* added notebook to the link of examples

* Update uplift.pyx with isinstance(v, numbers.Number):

…n UpliftRandomForestClassifier (#850)

…ndomForestRegressor (#852)

* Add outcome vector y preparation for multiple treatment groups
* Update CausalTreeRegressor class
* Add multiple groups support in cython part of causal trees building
* Add min_group_size parameter to control tree building
* Update python part of causal trees, keep consistent namings for tree builder
* Add an option to pass custom matplotlib axes in charts
* Update Jupyter notebooks with causal trees and forests

* Keep consistent codestyle with black

* Update validate_data arguments support for sklearn <1.6 & >=1.6

* Extend causal tree and forest tests for multiple treatment groups

* Keep consistent codestyle with black

* Fix description in causal trees notebok

* Keep consistent codestyle in tests

* Remove unused cython var in criterion header

* Add separate function for check_y_params

)

* fix: remove sklearn.utils._random import to avoid DEFAULT_SEED signature mismatch

- Copy our_rand_r and RAND_R_MAX implementations locally
- Avoids sklearn 1.6+ DEFAULT_SEED const qualifier change
- Maintains BSD-3-Clause license compatibility

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* fix: support scipy>=1.16.0 by removing sklearn internal dependency

Resolves #859

- Remove sklearn.utils._random import from causalml/inference/tree/_tree/_utils.pyx
- Copy our_rand_r and RAND_R_MAX implementations locally with BSD-3-Clause attribution
- Support scipy>=1.16.0, numpy>=1.25.2, statsmodels>=0.14.5
- Requires Python>=3.11
- Fixes TypeError with sklearn.utils._random.DEFAULT_SEED signature mismatch

The root cause was that Cython auto-imports ALL symbols when using cimport,
including DEFAULT_SEED which had a signature change in sklearn 1.6+. By
copying the needed functions locally, we eliminate this dependency and
ensure compatibility with current sklearn versions.

Tested with: Python 3.11.9, sklearn 1.7.0, scipy 1.17.0, numpy 2.1.3
All 109 tests passing.

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* ci: update Python test matrix to 3.11+ only

Remove Python 3.9 and 3.10 from test matrix to match updated
requires-python>=3.11 requirement.

* docs: add Python 3.11 ReadTheDocs environment file

Add environment-py311-rtd.yml for building docs with Python 3.11
and updated dependency versions:
- scipy>=1.16.0
- numpy>=1.25.2
- scikit-learn>=1.6.0
- statsmodels>=0.14.5
- cython==3.0.11

* docs: update ReadTheDocs config to use Python 3.11 environment

Switch from environment-py39-rtd.yml to environment-py311-rtd.yml
to match updated Python 3.11+ requirement.

* style: apply black 26.1.0 formatting

- Upgrade black from >=25.1.0 to >=26.1.0 in pyproject.toml
- Apply black 26.1.0 formatting to 13 files
- Fixes CI lint errors

* docs: fix copyright year to range 2019-2026
* fix: replace force_all_finite with ensure_all_finite for sklearn 1.6+

sklearn 1.6+ renamed the parameter from force_all_finite to ensure_all_finite.
Update _classes.py to use the new parameter name.

* Fix pandas 2.x + numpy 2.x string dtype compatibility

Replace dtype == "object" checks with is_numeric_dtype() checks to handle
both legacy object dtype and new str dtype introduced in pandas 2.x with
numpy 2.x.

This fixes test failures in CI where pd.options.future.infer_string=True
causes string columns to have dtype 'str' instead of 'object', making
the previous dtype checks fail to detect categorical columns.

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* Fix LabelEncoder to work with pandas 2.x string dtype

When pandas uses the new string dtype (instead of object), we cannot
assign numeric values to string-typed columns. LabelEncoder converts
categorical strings to numeric labels, so we need to:

1. Copy the input DataFrame to avoid modifying the input
2. Explicitly convert encoded values to float dtype with .astype(float)

This fixes the test_LabelEncoder failure in CI where pandas 2.x with
numpy 2.x uses dtype 'str' instead of 'object' for string columns.

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* update the readthedocs config to use the latest Ubuntu and miniforge to fix the build timeout
* update the docs config, and dependencies
* update readthedocs conda env file
* fix lint errors

---------

Co-authored-by: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

…hon version requirement (#864)

The cibuildwheel configuration was attempting to build wheels for Python 3.9 and 3.10,
but the package requires Python >=3.11 (as specified in requires-python). This mismatch
caused the Ubuntu packaging workflow to fail during validation.

Changes:
- Updated [tool.cibuildwheel] build list to only include cp311-* and cp312-*
- Removed cp39-* and cp310-* from the build list

Fixes #863

Co-authored-by: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* Fix Ubuntu packaging failure by aligning cibuildwheel config with Python version requirement

The cibuildwheel configuration was attempting to build wheels for Python 3.9 and 3.10,
but the package requires Python >=3.11 (as specified in requires-python). This mismatch
caused the Ubuntu packaging workflow to fail during validation.

Changes:
- Updated [tool.cibuildwheel] build list to only include cp311-* and cp312-*
- Removed cp39-* and cp310-* from the build list

Fixes #863

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* Pin scipy to <1.17.0 to fix Ubuntu packaging failure

The Ubuntu packaging build was failing because scipy 1.17.0 (released
January 2026) only provides manylinux_2_27/manylinux_2_28 wheels, which
are incompatible with the manylinux2014 (manylinux_2_17) containers used
by cibuildwheel. This caused pip to fall back to building scipy from
source, which failed due to missing OpenBLAS dependency.

By pinning scipy to <1.17.0, we ensure pip only selects versions that
have compatible manylinux2014 wheels (like scipy 1.16.3), preventing the
source build fallback and resolving the build failure.

Fixes #863

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* Address CoPilot review comments - ensure consistency across config files

1. Update scipy constraint in RTD environment file
   - Pin scipy to <1.17.0 in docs/environment-py311-rtd.yml
   - Prevents manylinux2014 compatibility issues during doc builds

2. Update GitHub Actions workflows for Python 3.11+ requirement
   - Remove Python 3.9 and 3.10 from test-build-from-source.yml
   - Remove Python 3.9 and 3.10 from test-pypi-install.yml
   - Aligns with requires-python = ">=3.11" in pyproject.toml

3. Explicitly set manylinux image in cibuildwheel config
   - Add manylinux-x86_64-image = "manylinux2014"
   - Add manylinux-aarch64-image = "manylinux2014"
   - Ensures consistent manylinux level across all Python versions

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

---------

Co-authored-by: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

Changes:
- Removed macos-13 from build matrix (GitHub has deprecated this runner)
- Upgraded cibuildwheel from v2.22 to v3.3.1 (latest as of Jan 2026)

cibuildwheel v3.0+ changes:
- Requires Python 3.11+ to run (compatible with CausalML's requirements)
- Dropped support for building Python 3.6/3.7 wheels (not needed)
- Added CPython 3.14 support

Co-authored-by: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

…onstraints (#869)

* Upgrade to manylinux_2_28 and remove scipy version constraints

This PR modernizes CausalML's wheel distribution by upgrading from
manylinux2014 to manylinux_2_28, enabling compatibility with both
scipy 1.16.x and 1.17.x without version pinning.

## Changes

### 1. Upgrade manylinux platform tag (pyproject.toml)
- **Before:** manylinux2014 (glibc 2.17, CentOS 7 base - EOL June 2024)
- **After:** manylinux_2_28 (glibc 2.28, modern standard)

### 2. Remove scipy version constraints
- **pyproject.toml:** `scipy>=1.16.0,<1.17.0` → `scipy>=1.16.0`
- **docs/environment-py311-rtd.yml:** Removed upper bound constraint
- **Benefit:** Support both scipy 1.16.x and 1.17.x automatically

### 3. Document system requirements (docs/installation.rst)
Added new "System Requirements" section documenting:
- Python 3.11+ requirement
- Minimum Linux distributions (Ubuntu 18.04+, RHEL 8+, Debian 10+)
- Build-from-source instructions for older systems

## Why This Change?

### Problems with manylinux2014:
- Based on CentOS 7 (EOL June 2024)
- Required version pinning to avoid scipy 1.17.0
- Prevented access to newer dependency features
- Maintenance burden with explicit constraints

### Benefits of manylinux_2_28:
- Modern, actively maintained standard
- Compatible with scipy 1.16.x AND 1.17.x
- No version pinning needed - pip selects the best version
- Future-proof for upcoming dependencies
- Cleaner dependency declarations

## Compatibility Impact

### Systems That Work:
- Ubuntu 18.04 LTS+ (glibc 2.27+)
- RHEL/CentOS 8+ (glibc 2.28+)
- Debian 10+ (glibc 2.28+)
- All recent macOS and Windows versions

### Systems Requiring Source Build:
- RHEL/CentOS 7 (glibc 2.17)
- Ubuntu 16.04 and earlier
- Debian 9 and earlier

**Note:** CentOS 7 reached EOL in June 2024, and Ubuntu 16.04 EOL
was April 2021, making this upgrade aligned with industry standards.

## scipy Compatibility

CausalML uses these scipy modules:
- scipy.sparse, scipy.stats
- scipy.optimize (fsolve, minimize)
- scipy.special (expit, logit)
- scipy.interpolate (UnivariateSpline)

**None of these are affected by scipy 1.17.0 breaking changes**, which
only impact scipy.spatial.transform. Both scipy 1.16.x and 1.17.x work
correctly with CausalML.

## Testing

Pre-built wheels will now use manylinux_2_28. Users can install with
either scipy version:
- scipy 1.16.3 (stable, has manylinux2014 wheels)
- scipy 1.17.0+ (latest, has manylinux_2_28 wheels)

pip will automatically select the appropriate version based on the
user's system capabilities.

Supersedes: #868 (build-system fix no longer needed with manylinux_2_28)
Closes: #863

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* Fix glibc version requirement: manylinux_2_28 requires glibc 2.28, not 2.27

Copilot correctly identified that Ubuntu 18.04 is NOT compatible with
manylinux_2_28 wheels. Ubuntu 18.04 has glibc 2.27, but manylinux_2_28
requires glibc 2.28 or later.

Changes:
- Updated glibc requirement from 2.27 to 2.28
- Changed minimum Ubuntu version from 18.04 to 20.04 LTS (has glibc 2.31)
- Added Ubuntu 18.04 to the list of distributions requiring source build

This ensures users have accurate information about system requirements.

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

* Bump version to 0.16.0 and update changelog for breaking changes

This release introduces breaking changes in Linux wheel compatibility
due to the manylinux_2_28 upgrade, warranting a minor version bump
from 0.15.6 to 0.16.0.

Changes:
- Updated version in pyproject.toml: 0.15.6 → 0.16.0
- Added comprehensive 0.16.0 changelog entry documenting:
  - Breaking change: manylinux_2_28 requirement (glibc 2.28+)
  - Affected systems and migration path
  - scipy version pin removal
  - Related PRs (#869, #867, #865, #864)

Breaking Changes:
- Pre-built wheels require Ubuntu 20.04+, RHEL 8+, Debian 10+ (glibc 2.28+)
- Users on Ubuntu 18.04, RHEL 7, etc. must build from source
- Python 3.11+ required (already enforced in previous release)

This follows semantic versioning: minor version bump for backward-
incompatible changes to wheel distribution.

Co-Authored-By: Claude (claude-sonnet-4-5) <noreply@anthropic.com>

---------

Co-authored-by: Claude (claude-sonnet-4-5) <noreply@anthropic.com>