:rocket: UniDep - Unified Conda and Pip Dependency Management :rocket:

unidep simplifies Python project dependency management by enabling a single requirements.yaml file to handle both Conda and Pip dependencies. This streamlined approach allows for creating a unified Conda environment.yaml, while also seamlessly integrating with setup.py or pyproject.toml. In addition, it can be used as a CLI to combine multiple requirements.yaml files into a single environment.yaml file. Simplify your setup and maintain all your dependencies in one place with unidep.

:books: Table of Contents

• :package: Installation
• :page_facing_up: requirements.yaml structure
  • Example
  • Key Points
  • Platform Selectors
    • Supported Selectors
    • Usage
    • Implementation
  • Conflict Resolution
    • How It Works
• :memo: Usage
  • With pyproject.toml or setup.py
  • :memo: As a CLI
• Limitations

:package: Installation

To install unidep, run the following command:

pip install -U unidep

Or just copy the script to your project:

wget https://raw.githubusercontent.com/basnijholt/unidep/main/unidep.py

:page_facing_up: requirements.yaml structure

unidep processes requirements.yaml files with a specific format:

• name (Optional): For documentation, not used in the output.
• channels: List of sources for packages, such as conda-forge.
• dependencies: Mix of Conda and Pip packages.

Example

Example of a requirements.yaml file:

name: example_environment
channels:
  - conda-forge
dependencies:
  - numpy  # same name on conda and pip
  - conda: python-graphviz  # When names differ between Conda and Pip
    pip: graphviz
  - pip: slurm-usage  # pip-only
  - conda: mumps  # conda-only
  # Use platform selectors; below only on linux64
  - conda: cuda-toolkit  # [linux64]

⚠️ unidep can process this file in pyproject.toml or setup.py and create a environment.yaml file.

Key Points

• Standard names (e.g., - numpy) are assumed to be the same for Conda and Pip.
• Use conda: <package> and pip: <package> to specify different names across platforms.
• Use pip: to specify packages that are only available through Pip.
• Use conda: to specify packages that are only available through Conda.

Using the CLI unidep will combine these dependencies into a single conda installable environment.yaml file.

Platform Selectors

This tool supports a range of platform selectors that allow for specific handling of dependencies based on the user's operating system and architecture. This feature is particularly useful for managing conditional dependencies in diverse environments.

Supported Selectors

The following selectors are supported:

• linux: For all Linux-based systems.
• linux64: Specifically for 64-bit Linux systems.
• aarch64: For Linux systems on ARM64 architectures.
• ppc64le: For Linux on PowerPC 64-bit Little Endian architectures.
• osx: For all macOS systems.
• osx64: Specifically for 64-bit macOS systems.
• arm64: For macOS systems on ARM64 architectures (Apple Silicon).
• macos: An alternative to osx for macOS systems.
• unix: A general selector for all UNIX-like systems (includes Linux and macOS).
• win: For all Windows systems.
• win64: Specifically for 64-bit Windows systems.

Usage

Selectors are used in requirements.yaml files to conditionally include dependencies based on the platform:

dependencies:
  - some-package  # [unix]
  - another-package  # [win]
  - special-package  # [osx64]
  - pip: cirq  # [macos]
    conda: cirq  # [linux]

In this example:

• some-package is included only in UNIX-like environments (Linux and macOS).
• another-package is specific to Windows.
• special-package is included only for 64-bit macOS systems.
• cirq is managed by pip on macOS and by conda on Linux. This demonstrates how you can specify different package managers for the same package based on the platform.

Implementation

The tool parses these selectors and filters dependencies according to the platform where it's being run. This is particularly useful for creating environment files that are portable across different platforms, ensuring that each environment has the appropriate dependencies installed.

Conflict Resolution

unidep features a conflict resolution mechanism to manage version conflicts and platform-specific dependencies in requirements.yaml files. This functionality ensures optimal package version selection based on specified requirements.

How It Works

• Version Pinning Priority: unidep gives priority to version-pinned packages when multiple versions of the same package are specified. For instance, if both foo and foo <1 are listed, foo <1 is selected due to its specific version pin.

• Minimal Scope Selection: unidep resolves platform-specific dependency conflicts by preferring the version with the most limited platform scope. For instance, given foo <1 # [linux64] and foo >1, it installs foo <1 exclusively on Linux-64 and foo >1 on all other platforms. This approach ensures platform-specific requirements are precisely met.

• Resolving Intractable Conflicts: When conflicts are irreconcilable (e.g., foo >1 vs. foo <1), unidep issues a warning and defaults to the first encountered specification.

:memo: Usage

With pyproject.toml or setup.py

To use unidep in your project, you can configure it in pyproject.toml. This setup works alongside a requirements.yaml file located in the same directory. The behavior depends on your project's setup:

• When using only pyproject.toml: The dependencies field in pyproject.toml will be automatically populated based on the contents of requirements.yaml.
• When using setup.py: The install_requires field in setup.py will be automatically populated, reflecting the dependencies defined in requirements.yaml.

Here's an example pyproject.toml configuration:

[build-system]
build-backend = "setuptools.build_meta"
requires = ["setuptools", "wheel", "unidep"]

[project]
dynamic = ["dependencies"]

In this configuration, unidep is included as a build requirement, allowing it to process the Python dependencies in the requirements.yaml file and update the project's dependencies accordingly.

:memo: As a CLI

Use unidep to scan directories for requirements.yaml file(s) and combine them into an environment.yaml file. See example for more information or check the output of unidep -h:

usage: unidep [-h] [-d DIRECTORY] [-o OUTPUT] [-n NAME] [--depth DEPTH]
              [--stdout] [-v]

Unified Conda and Pip requirements management.

options:
  -h, --help            show this help message and exit
  -d DIRECTORY, --directory DIRECTORY
                        Base directory to scan for requirements.yaml files, by
                        default `.`
  -o OUTPUT, --output OUTPUT
                        Output file for the conda environment, by default
                        `environment.yaml`
  -n NAME, --name NAME  Name of the conda environment, by default `myenv`
  --depth DEPTH         Depth to scan for requirements.yaml files, by default
                        1
  --stdout              Output to stdout instead of a file
  -v, --verbose         Print verbose output

Limitations

• Conda-Focused: Best suited for Conda environments.

Try unidep today for a streamlined approach to managing your Conda environment dependencies across multiple projects! 🎉👏