PyPhi

PyPhi is a Python library for computing integrated information.

To report issues, please use the issue tracker on the GitHub repository. Bug reports and pull requests are welcome.

For general discussion, you are welcome to join the pyphi-users group.

Installation

To install the latest stable release, run

pip install pyphi

To install the latest development version, which is a work in progress and may have bugs, run

pip install "git+https://github.com/wmayner/pyphi@develop#egg=pyphi"

Tip

For detailed instructions on how to install PyPhi on macOS, see the Detailed installation guide for macOS.

Note

PyPhi is only supported on Linux and macOS operating systems. Windows is not supported.

Usage and Examples

• Installation
• Getting started
• Basic Usage
• IIT 3.0 Paper (2014)
• Conditional Independence
• XOR Network
• Emergence (coarse-graining and blackboxing)
• Actual Causation
• Residue
• Magic Cuts
• Detailed installation guide for macOS

Conventions

• Transition probability matrix conventions
• Connectivity matrix conventions

Configuration

• Loading a configuration
• Approximations and theoretical options
• Parallelization and system resources
• Memoization and caching
• Logging
• Numerical precision
• The config API

API Reference

• actual
• cache
• compute
• compute.distance
• compute.network
• compute.parallel
• compute.subsystem
• conf
• connectivity
• constants
• convert
• direction
• distance
• distribution
• examples
• exceptions
• jsonify
• macro
• models
• models.actual_causation
• models.cuts
• models.mechanism
• models.subsystem
• network
• node
• partition
• subsystem
• timescale
• tpm
• utils
• validate