PyHealth icon indicating copy to clipboard operation
PyHealth copied to clipboard

Added a custom KG Radiology Dataset for DL4H

Open jatin-tahiliani opened this issue 1 month ago • 0 comments

Who I am: Jatin Tahiliani ([email protected])

Type of contribution: New dataset

Description of work: The dataset used for testing this loader is publicly available in my project repository: https://github.com/jatin-tahiliani/DL4H_research_project In this pull request, I added a custom radiology dataset loader under pyhealth/datasets/ and added test case under pyhealth/tests/ validating dataset loading

Files to review: pyhealth/datasets/kg_dataset.py pyhealth/tests/core/test_kg_dataset.py

How to access the csv dataset used in this research: https://github.com/jatin-tahiliani/DL4H_research_project/blob/main/mimic_reports.csv

Functionality:

  • Loads a CSV file with columns: report_id, report
  • Automatically parses records into PyHealth BaseDataset format
  • Supports splitting using PyHealth's splitter utilities

jatin-tahiliani avatar Dec 06 '25 23:12 jatin-tahiliani