User Guide Overview

This section provides comprehensive documentation on how to use Phentrieve's various features and interfaces.

Available Interfaces

Phentrieve offers multiple ways to interact with the system:

1. Command-Line Interface (CLI): The primary way to use Phentrieve's features, including text processing, querying, benchmarking, and more.

2. API: A FastAPI-based service that exposes Phentrieve's functionality through RESTful endpoints.

3. Web Frontend: A Vue.js-based web interface for user-friendly interaction with the system.

Core Functionality

Phentrieve provides several key functionalities:

HPO Term Retrieval

Map clinical text to relevant HPO terms using multilingual embedding models. This works across languages without requiring translation.

Text Processing

Process clinical text with advanced features: - Semantic chunking - Assertion detection (negated, normal, uncertain) - Evidence aggregation - Confidence scoring

Benchmarking

Evaluate and compare the performance of different embedding models for HPO term retrieval.

Ontology Similarity

Calculate semantic similarity between HPO terms using the ontology graph structure.

Section Contents

This User Guide section contains the following pages:

• CLI Usage: Detailed guide on using the Phentrieve command-line interface
• Text Processing Guide: How to process clinical text to extract HPO terms
• Benchmarking Guide: Running and interpreting benchmarks
• Configuration Profiles: Configuring Phentrieve for different use cases
• API Usage: Using the Phentrieve API
• Frontend Usage: Using the Phentrieve web interface