#!/usr/bin/env python
# coding: utf-8

# # Fitting
# <hr style="border-top-width: 4px; border-top-color: #34609b;">
# This ROOTbook shows how to create a [histogram](https://root.cern.ch/doc/master/classTH1F.html), [fill it](https://root.cern.ch/doc/master/classTH1.html#a77e71290a82517d317ea8d05e96b6c4a) and [fit it](https://root.cern.ch/doc/master/classTH1.html#a7e7d34c91d5ebab4fc9bba3ca47dabdd).
# 
# Let's start importing the [ROOT](https://root.cern.ch) module, which gives us access to all [ROOT](https://root.cern.ch) classes, and activate the [JSROOT](https://root.cern.ch/js/) visualisation.

# In[1]:


import ROOT


# In[2]:


get_ipython().run_line_magic('jsroot', 'on')


# Now we will create a [histogram](https://root.cern.ch/doc/master/classTH1F.html) specifying its title and axes titles. We'll also subsequently fill it with gaussian random numbers.

# In[3]:


h = ROOT.TH1F("myHisto","My Histo;X axis;Y axis",64, -4, 4)
h.FillRandom("gaus")


# We now setup our fit. ROOT provides a set of predefined functions, among which the Gaussian. You can have a look to the [TF1 class documentation](https://root.cern.ch/doc/master/classTF1.html) to learn more about how functions are implemented in ROOT. For example you will see that you can define functional forms with strings, lambda functions or other kind of functors.
# 
# We specify the option "S" to extract the information relative to the fit in a separate object: we will inspect it later.

# In[4]:


fitResultPtr = h.Fit("gaus","S")


# You can see above the output of the fit, which converged. The value, uncertainty and other useful quantities are reported for each parameter.
# 
# If more detail is needed, other information can be extracted, for example the final chi-square value or the number of free parameters:

# In[9]:


chi2_nparams = (fitResultPtr.Chi2(), fitResultPtr.NFreeParameters()) 
print "The final Chi2 value was %.2f and the number of free parameters was %d" %chi2_nparams


# So far so good. Now we can draw our fitted histogram. We have two options. 
#   1. Create a new, fresh canvas and draw the histogram in it.
#   2. Find back the canvas which was created (see the message above!) and draw it.
# The fitted function is now associated of the histogram: no special operation is needed.
# 
# Let's tackle this task following option 2. ROOT keeps track of all canvases and it is trivial to retrieve any of them by name.

# In[13]:


c1 = ROOT.gROOT.GetListOfCanvases().FindObject("c1")
c1.Draw()


# Congratulations! You accomplished your fit.