# TOMATO

A library to calculate parametric maps in MRI. It contains ShMOLLI implementation as in [this article](https://jcmr-online.biomedcentral.com/articles/10.1186/1532-429X-12-69).  
*   **Github** repository can be found [here](https://github.com/MRKonrad/tomato).  
*   **Documentation** in doxygen can be found [here](https://mrkonrad.github.io/tomato/html).

| System                                    |     Status                                                         |
|-------------------------------------------|:------------------------------------------------------------------:|
| [DOI][zenodo_link]                        | [![DOI][zenodo_badge]][zenodo_link]                                |
| [OSX/Linux build - Travis][travis_link]   | [![Travis (.org)][travis_badge]][travis_link]                      |
| [Windows build - AppVeyor][appveyor_link] | [![AppVeyor][appveyor_badge]][appveyor_link]                       |
| [Code quality - Codacy][codacy_link]      | [![Codacy Badge][codacy_badge]][codacy_link]                       |
| [Test coverage - Codecov][codecov_link]   | [![codecov][codecov_badge]][codecov_link]                          |
| [Downloads][downloads_link]               | [![GitHub Releases (by Release)][downloads_badge]][downloads_link] |

## Usage

*   App - download the latest version of `TomatoExe` from the [Releases][downloads_link] page and follow the youtube tutorial below
[![Tomato tutorial](https://img.youtube.com/vi/0tzNZNiZh18/0.jpg)](https://www.youtube.com/watch?v=0tzNZNiZh18)

*   Lib - download the latest version of `TomatoLib` from the [Releases][downloads_link] page

*   Code - if you want to build the code please take a look at the scripts in the `scriptsBuild` and the continuous integrations scripts `.appveyor.yml` and `travis.yml`

## Tomato and Numerical Recipes

**Important**  
Original ShMOLLI (and based on it Tomato) uses code based on Numerical Recipes book. Due to Numerical Recipes licence I cannot share this part of the code online. Please contact me if you would like to use this part of the code. To make up for this limitation we provide a number of alternative fitting procedures.

## Troubleshooting

*   missing msvcp140.dll  
Download <https://www.microsoft.com/en-us/download/details.aspx?id=48145> as suggested in <https://stackoverflow.com/questions/32998902/msvcp140-dll-missing>

## Software Requirements Specification

1.  Calculate ShMOLLI with exactly the same results as these from the previous implementation.
2.  API for easily choosing
    *   a map calculation algorithm,
    *   a fitting algorithm,
    *   a model function and corresponding energy function.
3.  Compatibility with:
    *   main operating systems (Windows, MacOS, Linux),
    *   C++98 (to include the library in older code).
4.  Handling DICOM files.
5.  Fast calculation.
6.  Easily interfaced to calculation environments MATLAB, IDL and Python.  

## Contributing and code of conduct

Please see [contributing.md](contributing.md) and [codeofconduct.md](codeofconduct.md).

## Notes

*   **Continuous integration** was tested [in this repo](https://github.com/MRKonrad/ContinousIntegrationPlayground).
*   **Gitmodules**: after `git clone` run `git submodule update --init --recursive` to get the submodules code.
*   It would be nice to follow [CppCoreGuidlines](https://github.com/isocpp/CppCoreGuidelines/blob/master/CppCoreGuidelines.md) as far as C++98 allows.
*   Maybe I can write a script for github releases with [github releases REST api](https://developer.github.com/v3/repos/releases/#create-a-release) and [curl REST api](http://www.codingpedia.org/ama/how-to-test-a-rest-api-from-command-line-with-curl/#12_HEAD_requests)?
*   LMFIT - somehow windows does not like the static library, so the dynamic is used

## TODO

*   add acceptance test for 2param
*   maybe rename calculatormolli to calculator3param? 
*   maybe add calculator2param to produce different results?
*   Deployment: exe and lib in one zip
*   Deployment: cmake to download ITK and lmfit
*   update comments in configuration yaml files
*   output SNR maps
*   make sure SNR and nAmoebaCalls is the same as in Shmolli
*   fix downloadITK_linux_osx
*   separate projects for building tomato dependencies (done for lmfit)
*   possibly I could make all FunctionsT1 object static
*   write user-friendly build scripts (and use them in CI)

[zenodo_link]: https://zenodo.org/badge/latestdoi/142749215
[zenodo_badge]: https://img.shields.io/badge/DOI-10.5281%2Fzenodo.1489791-blue.svg?style=flat-square
[travis_link]: https://travis-ci.org/MRKonrad/tomato
[travis_badge]: https://img.shields.io/travis/MRKonrad/tomato/master.svg?style=flat-square
[appveyor_link]: https://ci.appveyor.com/project/MRKonrad/tomato
[appveyor_badge]: https://img.shields.io/appveyor/ci/MRKonrad/tomato/master.svg?style=flat-square
[codacy_link]: https://www.codacy.com/app/MRKonrad/tomato?utm_source=github.com&amp;utm_medium=referral&amp;utm_content=MRKonrad/tomato&amp;utm_campaign=Badge_Grade
[codacy_badge]: https://img.shields.io/codacy/grade/1ca5a2f31ee040cc9258fc5018e51c1e?style=flat-square
[codecov_link]: https://codecov.io/gh/MRKonrad/tomato
[codecov_badge]: https://img.shields.io/codecov/c/github/MRKonrad/tomato.svg?style=flat-square
[downloads_link]: https://github.com/MRKonrad/tomato/releases
[downloads_badge]: https://img.shields.io/github/downloads/MRKonrad/tomato/total.svg?style=flat-square