Add example notebook for Kaggle Chest X-Ray Pneumonia dataset

[rz-coder]

Note

Who you are:

rzhou19

What type of contribution you are making:

New example / use case of PyHealth.

A high level description of what you’ve done:

This notebook demonstrates a complete end-to-end workflow for training a deep learning model on the Chest X-Ray Pneumonia dataset using the PyHealth framework inside Kaggle

At a high level, the notebook:

1. Loads and explores the Kaggle dataset
2. Builds a PyHealth-compatible dataset
3. Prepares train/val/test datasets for deep learning
4. Trains a pretrained ResNet18 model using PyHealth’s Trainer
5. Evaluates the model performance

A quick rundown of what files to look at to test the implementation

• To preview the implementation, you can download thsi commit and view locally for better format.

• To test the implementation, you can either:

  1. Run this notebook on Kaggle using the following dataset as input:
     https://www.kaggle.com/datasets/paultimothymooney/chest-xray-pneumonia/code
  2. Download the dataset and run this notebook locally on your own machine.

[EricSchrock]

Disclaimer: I am not a CS 598 TA, just a PyHealth reviewer. This is some initial high-level feedback to gauge your interest in actually getting your contribution merged into PyHealth.

Looks like this is a dataset we don't have in PyHealth yet and the way you're using PyHealth looks about right at first glance. If you're interested, I don't think it would be that hard to get this PR to a mergeable state and I'd be happy to provide further feedback and support if you're interested in making that happen. At a high level, here's the main items that need to be addressed.

1. Pull your dataset and task classes out of the example notebook
2. Add proper documentation (doc strings and .rst file updates)
3. Add unit tests

See PR #392 if you'd like an example of what is needed for adding a chest x-ray dataset and related tasks.