PyPulseq-CEST: Simulation tool and seq-file preparation for CEST saturation blocks

This repository contains the code and tools to build CEST saturation blocks with pypulseq, which is a python adaption of the matlab-based pulseq project. Please visit https://pulseq-cest.github.io/ for more information on the pulseq-cest project. The documentation of the pulseq open file format for MR sequences can be found here. Pulseq-cest specific seq file and parameter handling is done using the BMCTool python package.

Installation

For the installation of pypulseq-CEST you need to have Git installed. If this is the case, just perform the following steps:

1) Download the code:

Clone the GitHub repository OR download the latest version as a ZIP file and unzip it into a folder.

2) Installation using setup.py file:

We provide several pre-compiled distributions of the C++ based simulation code for Linux and Windows and python versions between 3.6 and 3.9. Using these pre-compiled distributions is the recommended installation. We further recommend using a clean python/anaconda environment. Two convenient ways to start the automatic installation via setup.py file are:

Open a terminal in the pypulseq-cest folder, activate your environment and execute python setup.py
Open the pypulseq-cest folder in your favorite IDE (e.g. PyCharm) and run the setup.py file.

The installation process should install the following packages into your local python environment:

pySimPulseSBB: package containing the C++ based simulation code
pypulseq_cest: package containing the python based parser functions
BMCTool: package for pulseq-cest specific seq file and config file handling
pypulseq: package for general seq file writing/reading
all other required packages like PyYAML, numpy, matplotlib, etc...

3) Download the pulseq-cest-library:

During the installation, a folder named pulseq-cest-library should have been created in the library subfolder. If this is not the case, download the latest version as a ZIP file and add it to the library folder manually.

If you run into any troubles during installation, please refer to the Troubleshooting section below.

Quick Start

To perform an example simulation, you can run the following code:

from pypulseq_cest.simulate import sim_example
sim_example()

As an alternative, we provide a quick_start.py script. The individual steps/lines are explained below.:

Import the simulate function

from pypulseq_cest.simulate import simulate

Define a config and a sequence file
```
sim_config = 'pypulseq_cest/example_library/config_example.yaml'
seq_file = 'pypulseq_cest/example_library/seq_example.seq'
```
For more information about the config and sequence files, check the Configuration and sequence file library section below or visit the pulseq-cest-library Github repository.
Start the simulation
```
sim = simulate(config_file=sim_config, seq_file=seq_file, show_plot=True, normalize=True)
```
The simulate function requires a config_file and seq_file argument. The show_plot argument (default=False) allows to toggle a plotting functionality. Furthermore, the simulate function accepts several additional keyword arguments (**kwargs), that allow to adjust the generated plot. These are for example normalize (bool: toggle normalization), norm_threshold (value/list/array: threshold for normalization offsets), offsets (list/array: manually defined x-values), invert_ax (bool: toggle invert ax), plot_mtr_asym (bool:toggle plot MTR_asym) and title, x_label, y_label to control the lables.
Additional data processing:

The simulate functions returns a SimulationParameters object that allows further processing of the simulated data. For example, you can retrieve the final magnetization vector of the water pool using:
```
m_out = sim.GetFinalMagnetizationVectors()
```
Afterwards, you can retrieve (and print) the offset values and the z-magnetization of the water pool using:
```
offsets, mz = get_zspec(m_out=m_out, sp=sim, seq_file=seq_file)
print(f'The offsets used for simulation are: \n{offsets}')
```

Configuration and sequence file library

All simulations in pypulseq-cest require a yaml file that includes all simulation settings and a seq file, which defines the pre-saturation block. An example seq-file, and an example yaml file file can be found in the library subfolder.

You will find further pre-defined and approved pre-saturation schemes and simulation configs in the pulseq-cest-library If you have not successfully used the above installation, please download it from the pulseq-cest-library repository.

Project structure

This project tree is supposed to clarify the project components, their functionality and the source.

pypulseq-cest
|
| -- src: source code and code handling
|     |
|     | -- compile (C++, SWIG): C++ code for simulation handling (SimPulseqSBB), 
|     |     |                   SWIG for direct wrapping (pySimPulseqSBB)
|     |     |
|     |     | -- Eigen (C++): library for matrix handling in C++
|     |     | -- dist: built distributions of pySimPulseqSBB for different environments
|     |     |
|     |     - BlochMcConnellSolver.h, ExternalSequence.cpp, ExternalSequence.h, SimPulseqSBB.cpp, 
|     |       SimPulseqSBB.h, SimulationParameters.cpp, SimulationParameters.h, 
|     |       SimPulseqSBBTemplate.h (C++): simulation handling (SimPulseqSBB)
|     |     - pySimPulseqSBB.i, numpy.i, eigen.i (SWIG): SWIG wrapper handling
|     |     - setup.py: installation of pySimPulseqSBB, the direct wrapped C++ code
|     |  
|     - readme.md: manual installation guide
|     - setup_pypulseq_cest.py: installation of the pypulseq_cest parser package
|     - MANIFEST.in: defining the pypulseq_cest package for setup_pypulseq_cest
|
| -- pypulseq_cest (python): package with parser functions for use of pySimPulseqSBB in python
|     |  
|     | -- example_library (seq, yaml): input files for an exemplary simulation
|     |     |
|     |     - config_example.yaml: example file for simulation configurations
|     |     - seq_example.seq: example sequence in the Pulseq format
|     |  
|     - parser.py: functions to parse simulation parameters to pySimPulseqSBB (SWIGged C++)
|     - simulate.py: pythonic implementation of the simulation function to simplify parsing
|
| (--) after successful installation: pulseq-cest-library (python, matlab, seq, yaml): contains
|      additional config and sequence files --> visit https://pulseq-cest.github.io/ for more info
|
- readme.md, CONTRIBUTING.md, LICENSE.md: project information
- setup.py: complete installation (pypulseq_cest, pySimPulseqSBB, dependencies)
- simulate.py: exemplary simulation script

Installation Troubleshooting

Permission problems:

To avoid permission problems, you can run the setup with administrative rights:

Windows: start the terminal with administrative rights
Linux: depending on the environment you want to install into, use sudo or hand the --user flag:
```
sudo python setup.py
```
or
```
python setup.py --user
```
You might need to adapt your python executable, e.g. python3

System independent installation using setup.py file:

If your first try to install pypulseq-cest via setup.py file didn't work, your system configuration is probably not included in the pre-compiled distributions we provide. For this case we provide a system independent installation. To run this, please ensure that the following prerequisites are fullfilled:

The BMCTool package has to be installed. This can be done by running pip install bmctool. Using pip ensures that the required pypulseq package and pyYaml package are installed as well.
The following programs have to be installed on your computer:
- Git
- SWIG (Installation for Windows)
- a working C++ compiler
  - for Linux, this comes with your operating system
  - for Windows, you need Visual C++ v.12.0 or higher, e.g. Microsoft Visual C++ Redistributable

If this is the case, simply re-run the setup.py file. This should execute all required steps including the compilation and installation of the C++ based code.

In the installation still didn't work, please try this Manual Installation Guide.

FAQ: I can't get any of this to work, what shall I do?

An alternative, pure python based tool that can handle the same input is the bmctool. You will find the same functionalities, but with lower performance (simulation speed).

KerstinKaspar / pypulseq-cest