Metadata-Version: 2.3
Name: samesame
Version: 0.3.1
Summary: Statistical tests for model monitoring, data validation, and drift detection.
Keywords: drift detection,data monitoring,dataset shift,data validation,model validation,noninferiority tests,covariate balance,statistical tests
Author: Vathy M. Kamulete
Author-email: Vathy M. Kamulete <vathymut@gmail.com>
License:                    GNU LESSER GENERAL PUBLIC LICENSE
                                Version 3, 29 June 2007
         
          Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
          Everyone is permitted to copy and distribute verbatim copies
          of this license document, but changing it is not allowed.
         
         
           This version of the GNU Lesser General Public License incorporates
         the terms and conditions of version 3 of the GNU General Public
         License, supplemented by the additional permissions listed below.
         
           0. Additional Definitions.
         
           As used herein, "this License" refers to version 3 of the GNU Lesser
         General Public License, and the "GNU GPL" refers to version 3 of the GNU
         General Public License.
         
           "The Library" refers to a covered work governed by this License,
         other than an Application or a Combined Work as defined below.
         
           An "Application" is any work that makes use of an interface provided
         by the Library, but which is not otherwise based on the Library.
         Defining a subclass of a class defined by the Library is deemed a mode
         of using an interface provided by the Library.
         
           A "Combined Work" is a work produced by combining or linking an
         Application with the Library.  The particular version of the Library
         with which the Combined Work was made is also called the "Linked
         Version".
         
           The "Minimal Corresponding Source" for a Combined Work means the
         Corresponding Source for the Combined Work, excluding any source code
         for portions of the Combined Work that, considered in isolation, are
         based on the Application, and not on the Linked Version.
         
           The "Corresponding Application Code" for a Combined Work means the
         object code and/or source code for the Application, including any data
         and utility programs needed for reproducing the Combined Work from the
         Application, but excluding the System Libraries of the Combined Work.
         
           1. Exception to Section 3 of the GNU GPL.
         
           You may convey a covered work under sections 3 and 4 of this License
         without being bound by section 3 of the GNU GPL.
         
           2. Conveying Modified Versions.
         
           If you modify a copy of the Library, and, in your modifications, a
         facility refers to a function or data to be supplied by an Application
         that uses the facility (other than as an argument passed when the
         facility is invoked), then you may convey a copy of the modified
         version:
         
            a) under this License, provided that you make a good faith effort to
            ensure that, in the event an Application does not supply the
            function or data, the facility still operates, and performs
            whatever part of its purpose remains meaningful, or
         
            b) under the GNU GPL, with none of the additional permissions of
            this License applicable to that copy.
         
           3. Object Code Incorporating Material from Library Header Files.
         
           The object code form of an Application may incorporate material from
         a header file that is part of the Library.  You may convey such object
         code under terms of your choice, provided that, if the incorporated
         material is not limited to numerical parameters, data structure
         layouts and accessors, or small macros, inline functions and templates
         (ten or fewer lines in length), you do both of the following:
         
            a) Give prominent notice with each copy of the object code that the
            Library is used in it and that the Library and its use are
            covered by this License.
         
            b) Accompany the object code with a copy of the GNU GPL and this license
            document.
         
           4. Combined Works.
         
           You may convey a Combined Work under terms of your choice that,
         taken together, effectively do not restrict modification of the
         portions of the Library contained in the Combined Work and reverse
         engineering for debugging such modifications, if you also do each of
         the following:
         
            a) Give prominent notice with each copy of the Combined Work that
            the Library is used in it and that the Library and its use are
            covered by this License.
         
            b) Accompany the Combined Work with a copy of the GNU GPL and this license
            document.
         
            c) For a Combined Work that displays copyright notices during
            execution, include the copyright notice for the Library among
            these notices, as well as a reference directing the user to the
            copies of the GNU GPL and this license document.
         
            d) Do one of the following:
         
                0) Convey the Minimal Corresponding Source under the terms of this
                License, and the Corresponding Application Code in a form
                suitable for, and under terms that permit, the user to
                recombine or relink the Application with a modified version of
                the Linked Version to produce a modified Combined Work, in the
                manner specified by section 6 of the GNU GPL for conveying
                Corresponding Source.
         
                1) Use a suitable shared library mechanism for linking with the
                Library.  A suitable mechanism is one that (a) uses at run time
                a copy of the Library already present on the user's computer
                system, and (b) will operate properly with a modified version
                of the Library that is interface-compatible with the Linked
                Version.
         
            e) Provide Installation Information, but only if you would otherwise
            be required to provide such information under section 6 of the
            GNU GPL, and only to the extent that such information is
            necessary to install and execute a modified version of the
            Combined Work produced by recombining or relinking the
            Application with a modified version of the Linked Version. (If
            you use option 4d0, the Installation Information must accompany
            the Minimal Corresponding Source and Corresponding Application
            Code. If you use option 4d1, you must provide the Installation
            Information in the manner specified by section 6 of the GNU GPL
            for conveying Corresponding Source.)
         
           5. Combined Libraries.
         
           You may place library facilities that are a work based on the
         Library side by side in a single library together with other library
         facilities that are not Applications and are not covered by this
         License, and convey such a combined library under terms of your
         choice, if you do both of the following:
         
            a) Accompany the combined library with a copy of the same work based
            on the Library, uncombined with any other library facilities,
            conveyed under the terms of this License.
         
            b) Give prominent notice with the combined library that part of it
            is a work based on the Library, and explaining where to find the
            accompanying uncombined form of the same work.
         
           6. Revised Versions of the GNU Lesser General Public License.
         
           The Free Software Foundation may publish revised and/or new versions
         of the GNU Lesser General Public License from time to time. Such new
         versions will be similar in spirit to the present version, but may
         differ in detail to address new problems or concerns.
         
           Each version is given a distinguishing version number. If the
         Library as you received it specifies that a certain numbered version
         of the GNU Lesser General Public License "or any later version"
         applies to it, you have the option of following the terms and
         conditions either of that published version or of any later version
         published by the Free Software Foundation. If the Library as you
         received it does not specify a version number of the GNU Lesser
         General Public License, you may choose any version of the GNU Lesser
         General Public License ever published by the Free Software Foundation.
         
           If the Library as you received it specifies that a proxy can decide
         whether future versions of the GNU Lesser General Public License shall
         apply, that proxy's public statement of acceptance of any version is
         permanent authorization for you to choose that version for the
         Library.
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Typing :: Typed
Classifier: Topic :: Scientific/Engineering
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Intended Audience :: Science/Research
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: GNU Lesser General Public License v3 (LGPLv3)
Classifier: Operating System :: OS Independent
Requires-Dist: numpy>=1.21
Requires-Dist: scipy>=1.15
Requires-Dist: scikit-learn
Requires-Dist: samesame[docs,test] ; extra == 'dev'
Requires-Dist: mkdocs-material ; extra == 'docs'
Requires-Dist: mkdocstrings-python ; extra == 'docs'
Requires-Dist: numpydoc ; extra == 'docs'
Requires-Dist: pandas ; extra == 'docs'
Requires-Dist: uv ; extra == 'test'
Requires-Dist: pytest ; extra == 'test'
Requires-Dist: pytest-cov ; extra == 'test'
Requires-Dist: ruff ; extra == 'test'
Maintainer: Vathy M. Kamulete
Maintainer-email: Vathy M. Kamulete <vathymut@gmail.com>
Requires-Python: >=3.12
Provides-Extra: dev
Provides-Extra: docs
Provides-Extra: test
Description-Content-Type: text/markdown

<!-- markdownlint-disable MD041 -->
<!-- markdownlint-disable MD033 -->

# samesame

<!-- badges: start -->
[![Development Status](https://img.shields.io/badge/status-early%20development-yellow)](https://github.com/vathymut/samesame)
[![Python](https://img.shields.io/badge/python-3.12+-blue.svg)](https://pypi.org/project/samesame/)
[![Downloads](https://static.pepy.tech/badge/samesame)](https://pepy.tech/project/samesame)
[![Static Badge](https://img.shields.io/badge/docs-link-blue)](https://vathymut.github.io/samesame/)
[![UAI 2022](https://img.shields.io/badge/paper-UAI%202022-yellow)](https://arxiv.org/abs/2107.02990)
[![uv](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/uv/main/assets/badge/v0.json)](https://github.com/astral-sh/uv)
[![Ruff](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/ruff/main/assets/badge/v2.json)](https://github.com/astral-sh/ruff)
<!-- badges: end -->

> Same, same but different ...

`samesame` helps you compare a source sample with a target sample.

It answers two practical questions:

- Did anything change? Use `test_shift(...)`.
- Did things get worse? Use `test_adverse_shift(...)`.

Use it for model monitoring, data validation, drift assessment, or any workflow where you need to compare two groups and determine whether the difference is practically important.

## Who is this for?

`samesame` is useful whenever you need to compare a source group and a target group, for example:

- **Model monitoring** — Does production data still look like training data?
- **Data validation** — Does this new batch look like the data I expect?
- **Drift detection** — Did something change between last month and this month?
- **Group comparison** — Do two customer groups, regions, or experiments look meaningfully different?

## Installation

```bash
python -m pip install samesame
```

## Quick Start

Suppose you already have one score per row for a source sample and a target sample.
Larger scores should indicate either worse outcomes or unusual ones.
The score usually comes from a (pre-trained) model. For example, you might train a classifier to distinguish between the source and target data, then use the predicted probabilities as scores. Or you might use a model's confidence or prediction errors as scores. The choice of score depends on your application and what kind of shift you want to detect.

```python
import numpy as np
from samesame import test_adverse_shift, test_shift

rng = np.random.default_rng(123_456)
source_scores = rng.normal(size=600)
target_scores = rng.normal(size=600)

shift = test_shift(source=source_scores, target=target_scores)
print(f"Did anything change?  p-value = {shift.pvalue:.4f}")

harm = test_adverse_shift(
    source=source_scores,
    target=target_scores,
    direction="higher-is-worse",
)
print(f"Did things get worse? p-value = {harm.pvalue:.4f}")
```

**How to read this:** a small p-value from `test_shift(...)` indicates evidence that the target sample differs from the source sample.
A small p-value from `test_adverse_shift(...)` indicates evidence that it has also shifted in a worse direction.
If the first is small and the second is large, the data changed but not in a clearly harmful way.

## How it works

`samesame` does not compare raw tables directly. The usual workflow is:

1. Turn each row into one score — typically from a classifier trained to distinguish the two groups.
2. Compare those scores with `test_shift(...)` (did anything change?) and `test_adverse_shift(...)` (did it get worse?).

Both tests are **permutation-based**, so no distributional assumptions are required.

When you know that source and target have different feature distributions — covariate shift —
you can supply per-sample importance weights to focus the test on the region where both groups
overlap. See [Adjust for covariate shift with importance weights](examples/tutorials/adjust-for-covariate-shift.md).

## Where to go next

Step-by-step examples are available in the [documentation](https://vathymut.github.io/samesame/):

**Tutorials**

- [Detect a distribution shift](https://vathymut.github.io/samesame/examples/tutorials/detect-distribution-shift/)
- [Check whether a shift is harmful](https://vathymut.github.io/samesame/examples/tutorials/check-shift-harm/)

**How-to guides**

- [Monitor a credit risk model](https://vathymut.github.io/samesame/examples/credit/monitor-credit-risk/)
- [Monitor prediction errors with per-sample scores](https://vathymut.github.io/samesame/examples/credit/monitor-prediction-errors/)
- [Monitor model confidence](https://vathymut.github.io/samesame/examples/credit/monitor-confidence-ood/)

## Dependencies

`samesame` has minimal dependencies. It is built on top of, and fully compatible with,
[scikit-learn][scikit-learn] and [numpy][numpy].

[numpy]: https://numpy.org/
[scikit-learn]: https://scikit-learn.org/stable
