v0.2.6

feat: add docs CLI subcommand via shared scitex-dev infrastructure

v0.2.3

Changes

• H0 field added to all 23 statistical tests (null hypothesis in plain text)
• Decision flowchart (PNG) added to README and RTD for choosing the right test
• Example figure showing t-test boxplot with significance bracket and effect size
• Elaborated RTD documentation — 6 new pages: choosing_tests, unified_output, cli, mcp, examples, plus improved index/quickstart
• SciTeX ecosystem table in README showing stats, plt (figrecipe), io, clew integration
• README title updated: Stats → SciTeX Stats

v0.2.2

Changes

• All examples now save both results.txt and results.json artifacts
• README unified result dictionary expanded with H0, alternative, n_x, n_y, alpha fields

v0.2.1

Fixes

• README: correct StatContext constructor, fix posthoc API name, add CI badge
• README: add unified result dictionary example (Table 3)
• Fix stale scitex.stats references across 78 source files
• Clean up merged branches

Full Changelog

v0.2.0...v0.2.1

v0.2.0 — Standalone Package

Breaking Change

scitex-stats is now a fully standalone package. Previously it was an empty shell that required scitex-python to be installed. Now pip install scitex-stats works out of the box.

What's New

• 97 Python source files extracted from scitex-python into standalone package
• Logging compatibility: uses scitex.logging when available, falls back to stdlib
• All 23 statistical tests, effect sizes, corrections, post-hoc, power analysis included
• Decision tree for automatic test recommendation
• MCP server with 10 tools for AI agents

Test Results

• 74 tests passing, 0 failures
• MCP doctor: all checks passed

Full Changelog

v0.1.1...v0.2.0

v0.1.1

What's Changed

Bug Fixes

• FastMCP v3 compatibility: replaced deprecated _tool_manager._tools and get_tools() with list_tools() across CLI and tests
• Mermaid diagram in README now uses high-contrast colors visible on both light and dark themes

Improvements

• Examples now save output artifacts to _out/ directories, tracked in git
• pytest coverage reporting enabled (--cov=scitex_stats --cov-report=term-missing)
• 74 tests passing, 0 failures

Full Changelog

v0.1.0...v0.1.1

v0.1.0 — Initial Release

scitex-stats v0.1.0

Publication-ready statistical testing framework — standalone package extracted from scitex.

Features

• 23 statistical tests with automatic recommendation
• Effect sizes (Cohen's d, Cliff's delta, eta squared)
• Power analysis and sample size calculation
• APA-formatted output
• Three interfaces: Python API, CLI, MCP server (10 tools)
• 74 passing tests

Install

pip install scitex-stats

Links

• PyPI
• Documentation