Advanced tutorials introduced

[vineetbansal]

Advanced tutorials from https://github.com/pathpy/2026-polnets-tutorials are introduced in docs/tutorials/advanced, and tested in the CI.

All notebooks are working locally for me except the model selection one - 02-04-model-selection.ipynb, because it uses some featuers not yet merged in main or dev.

Specifically, the notebooks were meant to be run against the trp-model-lselection branch. I'm not sure about the status of this branch and whether it is ready to be merged to main or dev.

I've excluded that notebook in this PR for now, and included all the remaining ones. We can discuss whether it belongs in this PR.

This partially addresses issue #331. Once CI passes then the notebooks/links can be moved around so that the current tutorials go under "getting started" and these "advanced" tutorials become the default ones.

TODO: Ping @IngoScholtes about the 02-04 notebook that doesn't work.

[M-Lampert]

Thank you for migrating all of this!
I saw that you added some new datasets to the repository. I think two of them are identical to datasets that we already have in docs/data (at least their names are identical).
For the existing tutorial notebooks, I tried to avoid using files that are saved in the repository because that will lead to problems if you run the notebook directly in Colab (see #129).

I know that the Sociopatterns high school dataset can be downloaded from Netzschleuder and we already do this in an existing tutorial notebook here. I also know that the temporal_clusters.tedges is a synthetic dataset for which it would anyway be nice to have a PathpyG function that can generate it. Maybe @IngoScholtes can provide the code for this and we could add it to algorithms.generative_models to generate it directly so we could remove this file dependency as well.
The small test graph files could be generated like here.
The other datasets can probably also be found online somewhere or maybe we could replace them with datasets that can be found if it's not used for something dataset-specific. For the others that are specifically needed and where I couldn't find the dataset easily downloadable online, I added an optional download from the repo like this.

[vineetbansal]

@M-Lampert - I've added a utility function pp.io.example_data that uses the local path to docs/data (if available, i.e. when running after a git clone), or downloads the data from github (if not available, i.e. running from Colab etc) - much like what you were doing, but now centralized in a single function.

Both the existing notebooks and the "advanced" notebooks use it now. The data folder for both have been merged into one (there was duplication, as you observed).

So I'd say this PR is ready to be merged pending the decision on the one notebook - [02-04-model-selection.ipynb](https://github.com/pathpy/2026-polnets-tutorials/blob/main/02-04-model-selection.ipynb) - whether it needs to be included or not.