Misha Mikhasenko

great. Accept PR? Something in this spirit? ```julia module Spin using SymPy import PyCall # # include("wigner_functions.jl") PyCall.pyimport_conda("sympy.physics.quantum.spin", "sympy") import_from(sympy.physics.quantum.spin, (:WignerD,), typ=:Any) # export clebsch_gordan # # include("clebshes.jl") PyCall.pyimport_conda("sympy.physics.wigner", "sympy")...

I found that `Optim` can do it ```julia f(p) = -sum(x->(x>0) ? log(x) : -Inf, notmalized_fit_function(data, p)) pfr = Optim.minimizer(Optim.optimize(f, init_pars, BFGS(), Optim.Options(show_trace = true); autodiff = :forwarddiff)) ``` perhaps,...

> Could you do > > ```julia > llh(model) = (data,p) -> - sum(log, model(data,p)) > result = curve_fit(llh(model), data, initial_p) > ``` > > ? @Magalame, that would be...

just checked that `x` vector can have a different length then `y` vector ```julia using LsqFit @. model(x, p) = p[1]*exp(-x*p[2]) xdata = range(0, stop=10, length=20) ydata = model(xdata, [1.0...

thanks for checking. The example with `fit1` and `fit2` somehow got away from the original issue. For the LLH minimization, we do not have `y` vector, I thought that `LsqFit`...

Let's check if we are talking about different things. - binned LLH, vs - unbinned LLH With your last passage on *exactly equivalent*, it seems to me that you have...

that is right. `log` was missing. Hm, I'm confused by what you said. ```julia llh(mu, sigma) = -sum(x->(x-mu)^2/sigma^2, sample) + length(sample)*log(sigma) ``` is not equivalent to minimizing ```julia lsq(mu, sigma)...

Mm, 1. think of estimating both, `mu` and `sigma` 2. in the `lsq`, the `sigma` is not the correct error for the residual of `y`. The `sigma` is the error...

Now it is getting clear for me, thanks a lot for the clarification. Could it be added to the documentation? From the outside, it looked to me that PlotlyJS is...

Also, concerning the documentation of the Plotly.jl. I did not get what the `end_point` is good for. I did not need it to post a plot (sorry for the ignorance)....