Ideas for tutorials for the NNPDF school

[scarlehoff]

Let's track ideas for tutorials in this issue.

Some general points (please add any others that you might think of!)

1. Make the pipeline easy to use. As mentioned in today's PC, users are lazy. We should embrace that.
2. The tutorials should end with a few-replicas fit but we don't need to explain the fit. We can just prepare pre-trained fits (I'll take care of that) so that 5 minutes after creating the new data etc the students will se a PDF that looks a bit different because of what they did with the pipeline.
3. All ekos should also already be prepared. Since potentially there will be student trying to run with very silly computers, let's sacrifice precision and use small ekos (and also small fktables of course).

(rather than having long discussions here, I'd rather treat this like a wiki, if you see something that can be improved in someone's idea just modify the comment, if it is more fundamental it can be discussed of course).

Please feel free to modify the tutorials, specially to add more information, expand some of the steps, etc.

I don't know how much time do we have of course so we might need to prioritize!

[scarlehoff]

Standard tutorial

Using the theories and runcards already included, simply install, run and fit.

1. Install pineline (this is here but maybe it is not complete)
2. Install nnpdf
3. Run a madgraph and a yadism dataset to have two different output examples.
4. Create the fktable out of that (this is here)
5. Put the files in the right folder (for pineko part, this is here)
6. Run the fit

[scarlehoff]

Modified theory parameters

Using the runcards already included, fit a PDF using a different value of some parameter (for instance, a different charm mass).
We assume the installation already happened for everyone at #2 (comment)

1. Create a new theory by copying an already existing one and modify it at will.
2. Rerun the desired dataset & create fktable.
3. Put files in the right folder <----- we actually need a bridge script pinefarm -> nnpdf, I volunteer for that
4. Run the fit.

[scarlehoff]

Standard new dataset

Everything standard model and standard parameters, but the dataset is not already included. I would do this two times, one for DIS and one for a Madgraph process (pp -> H is flashy and quick, so that's a good one!)

1. Create the new madgraph runcard
2. Run dataset and create fktable <--- maybe these two should be done together automagically since it will happen a lot...
3. Create the new yadism runcard
4. Run dataset and create fktable
5. Put files in right folder.
6. Do fit

[scarlehoff]

Exotic Madgraph

I believe this is something people will find interesting, although it requires (at least for me) learning how to run exotic models in Madgraph!

1. Create the new Madgraph runcard with exotic stuff (not sure what does this entail ????)
2. Modify the theory runcard to account for the exotic stuff
3. Run dataset and create fktable
4. Files in right folder
5. Fit