Run 3 (2022) Analysis Production Quickstart

Created for https://indico.cern.ch/event/1237092/

N.B. This tutorial will eventually be moved to https://lhcb-ap.docs.cern.ch/ and maintained there. Look forward to it!

You can use the following terminal commands to set up a new working directory for your Analysis Production based on this example template.

Just log in to LXPLUS and edit+copy the below lines into the terminal

# clone the AnalysisProductions data package (https://gitlab.cern.ch/lhcb-datapkg/AnalysisProductions/) and cd to it
git clone https://:@gitlab.cern.ch:8443/lhcb-datapkg/AnalysisProductions.git
cd AnalysisProductions

# create a new branch of your own
git fetch
git checkout -b $(whoami)-run3-example origin/master

# create a new production based on the example template
git checkout origin/roneil-run3-emtf-example -- EMTF_Run3Example

# Please choose a useful name that describes the purpose e.g. "MyWG_Run3Monitoring"
git mv EMTF_Run3Example !!YOUR_UNIQUE_PRODUCTION_NAME!!

# create an initial commit for your production and publish the branch
git commit -am "Starting my first run3 production"
git push -u origin $(whoami)-run3-example

To ensure that the test job runs correctly, edit the entrypoint: fields in info.yaml, replacing EMTF_Run3Example with the name of your production folder!

From here, it is assumed that any changes you make to your production will be committed and git pushed as needed.

What's next?

Open a Draft Merge Request

Your production may not be ready yet, but whatever stage your production is at, we encourage you to open a Draft merge request! https://gitlab.cern.ch/lhcb-datapkg/AnalysisProductions/

You should be able to see a new Continuous Integration (CI) pipeline has started, including a link to the Analysis Productions website. You can also look on this page to find your test pipeline.

Configure your own Run 3 tuples

Replace or modify dv_simple.py to produce your own tuples with DaVinci/v63r5. You could refer to the following resources:

• https://gitlab.cern.ch/lhcb/DaVinci/-/tree/v63r5/DaVinciExamples/python/DaVinciExamples/tupling
• https://lhcb-davinci.docs.cern.ch/configuration/davinci_configuration.html

DaVinci options: local development and testing (LbAPLocal now recommended)

You can easily run local DaVinci tests as you develop your tupling options:

cd AnalysisProductions
lb-ap test YOUR_PRODUCTION_NAME Lcp_Data_Simple_Tuple

Need to debug issues in around the environment? Try this command instead:

lb-ap debug YOUR_PRODUCTION_NAME Lcp_Data_Simple_Tuple

Having issues with DaVinci? ~DPA-WP3 Offline Analysis Tools on Mattermost may be able to help.

Edit info.yaml to your needs

1. Change the inform: field to your CERN email
2. Change the wg: field to your working group name
3. Ensure that the entrypoint: field in each job is correct!
4. Add some more checks, or change existing checks.

Analysis Productions documentation

Submit your production

If you are happy with the state of your tuples and the tests, you should think about submitting your Run3 production. Before you do:

• Please, don't just submit the template! We want to avoid submitting many productions of a similar or exact configuration.
• Does your tuple contain all the variables you or your WG are interested in?
• Are there any other decays you would like to look at?

If you are happy with the above points, assign your WG liaison and mark the merge request as ready!

N.B. As of 16 Jan 2023, there were only two runs from 2022 on the bookkeeping. There is a lot more by now, and as more runs are added, productions will need to be re-submitted to ensure everything available is processed. For real data taking (i.e. after commissioning) it will be possible to submit concurrent productions that process data in parallel to LHCb running.