Use Case #115
Use case ATLAS: ATLAS002
This issue describes ATLAS use case defined as ATLAS002
ATLAS user analysis pipeline tests on experimental particle physics using augmented open data (source codes http://opendata.atlas.cern/software/)
The ability to run user final data analysis using datasets hosted in the datalake.
- Run the current AOD analysis examples requesting data from the datalake.
- Allow the execution of the examples over the multiplied data to simulate longer analysis pipelines.
- Modify such analysis to write outputs (ROOT files) that can be store back into the datalake.
- We can also:
- Create a volatile JupyterHub that can be used for multiple-user usage of the datalake.
- CERN OpenStack resources are a feasible option.
- Try to integrate those exercises with volunteer computer services like MyBinder and Google Colab (?)
- Container and/or another way to use the RUCIO CLI and Datalake authentication.
- At least one Jupyter-kind interface with the RUVIO extension to use the analysis notebooks
- The proper storage of the datasets in the datalake, e.i. The ATLAS 001 above.
- Correct replication of the analysis examples in the OSSR for a complete analysis pipeline simulation at the user level.
Arturo Sánchez Pineda
WP2, WP3 (for software stored in a record attached to the Zenodo ESCAPE community).
The analysis examples will be correctly be executed in several instances.
Things to test
Localisation, retrieval and usage of the data in several user-kind particle physics analysis examples.
- To show the ability of the datalake to deliver the data to several analysis instances (VMs and personal computers) and store the final products back to the datalake.
- Test different interfaces that will be the standard way of interaction for many final users.