Add simple gaussian model

[hammannr]

This adds the simplest model, which can serve as a very basic example and for unittests. This closes #5.

You can run the following code to check the functionality:

from alea.examples.gaussian_model import GaussianModel
import matplotlib.pyplot as plt
import numpy as np

# Initialize model
model = GaussianModel(mu=0, sigma=1)

# Generate and set toy
data = model.generate_data()
model.set_data(data)
print(model.get_data(), model.get_data().dtype)

# Fit the data
best_fit, ll_val = model.fit(verbose=True)

# Make a plot of the ll scan
x = np.linspace(-2, 2, 1000)
y = - model.ll(x, 1)
y -=  min(y)
plt.plot(x, y)
mle = best_fit["mu"]
plt.axvline(mle, label=f"MLE={mle:.2f}", color="red")
plt.axvline(model.get_data()["hat_mu"][0],
            label=f"hat_mu={model.get_data()['hat_mu'][0]:.2f}",
            ls=":",
            color="k")
plt.axhline(.5, c="k")
plt.ylim(0)
plt.legend()
plt.xlabel("mu")
plt.ylabel("Log Likelihood Ratio")

[hammannr]

@kdund I just fixed some import issues but apart from that I'd now leave it at that so we can continue working with this. We already discussed a few changes in the structure last Friday but I think they come more natural once we have the Parameter class.

[kdund]

sounds good-- should we start making a list of all properties a parameter might have?

[hammannr]

Yes I'm now opening an issue for the Parameter and ParameterList class, maybe we can then discuss it there. But I think we can merge this PR, right?

[kdund]

yes agreed