Package index • RHealth

Dataset

Classes for representing EHR datasets and configuration.

BaseDataset: BaseDataset — R6 infrastructure for clinical event datasets

SampleDataset(): SampleDataset

MIMIC3Dataset(<initialize>): MIMIC3Dataset: Dataset class for MIMIC-III

MIMIC4EHRDataset(<initialize>): MIMIC4EHRDataset: Dataset class for MIMIC-IV EHR

MIMIC4NoteDataset(<initialize>): MIMIC4NoteDataset: Dataset class for MIMIC-IV Clinical Notes

Patient: Patient: R6 Class for a Sequence of Events

Event: Event: R6 Class for a Single Clinical Event

DatasetConfig: DatasetConfig: Root dataset configuration

load_yaml_config(): Load and validate dataset configuration from YAML

get_dataloader(): Create DataLoader

split_by_patient(): split_by_patient

split_by_visit(): split_by_visit

split_by_sample(): Dataset Split Functions

Processors

Preprocessors for handling input/output features.

Processor: Abstract Processor Base Classes

DatasetProcessor: DatasetProcessor: Processor applied to entire datasets

SampleProcessor: SampleProcessor: Processor for sample-level transformations

FeatureProcessor: FeatureProcessor: Base class for all data processors

TextProcessor: Text Processor

SequenceProcessor: Sequence Processor

TimeseriesProcessor: Time Series Processor

BinaryLabelProcessor: Binary Label Processor

MultiClassLabelProcessor: Multi-Class Label Processor

MultiLabelProcessor: Multi-Label Processor

RegressionLabelProcessor: Regression Label Processor

RawProcessor: Raw Processor

get_processor(): Get Processor Class (Hardcoded Version)

Tasks

Benchmark prediction tasks based on MIMIC-IV.

BaseTask: BaseTask (Abstract Base Class)

Readmission30DaysMIMIC4: Readmission30DaysMIMIC4 Task

InHospitalMortalityMIMIC4: InHospitalMortalityMIMIC4 Task

Models

Neural network models and layers built on torch.

BaseModel(): BaseModel Class

EmbeddingModel(): EmbeddingModel

RNN(): RNN Model Class

RNNLayer(): RNNLayer Class

Training & Evaluation

Trainer class and supporting functions for training deep learning models.

Trainer: Generic Trainer for torch models

collate_fn_dict_with_padding(): Collate Function with Padding

get_metrics_fn(): Get Metrics Function

binary_metrics_fn(): Binary Classification Metrics (Python‐style API)

is_best(): Check if Score is Best

set_logger(): Initialize Logger

ece_confidence_binary(): Expected Calibration Error for Binary Classification

MedCode

Tools in the MedCode Module.

atc_convert(): Truncate an ATC Code to a Specified Level

get_ancestors(): Get Ancestor Codes in the Hierarchy

get_ddi(): Load the Drug–Drug Interaction (DDI) Table for ATC Codes

get_descendants(): Get Descendant Codes in the Hierarchy

lookup_code(): Lookup a Medical Code Entry

map_code(): Map a Code from One System to Another

supported_cross(): List Supported Crosswalk Code Systems

supported_inner(): List Supported Medical Code Systems

Utilities

Other tools.

create_directory(): Create Directory if Not Exists

.csv2parquet_path(): Given a *.csv(.gz) path, return *.parquet path in a /subset folder

Event-from_list: from_list: Create Event from row

JoinConfig: JoinConfig: Configuration for joining tables in a dataset

rhealth.config: Dataset Configuration

TableConfig: TableConfig: Configuration for a single table