MINOTAUR v2.1.0

Update compared to version 2.0:
Data can be scale such that their average and standard deviation equals 0 and 1 respectively. This may avoid over-weighting issues in the MCMC. The scaling is done independently for the intensities and each type of rates.

v2.0.0

Here are some improvements compared to version 1:

1. when simulating the shuttling periods, MINOTAUR divides the sample trajectory into time-increment that are optimized (see original publication for further details). Since the sample speed for shuttling up or down are potentially different, the initial approach was to generate time lists at which the propagator is calculated for both shuttling times (up and down). The current approach no longer optimizes the increment time but a distance increment. The trajectories are divided into similarly to both shuttling periods, and the delay time between each point is calculated independently, depending on the shuttling speed up and down. Thus, only one set of relaxation matrices are needed for both trajectories. The distance increment is optimized similarly to version 1. The starting distance increment is 1 cm. This lead to a time saving close to a factor 2.
2. the shuttling points at which the relaxation matrix is computed are identical for each experiments, only the delay lists are adapted. Thus, we reduce the number of relaxation matrices being calculated to its bare minimum, speeding up the calculations even further.
3. MINOTAUR calculations can be started from the command line (detailed below).
4. better figure qualities are generated.
5. a wider range of file format are accepted (columns separated by tabs, comma and spaces, or combination of these are now accepted).

reference:
Nicolas Bolik-Coulon, Milan Zachrdla, Guillaume Bouvignies, Philippe Pelupessy, Fabien Ferrage, Comprehensive analysis of relaxation decays from high-resolution relaxometry, J. Magn. Reson. 355 (2023) 107555