Program reads resonance data (time, w, X, Y) from stdin, fits data with a Lorentzian function:

(X + iY) = (A + iB) + (C + iD)/(w0^2 - w^2 - iw*dw)

and prints a line with 14 values to stdout:

• time -- center of the time range
• f_error -- mean square difference between data and fit
• A, A_error, B, E_error -- shift in x and y component
• C, C_error, D, D_error -- driving force (not amplitude!
• w, w_error == resonance frequency (in Hz or rad/s dependind on input data)
• dw, dw_error -- width at 1/2 height of amplitude curve, or (appoximately) distance between dispersion minimum and maximum. If frequency data is in rad/s then df=2/tau, if data in Hz then df = pi/tau

Each parameter is followed by corresponding parameter error.

The program can be used as a filter in graphene_filter script (see https://github.com/slazav/graphene/ )

Program uses libgsl library for fitting.

There is an option (turned on by default) for detecting overloaded signals. Signals are fit twice: in a usual way and with points inside largest 5% of the data range removed. The fit with smaller error is chosen.