hypors 0.4.1__tar.gz

hypors-0.4.1/.gitignore

@@ -0,0 +1,22 @@
+# Rust
+
+## Cargo compilations
+debug/
+target/
+
+# Cargo files
+Cargo.lock
+**/*.rs.bk
+
+# Python
+
+## Cache
+__pycache__
+.pytest_cache
+
+## Virtual Environment
+hypenv/
+
+# Other
+
+notes.txt

hypors-0.4.1/CHANGELOG.md

@@ -0,0 +1,100 @@
+# Changelog
+
+All notable changes to this project are documented in this file. The format
+follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
+
+While this crate is pre-1.0, the **minor** version carries breaking changes:
+a dependency on `0.3` will not resolve to `0.4`.
+
+## [Unreleased]
+
+Nothing yet. Add entries here as they land; move them under a new version
+heading when releasing, since the release workflow takes its GitHub Release
+notes from the section matching the tag.
+
+## [0.4.1] - 2026-09-13
+
+The first release of the Python package. It ships at 0.4.1 rather than 0.4.0
+because versions track the crate, and the `v0.4.0` tag predates the bindings —
+there is no Python package in that tree to publish.
+
+### Added
+
+- **Python bindings** (`python/`), published to PyPI as `hypors` and versioned
+  in lockstep with this crate. All 20 public test and sample-size functions are
+  exposed across `hypors.t`, `hypors.z`, `hypors.proportion`, `hypors.anova`,
+  `hypors.chi_square` and `hypors.mann_whitney`, with `TailType` and
+  `TestResult` in `hypors.common` and at the top level.
+  - Any Python iterable of numbers is accepted — list, tuple, generator,
+    `numpy` array, `pandas`/`polars` Series — and the package depends on none
+    of them.
+  - `StatError` surfaces as `ValueError` for empty or insufficient data and
+    `RuntimeError` for a failed computation.
+  - The pytest suite asserts the same fixture values as the Rust integration
+    tests, so both layers are held to the scipy-verified numbers.
+- CI builds the wheel, lints the binding crate and runs the Python suite.
+- The release workflow now publishes both registries from one `vX.Y.Z` tag:
+  wheels for Linux, macOS and Windows plus a source distribution go to PyPI
+  behind the `pypi` environment, alongside the existing crates.io publish.
+  Each registry is skipped if it already holds the version, so a release can
+  be re-run, and `workflow_dispatch` allows running against an existing tag.
+  The version guard now checks all three manifests agree with the tag.
+
+### Changed
+
+- The crate no longer builds a `cdylib`. It was a leftover from an earlier
+  attempt at putting the bindings in this crate, produced a `libhypors.so`
+  nothing consumed, and collided with the extension module's output filename.
+
+## [0.4.0] - 2026-09-07
+
+### Changed
+
+- **Mann-Whitney U p-values now match `scipy.stats.mannwhitneyu`
+  (`method="asymptotic"`).** Existing p-values change, and one-sided results
+  change substantially. Thanks to @guilherme-n-l (#7, reported in #6).
+  - The variance carries the tie correction
+    `n1*n2/12 * ((N + 1) - sum(t^3 - t) / (N * (N - 1)))`. The previous
+    no-ties variance overestimated sigma and inflated p-values on tied data.
+  - A 0.5 continuity correction is applied toward the tail, matching scipy's
+    default `use_continuity=True`.
+  - One-sided p-values derive from `U1` rather than `min(U1, U2)`. The old
+    statistic was sign-blind, so `TailType::Right` and `TailType::Left`
+    returned identical p-values regardless of which group dominated.
+  - When every observation is tied the variance is zero: the two-sided
+    p-value is now `NaN` and the one-sided p-values are `1.0`, matching
+    scipy 1.18's signed continuity correction. Previously this returned
+    `1.0` two-sided and `0.5` one-sided.
+- `test_statistic` is documented explicitly as `min(U1, U2)` while the
+  p-value derives from `U1`, so compare p-values rather than statistics
+  when checking against scipy.
+- **`mann_whitney::u_test` and `chi_square::variance` now return
+  `Result<TestResult, StatError>`** instead of `Result<TestResult, String>`,
+  matching every other test in the crate. Code matching on the error needs
+  updating; code using `?` or `unwrap()` does not. Empty input now yields
+  `StatError::EmptyData`, too few observations `StatError::InsufficientData`,
+  and distribution failures `StatError::ComputeError`.
+
+### Added
+
+- `rust-version = "1.85"`, the minimum supported Rust version for
+  edition 2024.
+- GitHub Actions CI running `cargo fmt --check`,
+  `cargo clippy --all-targets -- -D warnings`, `cargo test` (doc tests
+  included) and `cargo package` on every pull request (#8, #9).
+
+### Fixed
+
+- The manifest is named `Cargo.toml` rather than `cargo.toml`, which only
+  resolved on case-insensitive filesystems and broke every cargo command on
+  Linux (#8).
+
+### Internal
+
+- The published crate no longer ships the Python bindings (`hypopy`),
+  `py_tests`, or the CI configuration — 54 files down to 41 (#8).
+
+## [0.3.0] - 2025-04-19
+
+See the [release history](https://github.com/astronights/hypors/releases) for
+versions before this changelog was introduced.

hypors-0.4.1/Cargo.toml

@@ -0,0 +1,25 @@
+[package]
+name = "hypors"
+authors = ["Shubhankar Agrawal <shubhankar.a31@gmail.com>"]
+version = "0.4.1"
+edition = "2024"
+rust-version = "1.85"        # edition 2024 requires 1.85+
+license-file = "LICENSE"
+repository = "https://github.com/astronights/hypors"
+description = "Hypothesis Testing with Rust"
+keywords = ["statistics", "hypothesis", "testing", "data", "p-value"]
+
+# Keep the published crate to the Rust library: the Python bindings
+# (which carry their own manifest and pyproject.toml), their tests, and
+# the CI config are of no use to a crates.io consumer.
+exclude = [".github", "python"]
+readme = "README.md"
+
+[lib]
+name = "hypors"
+path = "src/lib.rs" 
+crate-type = ["lib"]
+
+[dependencies]
+serde = {version = ">=1.0.210", features = ["derive"]}
+statrs = ">=0.17.1"

hypors-0.4.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 Shubhankar Agrawal
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

hypors-0.4.1/PKG-INFO

@@ -0,0 +1,93 @@
+Metadata-Version: 2.4
+Name: hypors
+Version: 0.4.1
+Classifier: Programming Language :: Rust
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Topic :: Scientific/Engineering :: Mathematics
+Summary: Hypothesis testing for Python, backed by the hypors Rust crate
+Keywords: statistics,hypothesis,testing,p-value
+Author-email: Shubhankar Agrawal <shubhankar.a31@gmail.com>
+License-Expression: MIT
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM
+Project-URL: Changelog, https://github.com/astronights/hypors/blob/main/CHANGELOG.md
+Project-URL: Repository, https://github.com/astronights/hypors
+
+# hypors
+
+Hypothesis testing for Python, backed by the
+[`hypors`](https://crates.io/crates/hypors) Rust crate.
+
+Every statistic is computed by the Rust library — these bindings only convert
+values across the boundary, so Python and Rust always agree.
+
+## Install
+
+```bash
+pip install hypors
+```
+
+## Use
+
+```python
+from hypors import TailType
+from hypors.t import t_test
+
+result = t_test([1.2, 2.3, 1.9, 2.5, 2.8], pop_mean=2.0, tail=TailType.Two, alpha=0.05)
+
+print(result.test_statistic)
+print(result.p_value)
+print(result.reject_null)
+print(result.to_dict())
+```
+
+Any iterable of numbers works — a `list`, `tuple`, generator, `numpy` array,
+`pandas.Series` or `polars.Series`. The package itself depends on none of
+them:
+
+```python
+import numpy as np
+import polars as pl
+
+t_test(np.array([1.2, 2.3, 1.9]), 2.0, TailType.Two, 0.05)
+t_test(pl.Series([1.2, 2.3, 1.9]), 2.0, TailType.Two, 0.05)
+```
+
+## Tests available
+
+| Module | Functions |
+| --- | --- |
+| `hypors.t` | `t_test`, `t_test_paired`, `t_test_ind`, `t_sample_size` |
+| `hypors.z` | `z_test`, `z_test_paired`, `z_test_ind`, `z_sample_size` |
+| `hypors.proportion` | `z_test`, `z_test_ind`, `prop_sample_size` |
+| `hypors.anova` | `anova`, `f_sample_size` |
+| `hypors.chi_square` | `independence`, `goodness_of_fit`, `variance`, `chi2_sample_size_gof`, `chi2_sample_size_ind`, `chi2_sample_size_variance` |
+| `hypors.mann_whitney` | `u_test` |
+| `hypors.common` | `TailType`, `TestResult` |
+
+`TailType` and `TestResult` are also importable from the top level.
+
+## Results
+
+Every test returns a `TestResult` with `test_statistic`, `p_value`,
+`confidence_interval`, `null_hypothesis`, `alt_hypothesis` and `reject_null`,
+plus `to_dict()` for serialising. Attributes are read-only.
+
+Errors surface as ordinary Python exceptions: `ValueError` for empty or
+insufficient data, `RuntimeError` for a computation that could not be
+completed.
+
+## Notes
+
+- The Mann-Whitney U p-value matches `scipy.stats.mannwhitneyu` with
+  `method="asymptotic"`. Its `test_statistic` is `min(U1, U2)` while the
+  p-value derives from `U1`, so compare p-values rather than statistics when
+  checking against scipy.
+- Versions track the Rust crate: `hypors` on PyPI and `hypors` on crates.io
+  share a version number.
+
+## License
+
+MIT
+

hypors-0.4.1/README.md

@@ -0,0 +1,124 @@
+# HypoRS: A Statistical Hypothesis Testing Library
+
+`hypors` is a Rust library designed for performing a variety of hypothesis tests, including t-tests, z-tests, proportion tests, ANOVA, Chi-square tests, and Mann-Whitney tests. This library leverages the `statrs` crate for statistical distributions.
+
+Rust Crate: https://crates.io/crates/hypors
+
+PyPI Package: Work in Progress
+
+## Features
+
+### Hypothesis Tests
+
+Hypothesis testing is available for this suite of common distributions.
+
+- **T-Tests**: One-sample, two-sample paired, and two-sample independent t-tests.
+- **Z-Tests**: One-sample, two-sample paired, and two-sample independent z-tests.
+- **Proportion Tests**: One-sample and two-sample proportion tests.
+- **ANOVA**: One-way ANOVA for comparing means across multiple groups.
+- **Chi-Square Tests**: Chi-square test for independence and goodness-of-fit tests.
+- **Mann-Whitney U Test**: Non-parametric test for comparing two independent samples.
+
+### Sample Size Calculation
+
+All parametrized distributions have respective modules to calculate minimum sample size required with customizable parameters for alpha and statistical power.
+
+
+### **Additional Features**:
+  - Customizable tail type (left, right, and two-tailed).
+  - Customizable alpha value for all tests.
+  - Confidence interval calculations for all tests.
+  - p-value is generated along with each statistic.
+  - Null and alternate hypotheses strings are also generated.
+
+## Installation
+
+To use this library in your Rust project, add the following to your `Cargo.toml`:
+
+```toml
+[dependencies]
+hypors = "0.4.0"
+```
+
+**Note** HypoRS relies on the following dependencies, which will be automatically included:
+
+```
+serde (version >=1.0.210)
+statrs (version >=0.17.1)
+```
+
+## Example Usage
+
+### Rust
+
+Here are some examples of running tests with Rust.
+
+#### T - Test
+
+```rust
+use hypors::{t::t_test, common::TailType};
+
+let data = vec![1.2, 2.3, 1.9, 2.5, 2.8];
+let population_mean = 2.0;
+let tail = TailType::Two;
+let alpha = 0.05;
+
+let result = t_test(data, population_mean, tail, alpha).unwrap();
+println!("Test Statistic: {}", result.test_statistic);
+println!("p-value: {}", result.p_value);
+println!("Confidence Interval: {}", result.confidence_interval);
+println!("Null Hypothesis: {}", result.null_hypothesis);
+println!("Alternate Hypothesis: {}", result.alt_hypothesis);
+println!("Reject Null Hypothesis?: {}", result.reject_null);
+```
+
+#### Chi Square Test
+
+```rust
+use hypors::{chi_square::independence};
+
+let observed = vec![10, 20, 30]; // Observed frequencies
+let expected = vec![15, 15, 30]; // Expected frequencies
+
+let result = independece(observed, expected).unwrap();
+println!("Chi-Square Statistic: {}", result.test_statistic);
+println!("p-value: {}", result.p_value);
+println!("Null Hypothesis: {}", result.null_hypothesis);
+println!("Alternate Hypothesis: {}", result.alt_hypothesis);
+println!("Reject Null Hypothesis?: {}", result.reject_null);
+```
+
+### Usage with Polars
+
+With the `polars` crate having frequent minor version updates, this library has been updated to not use Polars inherently, and relies on `Vec` types.
+
+Any use of `polars` `Series` types should be converted before and after using the `hypors` functions.
+
+**Polars to Vec Conversion:**
+```rust
+use polars::prelude::*;
+let series = Series::new("series_name", &[1, 2, 3, 4, 5]);
+let vec = series.to_vec().unwrap();
+```
+
+**Vec to Polars Conversion:**
+```rust
+use polars::prelude::*;
+let vec = vec![1, 2, 3, 4, 5];
+let series = Series::new("series_name", vec);
+```
+
+
+### Python
+
+Work in Progress
+
+## Future Plans
+
+The next step for `hypors` is to add Python bindings to make it accessible to the Python community. This work is currently in progress. Stay tuned for updates!
+
+## Contributing
+Contributions are always welcome! If you have suggestions for improvements, bug fixes, or new features, please feel free to open an issue or submit a pull request.
+
+## License
+This project is licensed under the MIT License.

hypors-0.4.1/pyproject.toml

@@ -0,0 +1,27 @@
+[build-system]
+requires = ["maturin>=1.7,<2.0"]
+build-backend = "maturin"
+
+[project]
+name = "hypors"
+version = "0.4.1"
+description = "Hypothesis testing for Python, backed by the hypors Rust crate"
+readme = "README.md"
+authors = [{ name = "Shubhankar Agrawal", email = "shubhankar.a31@gmail.com" }]
+license = "MIT"
+requires-python = ">=3.8"
+keywords = ["statistics", "hypothesis", "testing", "p-value"]
+classifiers = [
+    "Programming Language :: Rust",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: Implementation :: CPython",
+    "Topic :: Scientific/Engineering :: Mathematics",
+]
+
+[project.urls]
+Repository = "https://github.com/astronights/hypors"
+Changelog = "https://github.com/astronights/hypors/blob/main/CHANGELOG.md"
+
+[tool.maturin]
+features = ["pyo3/extension-module"]
+manifest-path = "python/Cargo.toml"