-
Notifications
You must be signed in to change notification settings - Fork 222
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Save datasets in Galaxy history from jupyterlab notebook #1157
Conversation
What do you think about using a collection with |
@@ -0,0 +1,80 @@ | |||
<tool id="run_jupyter_job" name="Run long running jupyterlab job" version="0.0.1"> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@anuprulez can you check if there exists a hidden=true option here?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please also add a recent profile=21.05
@@ -0,0 +1,80 @@ | |||
<tool id="run_jupyter_job" name="Run long running jupyterlab job" version="0.0.1"> | |||
<description>on GPUs</description> | |||
<requirements> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I thought this needs to run in a Docker container? We should make this explicit, as we don't want to run this in conda, for security reasons?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ping @anuprulez
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This dynamic code execution is the same as executing code written in any other interactive tool. The execution command from this tool can be sent to any secluded VM for running as far as I understand.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, but this tool is way more insecure than any other tool, left? So making sure it runs in Docker and not by accident in Conda is making it a little bit more secure?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I agree, having a docker container running this tool will automatically enable it to run in a secluded environment.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Moving to an explicit container requirement does not in any way guarantee this is safe. The right way right now is to make it an interactive tool. The medium term solution is a tag that means we require containerized execution and pick a destination that can run it (so effectively an interactive tool without open ports).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Current tool is supposed to be executed on jupyterlab notebook that is already an interactive tool. I am not sure if interactive tools interact with one another?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
No problem, they're just tools.
<data format="h5" name="outfile_output_arrays" label="Saved arrays"></data> | ||
</outputs> | ||
<tests> | ||
<test> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
can you a test for num expected outputs?
<output name="outfile_output_model" file="scikit-script-model.onnx" ftype="onnx" compare="sim_size" delta="50" /> | ||
</test> | ||
<test> | ||
<param name="select_file" value="tf-script.py"/> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why not combining this test with the first one?
</output> | ||
</test> | ||
<test> | ||
<param name="select_file" value="scikit-script.py"/> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
and this test with the second?
Steps to run this tool (after this is merged to Galaxy EU):
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nice, thanks a lot @anuprulez
This PR adds a Galaxy tool for processing long-running jobs emitted from a script in the Jupyterlab notebook such as created here. This tool will be useful for running scripts as a job such as training a machine/deep learning model that takes a long time to finish. The trained model (as ONNX format) will be available in Galaxy history after the job finishes.
Steps to run this tool:
Currently, this work is in progress, tools tests are not yet written. Your comments are welcome!! Thanks!