HypoTestX Documentation

Natural Language Hypothesis Testing — Powered by LLMs or Pure Regex

HypoTestX lets you ask a plain-English question and a DataFrame and returns a full structured HypoResult — statistic, p-value, effect size, confidence interval, and human-readable interpretation. Zero mandatory dependencies, pure Python math core.

import hypotestx as hx
import pandas as pd

df = pd.read_csv("survey.csv")
result = hx.analyze(df, "Do males earn more than females?")
print(result.summary())

Getting Started

• Quick Start
  • Installation
  • Your First Test
  • Reading the HypoResult
  • Using a Real LLM Backend
  • Direct API
• Installation
  • Requirements
  • Basic Installation
  • Optional Extras
  • Installing from Source
  • Ollama (Local LLM Backend)
  • HuggingFace Local Inference
• Why HypoTestX?
  • The Problem
  • When to Use the Direct API Instead

User Guide

• Natural Language Interface
  • Signature
  • Supported Question Patterns
  • How the Regex Fallback Works
  • Using LLM Backends for Better Routing
  • The routing_confidence Field
  • verbose=True Output
• Direct API
  • Parametric Tests
  • Non-parametric Tests
  • Categorical Tests
  • Correlation Tests
• LLM Backends
  • Backend Summary
  • Common kwargs
  • Code Examples
  • Custom OpenAI-compatible Endpoint
  • Security: API Key Best Practices
• HypoResult Reference
  • Fields
  • summary() Output Format
  • to_dict()
  • plot()
  • Accessing Individual Fields
• Assumption Checking
  • check_normality()
  • check_equal_variances()
  • Individual Tests
  • Full Decision Workflow
• Power Analysis
  • Required sample size — two-sample t-test
  • Post-hoc power — two-sample t-test
  • Worked Example: Planning a Study
  • Power Curves
• Visualization
  • result.plot()
  • plot_result()
  • plot_distributions()
  • plot_p_value()
  • generate_report()
  • Reporting module helpers
  • Notes
• Reporting
  • generate_report()
  • export_html()
  • export_pdf()
  • export_csv()
  • apa_report()
  • APA Formatting in summary()
  • Multiple Results Example

API Reference

• analyze() — Natural Language Entry Point
• Statistical Test Functions
• HypoResult
• LLM Backends
• Power Analysis
• Visualization

Project

• Changelog
• Contributing
• License

Indices and tables

• Index

• Module Index

• Search Page