Zhenbang Wu

Core structure updates to support image and text data. I will elaborate more on the changes in tonight's meeting. Framework: - Re-design base and sample dataset to handle multi-modality data...

Add data caching

4

Add caching supports for - BaseDataset (caching dataframe) - SampleDataset (caching list of dictionaries)

In https://github.com/sunlabuiuc/PyHealth/blob/9305a6f6a7f338ce3c5a64d6a3cf596f87116b6c/pyhealth/processors/timeseries_processor.py#L12, we should allow the user to normalize the timeseries based on statistics computed on the training set.

Fix sample dataset indexing

2

We should probably add a check. But, it's mostly so we don't break compatibility with legacy code like split_by_visit() in https://github.com/sunlabuiuc/PyHealth/blob/master/pyhealth/datasets/splitter.py _Originally posted by @jhnwu3 in https://github.com/sunlabuiuc/PyHealth/pull/335#discussion_r2059140897_

Update document to reflect the changes in v2.0

3

1. Update the doctoring in python files 2. Update the GitHub README 3. Update the readthedoc web