Skip to content

cellistigs/neurocaas_interface

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

NeuroCAAS Local Interface for efficient data transfer.

Info for Joao:

General Testing Stack Structure

The interface can be tested at the bucket cianalysispermastack. I have added your website enabled credentials (jcouto) to the group traviscipermagroup, allowing you to upload datasets in the corresponding bucket.

The bucket has the following structure:

s3://cianalysispermastack
|-traviscipermagroup 
  |-inputs
    |-dataset1.json
    |-dataset2.json
  |-configs
    |-config_None_None.json
    |-config.json
    |-...
  |-submissions
    |-submit.json
    |-condition_0_0submit.json
    |-...
  |-results
    |-job__cianalysispermastack_23:05:47.219155condition_0_2/
      |-logs/
      |-process_results/
        |-end.txt/

We can run many different test cases in this stack, but I will focus on the simplest one for now.

Paths to monitor

In the simplest case, we can trigger a job by uploading a submit file that looks like this:

{
"dataname":"traviscipermagroup/inputs/dataset1.json",
"configname": "traviscipermagroup/configs/config_None_None.json",
"timestamp": "value_text"}

Here the dataname and configname fields refer to data and config files that exist in the inputs and configs directories. The interesting part is the timestamp field: this dictates a location in the results directory where job results will be written. In particular, if timestamp has a value "value_text", then the job results will be written to "s3://cianalysispermastack/traviscipermagroup/results/job__cianalysispermastack_value_text/". I usually monitor for this location to be created to start showing users logs (which are updated live) and results. To monitor for the end of the job, I have configured analyses available through the website to create a subdirectory "process_results", and to write a file end.txt when the job is finished.

In brief, I monitor for the directory {bucketname}/{groupname}/job__{bucketname}_{timestamp} to start showing users job data, and file {bucketname}/{groupname}/job__{bucketname}_{timestamp}/process_results/end.txt to signal job completion.

You can find an example submit file that should work in this repo at ./files/condition_test_interface_submit.json

About

Local interface for batch data transfer with NeuroCAAS.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors