Modifications made to files that are located in a directory that is provided as the source argument of the submit_job method are not transmitted to the site local agent. Only the originally submitted files are available although modifications were made or new files were added to the directory.
The files are uploaded and updated correctly to a source_code folder in the EBRAINS-Drive, but the URL provided in the code field of the corresponding job on the site local agent directs to a zip file that still contains the originally uploaded files.
Even after several re-submits the files are not updated.

The idea is to provide a workflow that is as similar as possible to running a simulation locally, i.e. the code is on the local machine, results are saved to a local directory, and the log is printed to stdout.

The most recent client has an option to "wait" for the execution to complete. A nice companion option would be to have the logs collected and displayed during the execution, as they are generated. This would require the client to keep up to date with the log and only display the new parts, as I don't think it is possible to get the log in parts currently (or else the server could be updated to allow this).

Currently it says "You do not have permission to create a job", which is not very helpful for a user trying to understand what the problem is.

In nmpi_saga.JobRunner._get_input_data() self.client.download_data_url is called but the method nmpi_saga.HardwareClient.download_data_url() is not implemented.

There are a few spurious print statements in the library code, which ideally should be converted into log output, one example is

... this will be uploaded to a temporary location from where SpiNNaker/BrainScaleS can download it.

As an additional feature, we should allow the user to specify a directory, which will be tar-ed up (not including any .git or other hidden subdirectories, and uploaded.

Hi,

I am trying to use nmpi.Client("username") to login from my local mechine, but when I running this python file, I got json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0).

This error located in nmpi_user.py file, line 194, config = r_val.json(). It seems that the r_val is a text/html file, which cannot be coverted to json using json() function.

My python version is 3.8.9, and hbp_neuromorphic_platform package version is 0.10.1

It would be nice in Jupyter notebooks to have a "pretty" representation of the Job data structures.

The documentation for this is here: http://ipython.readthedocs.io/en/stable/api/generated/IPython.lib.pretty.html

(suggestion by Michael Schmuker)