integration of KNN approach

[leslieDLcy]

To integrate the data-driven KNN approach into the calibration capability

[leslieDLcy]

1. Plan of integration:

• an abstract KNN calibration approach for black-box models;
• an abstracttmcmc workflow for Bayesian model updating

2. Please find below for the instructions_on_fork.pdf to do pull request of this integration.

3. Docstrings are recommended such that users can follow the usage better. Generally, the Google style docstring is used throughout the package. But I have utilised "myst" to enrich the docstring, see for a template of how docstring will be rendered. Please find the source code of that function on Github repo to get the template.

[leslieDLcy]

@Roberock Many thanks for the latest commits, which are fantastic! Only two small issues:

1. There are two KNNCalibrator class defined, one in knn.py while the other in calibration.py. From the example, I can guess the one in knn.py is taking precedence.
2. utils.py module seems to be missing, which results in import exceptions. as seen here.

[Roberock]

@leslieDLcy

thank you:

1. I removed KNNCalibrator from calibration.py (to avoid duplicates). It now contains only abstract class Calibrator & a pointer for the future class MCMCCalibrator
2. utils.py has been removed and now __init__.py updated accordingly (empty now).

[leslieDLcy]

@Roberock I can successfully run the example code for KNN to generate the figures below. This means that the functionality is there. But it seems the two problems mentioned above are still there. Or, maybe you solved them but have not done a pull request for the latest changes yet?

[Roberock]

Hi @leslieDLcy,
I thought I pushed.

[leslieDLcy]

Ciao Roberto. I've integrated your latest pull request. Many thanks for the great work. Just for future reference, it works the best when you do pull request next time, choose the calibration branch as the destination rather than main. I can change it on my side as well, so it is not big deal, just for your reference.

[leslieDLcy]

@Roberock FYI, a few updates.

1. new notebook example added ==>

• TMCMC
• KNN

2. Associated API reference added:

• TMCMC
• KNN

Let me know how you make of it. Any comments/suggestions/criticism is appreciated!