Implementation:Scikit learn Scikit learn UpdateEnvironments

Knowledge Sources	Scikit_learn Scikit-learn Docs
Domains	Build Tools, CI/CD
Last Updated	2026-02-08 15:00 GMT

Overview

Concrete tool for updating CI environment files and associated lock files provided by scikit-learn.

Description

This script manages the CI environment configuration by updating conda environment.yml files, pip requirements.txt files, and their associated lock files (conda-lock and pip-compile). It reads minimum dependency specifications from sklearn/_min_dependencies.py and generates environment files for all CI build configurations. The script supports selective updates via regex-based build filtering and uses Jinja2 templating to generate environment files.

Usage

Use this script when bumping minimum dependency versions, pinning packages to work around regressions, or ensuring that CI uses the latest compatible versions of all dependencies.

Code Reference

Source Location

• Repository: scikit-learn
• File: build_tools/update_environments_and_lock_files.py

Signature

# Command-line script using click
# python build_tools/update_environments_and_lock_files.py [--select-build REGEX]

# Key configuration variables:
common_dependencies_without_coverage = [...]
# Uses Jinja2 Environment for template rendering
# Uses packaging.version.Version for version comparison

Import

import click
from jinja2 import Environment
from packaging.version import Version

I/O Contract

Inputs

Name	Type	Required	Description
--select-build	str (regex)	No	Regex pattern to filter which build configurations to update
sklearn/_min_dependencies.py	file	Yes	Source of minimum dependency version specifications
build_tools/*/environment.yml templates	files	Yes	Jinja2 templates for conda environment files

Outputs

Name	Type	Description
environment.yml files	YAML files	Updated conda environment configuration files
requirements.txt files	text files	Updated pip requirements files
lock files	text files	Generated conda-lock or pip-compile lock files

Usage Examples

Basic Usage

# Run from the root of the scikit-learn repository
# Update all environments:
# python build_tools/update_environments_and_lock_files.py

# Update only documentation builds:
# python build_tools/update_environments_and_lock_files.py --select-build doc

Related Pages

• Principle:Scikit_learn_Scikit_learn_Metric_Evaluation