latest sleepqa

[joyan2]

Contributor: Jeffrey Yan
Contribution Type: Full Pipeline Replication (Dataset + Task + Model + Ablation)

Description:
Implements the SleepQA dataset and maps it to an Extractive Question Answering task using a regression schema for character-level span prediction. It also introduces a SleepQABioBERT reader model and provides a full pipeline ablation study comparing a specialized medical reader against a general-purpose reader.

Files to Review:

• pyhealth/datasets/sleepqa.py - Core dataset and indexing logic.
• pyhealth/datasets/configs/sleepqa.yaml - Table configuration.
• pyhealth/tasks/sleepqa_extractive_qa.py - Task definition and processor mapping.
• pyhealth/models/sleepqa_biobert.py - BioBERT model implementation.
• examples/sleepqa_extractive_pipeline_biobert.py - Full pipeline ablation study.
• tests/core/test_sleepqa.py - Synthetic data test suite.
• pyhealth/datasets/__init__.py
• pyhealth/tasks/__init__.py
• pyhealth/models/__init__.py
• docs/api/datasets/pyhealth.datasets.sleepqa.rst
• docs/api/tasks/pyhealth.tasks.sleepqa_extractive_qa.rst
• docs/api/models/pyhealth.models.sleepqa_biobert.rst
• docs/api/datasets.rst - TOC update.
• docs/api/tasks.rst - TOC update.
• docs/api/models.rst - TOC update.