correct? @haesleinhuepf @imagejan do you maybe know how to install via command line a specific version of Fiji and a specific version of an update site? |
|
Hi, |
|
The The problem is possible the outdated Fiji available from conda, maybe this can be updated? I suggest discussing this on forum.image.sc. To my knowledge, the updater is currently not really suitable to achieve the goal of getting a specific version of all Fiji components from a given point in time. If you need a reproducible Fiji installation, I'd suggest building I'll ping @ctrueden here, as I'm sure he'll be able to give better advice than me. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Dear developers,
I would like to use bigdataprocessor2 on fiji on command line on a HPC. I would like you know if you would agree to put your plugin on conda? If yes, do you want to write the recipe or help me to do so?
Here is an example of a conda recipe for a fiji plugin: https://github.com/bioconda/bioconda-recipes/tree/master/recipes/fiji-morpholibj
Thanks
The text was updated successfully, but these errors were encountered: