Installation¶
Pyneal is built and tested using Python 3.6 in MacOS 10.13.4
Download Python from https://www.python.org/downloads/ or via a distribution like Anaconda
Pyneal requires additional libraries beyond the standard library. Instructions below use pip to install these libraries. Verify that you have pip installed:
which pip
If not, download and install pip from https://pip.pypa.io/en/stable/installing/
Download Pyneal¶
Download the Pyneal repository from GitHub, or clone the repository to your local machine:
git clone https://github.com/jeffmacinnes/pyneal.git
The installation instructions are broken down by Pyneal Scanner and Pyneal. If you haven't already, read the section on definitions, as those definitions are used throughout these instructions.
Pyneal-Scanner¶
The pyneal_scanner directory needs to be on the scanner computer. If the scanner computer is different from the analysis computer in your environment, copy the pyneal_scanner directory to the scanner computer.
dependencies¶
The scanner computer requires additional python libraries in order to run Pyneal Scanner.
You can attempt to install all required libraries at once by navigating into the pyneal_scanner directory and typing:
pip install -r requirements.txt
If that fails for any reason, you can install manually one at a time:
pip install numpy==1.17
pip install pydicom>=1.0.2
pip install nibabel>=2.1.0
pip install pyzmq>=16.0.2
pip install pyyaml>=3.12
These versions reflect the primary environment in which Pyneal is tested. It is likely that other versions maintain compatibility, but use at your own risk.
Pyneal¶
The pyneal directory needs to be on the analysis computer.
dependencies¶
Note: Pyneal relies on Nipy, which currently has compatibility issues with deprecated functions in Numpy and Scipy. Currently, the requirements files install older versions of Numpy and Scipy. If you are still experiencing errors, build the developer version of Nipy from source. Version 0.5.0.dev has been tested and works
The analysis computer requires additional python libraries in order to run Pyneal.
You can attempt to install all required libraries at once by navigating into the pyneal directory and typing:
pip install -r requirements.txt
If that fails for any reason, you can install manually one at a time:
pip install numpy==1.17
pip install scipy==1.1.0
pip install nibabel>=2.1.0
pip install nipy>=0.4.1
pip install pyzmq>=16.0.2
pip install pyyaml>=3.12
pip install wxPython>=4.0.1
pip install flask>=0.12.2
pip install flask_socketio>=2.9.2
pip install eventlet>=0.21.0
These versions reflect the primary environment in which Pyneal is tested. It is likely that other versions maintain compatibility, but use at your own risk.
Anaconda installation¶
If you are running Python via Anacona installation, you may need to run the following additional commands during installation and setup.
Prior to install¶
Create new conda environment:
conda create --name pynealEnv
conda activate pynealEnv
conda install pip
Launching Pyneal¶
Launch GUI with pythonw command
sudo pythonw pyneal.py
Additional Tools¶
Pyneal itself does not require any additional libaries beyond what is listed above. However, there are various tools included that you may find useful during a real-time scan session. For instance, the tool createMask can be used to transform a standard space ROI mask to the subject's functional space, which can then be used as a mask for analysis during a real-time scan.
In order to use these additional tools, make sure you have installed the following: