gaussfit

Command line utility and library to decompose spectra in gaussians and lorentzians, easily used for automatic processing of multiple similar spectra. Minimum dependencies and written in C. Tested only on Linux, but should work on other POSIX-compliant systems.

Installing:

Distributions with apt (Debian, Ubuntu, Linux Mint):

Add repository:

sudo apt-get install curl gpg
echo "deb http://packages.luisvmf.com/ stable main" | sudo tee /etc/apt/sources.list.d/luisvmf.com.list
curl -fsSL https://luisvmf.com/pgp-key.public | gpg --dearmor | sudo tee /etc/apt/trusted.gpg.d/luisvmfcom.gpg > /dev/null
sudo apt-get update

Install:

sudo apt-get install luisvmf.com-gaussfit

Building from source:

        git clone https://git.luisvmf.com/gaussfit/
           cd gaussfit
        make

External dependencies:

        Standard C libs.
        C compiler and make (to build).
        gnuplot on path and /dev/shm mounted (optional, for option -p to work. If this are not present the other options will work with no modification.)

See tests.sh for examples. Example for usage in C code is in file demogaussfitapi.c.

        Usage:
         ./gaussfit [-p] [-m] [-a] [-g A=XX,sigma=XX,mu=XX] [-g A=XX,sigma=XX,mu=XX] [-f file]

           -w X  File save mode.
                     X=0-> Don't save results (default).
                     X=1-> Saves final fit parameters
                           (filename.gaussfit).
                     X=2-> Saves final fit parameters
                           and spectrum with boxcar and
                           baselined removed (filename.gaussfit
                           and filename.proc).
           -b X  Baseline polynomial order (X is an integer).
           -s X  Boxcar smoothing (X is an integer).
           -p    Plot graph and gaussians with gnuplot.
                 Gnuplot must be installed and on path
                 for this option to be used.
           -m    Manualy specify gaussians/Lorentzians initial
                 guess with option -g or -l. Must not be used
                 together with option -a.
           -a rth=0.2,tyf=1,nt=95,faf=0.95,risaf=1.1
                    Automaticaly set gaussians initial guess. Parameters:
                    tyf:0 for gaussian and 1 for lorentzian.
                    nt: Noise tolerance, larger means less sensitive to noise.
                    rth: Minimum height to consider a peak, between 0 and 1.
                    faf: Minimum fall after peak before adding another peak.
                    risaf: Minimum rise after minimum before adding another peak.
           -g A=XX,sigma=XX,mu=XX
                 Set gaussians initial guess, where XX are
                 float point values. Options -g or -l must
                 be used if option -m is used. 
           -l A=XX,sigma=XX,mu=XX
                 Set Lorentzian initial guess, where XX are
                 float point values. Options -g or -l must
                 be used if option -m is used. 
           -f file
                 Set file containing data to
                 be fitted. Must be used.