scitex 2.1.2__py3-none-any.whl → 2.4.2__py3-none-any.whl

scitex/stats/docs/SUMMARY.md

@@ -1,241 +0,0 @@
-# SciTeX Stats Module - Complete Summary
-
-## Module Structure
-
-```
-stx.stats/
-├── tests/              # All statistical tests (test_* prefix)
-│   ├── parametric/     # t-tests, ANOVA
-│   ├── nonparametric/  # Brunner-Munzel, Wilcoxon, Kruskal, Mann-Whitney
-│   ├── normality/      # Shapiro-Wilk, Kolmogorov-Smirnov
-│   ├── correlation/    # Pearson, Spearman
-│   └── categorical/    # Chi-square, Fisher's exact
-├── correct/            # Multiple comparison corrections
-├── effect_sizes/       # Effect size measures
-└── power/              # Power analysis
-```
-
-## Implemented Tests (14 total)
-
-### Parametric (4)
-1. `test_ttest_ind()` - Independent samples t-test
-2. `test_ttest_rel()` - Paired t-test
-3. `test_ttest_1samp()` - One-sample t-test
-4. `test_anova()` - One-way ANOVA
-
-### Non-parametric (4)
-5. `test_brunner_munzel()` - Robust 2-sample comparison
-6. `test_wilcoxon()` - Paired samples (non-parametric)
-7. `test_kruskal()` - 3+ independent groups
-8. `test_mannwhitneyu()` - 2 independent samples
-
-### Normality (4)
-9. `test_shapiro()` - Shapiro-Wilk test (single sample)
-10. `test_normality()` - Multiple samples (Shapiro-Wilk)
-11. `test_ks_1samp()` - One-sample Kolmogorov-Smirnov
-12. `test_ks_2samp()` - Two-sample Kolmogorov-Smirnov
-
-### Correlation (2)
-13. `test_pearson()` - Linear correlation
-14. `test_spearman()` - Rank correlation (monotonic)
-
-### Categorical (2)
-15. `test_chi2()` - Chi-square independence test
-16. `test_fisher()` - Fisher's exact test (2×2)
-
-## Multiple Comparison Corrections (3)
-
-1. `correct_bonferroni()` - Conservative FWER control
-2. `correct_fdr()` - False Discovery Rate (BH/BY)
-3. `correct_holm()` - Sequential Bonferroni (more powerful)
-
-## Effect Sizes
-
-- **Parametric**: Cohen's d, eta-squared (η²)
-- **Non-parametric**: Cliff's delta, epsilon-squared (ε²), rank-biserial, P(X>Y)
-- **Categorical**: Cramér's V, odds ratio
-- **Correlation**: Pearson's r, Spearman's ρ, R², ρ²
-
-## Key Features
-
-### 1. Normalized Naming
-- ALL test functions use `test_*` prefix
-- Clear, consistent, discoverable API
-- No backward compatibility clutter
-
-### 2. Standardized Output
-Every test returns:
-- Test statistic and p-value
-- Effect size with interpretation
-- Confidence intervals (where applicable)
-- Sample sizes and degrees of freedom
-- Significance stars
-- Assumptions met flag (where applicable)
-
-### 3. Flexible Export (9 formats)
-1. dict
-2. DataFrame
-3. CSV
-4. JSON
-5. Excel (styled)
-6. LaTeX
-7. Markdown
-8. HTML
-9. Plain text
-
-### 4. Publication-Ready Visualizations
-- Automatic plot generation with `plot=True`
-- Professional styling
-- Significance annotations
-- Multiple panel layouts
-
-### 5. Consistent API
-```python
-test_xxx(
-    data,                    # Input data
-    var_x='x',              # Variable names
-    var_y='y',
-    alternative='two-sided', # Hypothesis
-    alpha=0.05,             # Significance level
-    plot=False,             # Visualization
-    return_as='dict',       # Output format
-    decimals=3              # Rounding
-)
-```
-
-### 6. Automatic Assumption Checking
-- ANOVA: normality + homogeneity of variance
-- Chi-square: expected frequencies
-- Parametric tests: optional normality checks
-- Warnings and recommendations when violated
-
-### 7. Power Analysis
-- Compute statistical power
-- Determine required sample sizes
-- Currently for t-tests (expandable)
-
-## Design Principles
-
-1. **Clarity over brevity**: Explicit, self-documenting names
-2. **Consistency over flexibility**: Same API across all tests
-3. **Quality over quantity**: 16 well-implemented tests > 100 half-done
-4. **Professional output**: Publication-ready results and plots
-5. **No legacy burden**: Clean slate, modern Python
-
-## Usage Patterns
-
-### Single Test
-```python
-import scitex as stx
-import numpy as np
-
-x = np.random.normal(0, 1, 30)
-y = np.random.normal(0.5, 1, 30)
-
-result = stx.stats.tests.test_ttest_ind(x, y, plot=True)
-print(result)
-```
-
-### Multiple Tests with Correction
-```python
-results = []
-for i in range(10):
-    r = stx.stats.tests.test_ttest_ind(data_a[i], data_b[i])
-    results.append(r)
-
-corrected = stx.stats.correct.correct_holm(results, alpha=0.05)
-```
-
-### Effect Size Computation
-```python
-d = stx.stats.effect_sizes.cohens_d(x, y)
-interp = stx.stats.effect_sizes.interpret_cohens_d(d)
-print(f"d = {d:.3f} ({interp})")
-```
-
-### Export Results
-```python
-from scitex.stats.utils import convert_results
-
-result = stx.stats.tests.test_anova(group_a, group_b, group_c)
-convert_results(result, return_as='latex', path='results.tex')
-convert_results(result, return_as='excel', path='results.xlsx')
-```
-
-## Comparison with Other Libraries
-
-| Feature | SciTeX Stats | scipy.stats | statsmodels | pingouin |
-|---------|--------------|-------------|-------------|----------|
-| Normalized naming | ✓ | ✗ | ✗ | ✓ |
-| Effect sizes | ✓ | ✗ | Partial | ✓ |
-| Auto plots | ✓ | ✗ | ✗ | ✓ |
-| Multiple formats | ✓ | ✗ | Partial | Partial |
-| Assumption checks | ✓ | ✗ | Manual | ✓ |
-| Power analysis | ✓ | Partial | ✓ | ✓ |
-| Categorical tests | ✓ | Basic | ✓ | Basic |
-| Professional export | ✓ | ✗ | ✗ | ✗ |
-
-## Future Roadmap
-
-### High Priority
-- [ ] Repeated measures ANOVA
-- [ ] Šidák correction
-- [ ] Two-way ANOVA
-- [ ] Partial correlation
-
-### Medium Priority
-- [ ] McNemar's test (paired categorical)
-- [ ] Kendall's tau (correlation)
-- [ ] Friedman test (non-parametric RM ANOVA)
-- [ ] Bootstrap confidence intervals
-
-### Low Priority
-- [ ] Mixed-effects models interface
-- [ ] Bayesian alternatives (Bayes factors)
-- [ ] Time series tests
-- [ ] Survival analysis basics
-
-## Documentation
-
-- **API Reference**: `API.md` - Complete function listing
-- **Examples**: Each test file includes 7-10 comprehensive examples
-- **Theory**: Docstrings include formulas, assumptions, interpretations
-- **Progress**: `PROGRESS.md` - Implementation history
-- **Architecture**: `ARCHITECTURE.md` - Design decisions
-
-## Testing Strategy
-
-Each test module includes `if __name__ == '__main__':` block with:
-- Basic usage
-- Edge cases (small samples, outliers, ties)
-- Assumption violations
-- Comparison with alternative tests
-- Export demonstrations
-- Large dataset performance
-
-Run with: `python -m scitex.stats.tests.xxx._test_yyy`
-
-## Professional Branding
-
-All exports include:
-```
-🤖 Generated with SciTeX Stats
-Professional Statistical Analysis Framework
-https://github.com/ywatanabe/scitex
-```
-
-## Development Completed
-
-**Session 1**: Infrastructure + basic tests (t-tests, Brunner-Munzel, Shapiro-Wilk, Bonferroni, FDR)
-
-**Session 2**: Expansion (Wilcoxon, Kruskal-Wallis, ANOVA, Holm, KS tests, Mann-Whitney, Pearson, Spearman, Chi-square, Fisher's exact) + Module reorganization
-
-**Total**: 16 tests + 3 corrections + comprehensive effect sizes + power analysis + 9 export formats
-
-**Lines of code**: ~8,000+ (excluding examples and tests)
-
-**Documentation**: 4 markdown files + inline docstrings with examples
-
----
-
-**Status**: Production-ready for scientific analysis and publication-quality reporting.

scitex/stats/docs/TODO.md

@@ -1,181 +0,0 @@
-<!-- ---
-!-- Timestamp: 2025-10-01 12:55:00
-!-- Author: ywatanabe + Claude Code
-!-- File: /home/ywatanabe/proj/scitex_repo/src/scitex/stats/TODO.md
-!-- --- -->
-
-# SciTeX Stats - TODO
-
-## ✅ Completed (Version 2.0.0)
-
-### Architecture
-- ✅ Organized module structure
-  - ✅ `scitex.stats.tests` - 23 statistical tests
-  - ✅ `scitex.stats.correct` - 4 multiple comparison corrections
-  - ✅ `scitex.stats.posthoc` - 3 post-hoc tests
-  - ✅ `scitex.stats.effect_sizes` - Effect size computations
-  - ✅ `scitex.stats.power` - Power analysis
-  - ✅ `scitex.stats.utils` - Formatters and normalizers
-- ✅ Normalized naming (all tests use `test_*` prefix, posthoc use `posthoc_*`)
-- ✅ No backward compatibility (clean slate)
-- ✅ Consistent API across all tests
-
-### Tests Implemented (23)
-- ✅ **Parametric (6)**: t-tests (ind, rel, 1samp), ANOVA (one-way, repeated measures, two-way)
-- ✅ **Non-parametric (5)**: Brunner-Munzel, Wilcoxon, Kruskal, Mann-Whitney, Friedman
-- ✅ **Normality (4)**: Shapiro-Wilk, test_normality, KS (1samp, 2samp)
-- ✅ **Correlation (3)**: Pearson, Spearman, Kendall's tau
-- ✅ **Categorical (4)**: Chi-square, Fisher's exact, McNemar, Cochran's Q
-
-### Corrections (4)
-- ✅ **Bonferroni**: Conservative FWER control
-- ✅ **Holm**: Sequential Bonferroni
-- ✅ **FDR (Benjamini-Hochberg)**: False discovery rate
-- ✅ **Šidák**: More powerful than Bonferroni under independence
-
-### Post-hoc Tests (3)
-- ✅ **Tukey HSD**: All pairwise comparisons (equal variances)
-- ✅ **Games-Howell**: Unequal variances (Welch-Satterthwaite)
-- ✅ **Dunnett**: Multiple treatments vs control
-
-### Features
-- ✅ Standardized output format (dict and DataFrame)
-- ✅ 9 export formats (CSV, LaTeX, Excel, JSON, etc.)
-- ✅ Publication-ready visualizations
-- ✅ Automatic assumption checking
-- ✅ Comprehensive effect sizes with interpretations
-- ✅ Power analysis for t-tests
-- ✅ `decimals=3` default rounding
-
-### Documentation
-- ✅ README.md - Professional overview
-- ✅ QUICKSTART.md - Quick start guide with examples
-- ✅ docs/API.md - Complete API reference
-- ✅ docs/PROGRESS.md - Detailed implementation with examples
-- ✅ docs/SUMMARY.md - Comprehensive summary and comparisons
-
----
-
-## 🔄 In Progress
-
-None - Version 2.0.0 is complete and production-ready!
-
----
-
-## 📋 Future Enhancements (Version 2.1+)
-
-### High Priority (Version 2.1+)
-- ✅ **Šidák correction** (COMPLETED)
-  - Alternative to Bonferroni assuming independence
-  - Formula: 1 - (1-α)^(1/m)
-  - More powerful than Bonferroni under independence
-
-- ✅ **McNemar's test** (COMPLETED)
-  - Paired categorical data
-  - 2×2 contingency tables for matched pairs
-  - Odds ratio effect size
-
-- ✅ **Repeated measures ANOVA** (COMPLETED)
-  - Within-subjects factor
-  - Sphericity testing (Mauchly's test)
-  - Greenhouse-Geisser correction
-  - Partial eta-squared effect size
-
-- ✅ **Two-way ANOVA** (COMPLETED)
-  - Factorial design (2 factors)
-  - Main effects and interaction effects
-  - Partial eta-squared for each effect
-  - Interaction plots
-
-### Medium Priority (Version 2.1+)
-- ✅ **Kendall's tau correlation** (COMPLETED)
-  - Alternative to Spearman
-  - Better for small samples with ties
-  - tau-b and tau-c variants
-
-- ✅ **Friedman test** (COMPLETED)
-  - Non-parametric repeated measures ANOVA
-  - 3+ related samples
-  - Kendall's W effect size
-
-- ✅ **Cochran's Q test** (COMPLETED)
-  - Binary repeated measures
-  - Extension of McNemar's test to 3+ conditions
-  - Kendall's W for binary data
-
-- [ ] **Bootstrap confidence intervals**
-  - For effect sizes
-  - Non-parametric CI estimation
-
-- ✅ **Post-hoc framework enhancement** (COMPLETED)
-  - ✅ Tukey HSD - All pairwise comparisons
-  - ✅ Games-Howell - Unequal variances
-  - ✅ Dunnett - Multiple treatments vs control
-
-### Low Priority
-- [ ] **Permutation tests**
-  - General non-parametric alternative
-  - Exact p-values
-
-- [ ] **Mixed-effects models interface**
-  - Basic linear mixed models
-  - Random effects
-
-- [ ] **Bayesian alternatives**
-  - Bayes factors
-  - Credible intervals
-
-- [ ] **Time series tests**
-  - Autocorrelation tests
-  - Stationarity tests
-
-- [ ] **Survival analysis basics**
-  - Log-rank test
-  - Kaplan-Meier curves
-
-### Infrastructure Improvements
-- [ ] **Performance optimization**
-  - Lazy imports to speed up module loading
-  - Caching for repeated operations
-
-- [ ] **Extended export formats**
-  - APA-style tables
-  - R-compatible output
-
-- [ ] **Enhanced visualizations**
-  - Interactive plots (plotly)
-  - Customizable themes
-
-- [ ] **Batch processing**
-  - Apply same test to multiple datasets
-  - Parallel computation support
-
----
-
-## 💡 Ideas for Discussion
-
-- [ ] Auto-test selection based on data properties
-- [ ] Integration with pandas DataFrames (native support)
-- [ ] Effect size visualization (forest plots, etc.)
-- [ ] Sample size calculator web interface
-- [ ] Automatic report generation (full analysis pipeline)
-
----
-
-## 🐛 Known Issues
-
-None currently - all tests working as expected!
-
----
-
-## 📝 Notes
-
-- Version 2.0.0 represents a complete rewrite with clean architecture
-- All legacy code moved to `.old/` directory
-- No backward compatibility with version 1.x (by design)
-- Focus on quality over quantity - 16 well-implemented tests
-- All tests have 7-10 comprehensive examples in `__main__` blocks
-
----
-
-<!-- EOF -->

scitex/stats/docs/ax_parameter_migration.md

@@ -1,125 +0,0 @@
-# Migration to `ax=` Parameter Pattern
-
-## Summary
-Statistical test functions now use an `ax=` parameter instead of `plot=` for visualization. This provides better flexibility for users to control figure layout while keeping the API simple.
-
-## Key Changes
-
-### Before (old pattern with `plot=`)
-```python
-def test_function(..., plot: bool = False, ...) -> Union[dict, pd.DataFrame, Tuple]:
-    # ... test logic ...
-
-    fig = None
-    if plot:
-        fig = _plot_function(...)
-
-    if plot:
-        return result, fig
-    else:
-        return result
-```
-
-### After (new pattern with `ax=`)
-```python
-def test_function(..., ax: Optional[matplotlib.axes.Axes] = None, ...) -> Union[dict, pd.DataFrame]:
-    # ... test logic ...
-
-    if ax is not None:
-        _plot_function(..., ax)
-
-    return result  # Always return result only
-```
-
-## Benefits
-
-1. **Simpler return type**: Functions always return test results only (never tuples)
-2. **User controls figures**: Users create their own figures and axes
-3. **Subplot integration**: Easy to place multiple tests in one figure
-4. **Matplotlib convention**: Follows standard matplotlib pattern
-
-## Updated Files
-
-### Completed
-- ✅ `tests/parametric/_test_ttest.py` (all 3 functions: ind, rel, 1samp)
-- ✅ `tests/parametric/_test_anova.py`
-
-### Remaining (16 files with `plot=` parameter)
-- `tests/nonparametric/_test_brunner_munzel.py`
-- `tests/nonparametric/_test_friedman.py`
-- `tests/nonparametric/_test_kruskal.py`
-- `tests/nonparametric/_test_mannwhitneyu.py`
-- `tests/nonparametric/_test_wilcoxon.py`
-- `tests/parametric/_test_anova_2way.py`
-- `tests/parametric/_test_anova_rm.py`
-- `tests/correlation/_test_kendall.py`
-- `tests/correlation/_test_pearson.py`
-- `tests/correlation/_test_spearman.py`
-- `tests/categorical/_test_chi2.py`
-- `tests/categorical/_test_cochran_q.py`
-- `tests/categorical/_test_fisher.py`
-- `tests/categorical/_test_mcnemar.py`
-- `tests/normality/_test_ks.py`
-- `tests/normality/_test_shapiro.py`
-
-## Usage Examples
-
-### Old way (with `plot=True`)
-```python
-# Returns tuple - awkward
-result, fig = test_ttest_ind(x, y, plot=True)
-plt.show()
-```
-
-### New way (with `ax=`)
-```python
-# Simple: always returns result
-result = test_ttest_ind(x, y)
-
-# With visualization
-fig, ax = plt.subplots()
-result = test_ttest_ind(x, y, ax=ax)
-plt.show()
-
-# In subplots
-fig, axes = plt.subplots(2, 2)
-result1 = test_ttest_ind(x1, y1, ax=axes[0, 0])
-result2 = test_ttest_ind(x2, y2, ax=axes[0, 1])
-result3 = test_anova([g1, g2, g3], ax=axes[1, 0])
-plt.tight_layout()
-```
-
-## Implementation Checklist
-
-For each function that needs migration:
-
-1. **Function signature**:
-   - [ ] Replace `plot: bool = False` with `ax: Optional[matplotlib.axes.Axes] = None`
-   - [ ] Add `import matplotlib.axes` at top
-   - [ ] Remove `Tuple` from return type annotation
-
-2. **Docstring**:
-   - [ ] Update Parameters section to describe `ax=`
-   - [ ] Update Returns section to remove figure return
-   - [ ] Update Examples to show new usage pattern
-
-3. **Function body**:
-   - [ ] Replace `if plot:` with `if ax is not None:`
-   - [ ] Pass `ax` to plotting helper function
-   - [ ] Remove `fig = None` and tuple return logic
-   - [ ] Always `return result` (never return tuple)
-
-4. **Helper plotting function**:
-   - [ ] Add `ax` as last parameter
-   - [ ] Remove `fig, axes = plt.subplots(...)` creation
-   - [ ] Plot directly on provided `ax`
-   - [ ] Remove `return fig`
-
-5. **Examples in main()**:
-   - [ ] Update any `plot=True` usage to create fig/ax first
-
-## Notes
-
-- Posthoc functions (tukey, games-howell, dunnett) don't have plotting - no changes needed
-- The `ax=` parameter is always optional - if not provided, no plotting occurs
-- Functions should not create figures internally - user's responsibility

scitex/stats/template.py

@@ -1,119 +0,0 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-# Timestamp: "2025-10-01 12:58:28 (ywatanabe)"
-# File: /ssh:sp:/home/ywatanabe/proj/scitex_repo/src/scitex/stats/template.py
-# ----------------------------------------
-from __future__ import annotations
-import os
-__FILE__ = __file__
-__DIR__ = os.path.dirname(__FILE__)
-# ----------------------------------------
-
-"""
-Functionalities:
-  - Does XYZ
-  - Does XYZ
-  - Does XYZ
-  - Saves XYZ
-
-Dependencies:
-  - scripts:
-    - /path/to/script1
-    - /path/to/script2
-  - packages:
-    - package1
-    - package2
-IO:
-  - input-files:
-    - /path/to/input/file.xxx
-    - /path/to/input/file.xxx
-
-  - output-files:
-    - /path/to/input/file.xxx
-    - /path/to/input/file.xxx
-
-(Remove me: Please fill docstrings above, while keeping the bulette point style, and remove this instruction line)
-"""
-
-"""Imports"""
-import argparse
-
-import scitex as stx
-from scitex import logging
-
-logger = logging.getLogger(__name__)
-
-"""Warnings"""
-# stx.pd.ignore_SettingWithCopyWarning()
-# warnings.simplefilter("ignore", UserWarning)
-# with warnings.catch_warnings():
-#     warnings.simplefilter("ignore", UserWarning)
-
-"""Parameters"""
-# CONFIG = stx.io.load_configs()
-
-"""Functions & Classes"""
-def main(args):
-    return 0
-
-
-def parse_args() -> argparse.Namespace:
-    """Parse command line arguments."""
-    import scitex as stx
-
-    parser = argparse.ArgumentParser(description="")
-    # parser.add_argument(
-    #     "--var",
-    #     "-v",
-    #     type=int,
-    #     choices=None,
-    #     default=1,
-    #     help="(default: %(default)s)",
-    # )
-    # parser.add_argument(
-    #     "--flag",
-    #     "-f",
-    #     action="store_true",
-    #     default=False,
-    #     help="(default: %%(default)s)",
-    # )
-    args = parser.parse_args()
-    return args
-
-
-def run_main() -> None:
-    """Initialize scitex framework, run main function, and cleanup."""
-    global CONFIG, CC, sys, plt, rng
-
-    import sys
-
-    import matplotlib.pyplot as plt
-    import scitex as stx
-
-    args = parse_args()
-
-    CONFIG, sys.stdout, sys.stderr, plt, CC, rng = stx.session.start(
-        sys,
-        plt,
-        args=args,
-        file=__FILE__,
-        sdir_suffix=None,
-        verbose=False,
-        agg=True,
-    )
-
-    exit_status = main(args)
-
-    stx.session.close(
-        CONFIG,
-        verbose=False,
-        notify=False,
-        message="",
-        exit_status=exit_status,
-    )
-
-
-if __name__ == "__main__":
-    run_main()
-
-# EOF

scitex/stats/tests/categorical/__init__.py

@@ -1,20 +0,0 @@
-#!/usr/bin/env python3
-# File: ./scitex_repo/src/scitex/stats/tests/categorical/__init__.py
-
-"""
-Categorical data tests.
-
-Tests for analyzing contingency tables and categorical associations.
-"""
-
-from ._test_chi2 import test_chi2
-from ._test_fisher import test_fisher
-from ._test_mcnemar import test_mcnemar
-from ._test_cochran_q import test_cochran_q
-
-__all__ = [
-    'test_chi2',
-    'test_fisher',
-    'test_mcnemar',
-    'test_cochran_q',
-]