Enso example development

[sana-ccrc]

Changes made to ENSO_Forecast and ENSO_Pipeline notebooks:

• Added a script that prepares a dataframe ready for machine learning, including target variables needed for 1-month lag forecasts with XGBoost and 3-month lag forecasts with MLP
• Added a script that encodes 'month' into cyclic features; this is useful if we need to include month as a predictor, using 'month_sin' and 'month_cos'
• minor: edited subheadings and cleaned comments

Issue: I am using pet/2025.06, and there is no plotly library

[coveralls]

Pull Request Test Coverage Report for Build 17001481280

Details

• 1 of 13 (7.69%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.3%) to 60.744%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
packages/pipeline/src/pyearthtools/pipeline/operations/xarray/normalisation.py	1	13	7.69%

Totals
Change from base Build 16385043506:	-0.3%
Covered Lines:	9955
Relevant Lines:	15896

---

💛 - Coveralls

[tennlee]

Thanks, I've had a read through of this. Let's have a chat early next week about making the ML objectives jump out just a little more, set up the XGB model and work up a good test/train/validate split etc.

[sana-ccrc]

Sounds good. Are you available on Tuesday 29 July anytime between 12 and 5pm?

[sana-ccrc]

Added more background
Preprocessed the data further, now it is ready for XGBoost training

[sana-ccrc]

Computed monthly climatology of Nino3.4, which can be used as a benchmark to evaluate our model.

[tennlee]

(note - the unit tests have started failing. This is completely unrelated to this change, and pertains to how the version is set in the nci site archive subpackage, and I think it's a change in strictness in the tooling. I will fix it separately, ignore it)

[tennlee]

@sana-ccrc Thanks so much for the work you did here. I have merged this pull request into PyEarthTools. There were duplicate copies of the notebooks which I removed, feel free to confirm I did the right thing and I can restore the deleted notebooks easily if needed. I will also go ahead and link these tutorials into the documentation so that they can be found easily, and I'll tweak the gridded MLP notebook as well as we discussed.

In addition, consider whether you would like to be listed in the .zenodo.json for the repository.

When we release a new version of PyEarthTools, that version is archived on Zenodo. See: https://doi.org/10.5281/zenodo.15760768

As you have contributed to PyEarthtools, would you like to be listed on Zenodo as an author the next time PyEarthTools is archived?

If so, please open a new pull request. In that pull request please add your details to .zenodo.json (which can be found in the PyEarthTools root directory).

In .zenodo.json, please add your details at the bottom of the “creators” section. The fields you will need to complete are:

1. “orcid”. This is an optional field. If you don’t have an ORCID, but would like one, you can obtain one here: https://info.orcid.org/researchers/ .
2. “affiliation”. Options include: the institution you are affiliated with, “Independent Researcher” or “Independent Contributor”.
3. “name”. Format: surname, given name(s).