A short introduction to Python

Prof. Stefano Carrazza

Summary: In this lecture we go step by step through the basics of Python, Numpy, Matplotlib and Pydicom.

Exercise 1 - Environment preparation

For this tutorials we have setup in the laboratory machines an anaconda environment with all required packages.

In order to activate the environment open a terminal and do:

module load python3/anaconda

This will select a python 3.x interpreter from the anaconda distribution.

Note

If you are working in our own machine we suggest to install anaconda3 and then perform the following installations with:

conda install numpy matplotlib pydicom

Exercise 2 - Hello World!

Write a python script which outputs the string "Hello World!" using the built-in python function print.
Python is a dynamic typing language, allocate the following three variables 5, 6.5 and "hello" using the function input and print to terminal using f-strings.

Exercise 3 - Translating from C++

Translate the following C++ code in a python script:

#include <iostream>

int main()
{
    int a = 1, b = 1;
    int target = 48;
    for(int n = 3; n <= target; ++n)
    {
        int fib = a + b;
        std::cout << "F("<< n << ") = " << fib << std::endl;
        a = b;
        b = fib;
    }

    return 0;
}

Compare the output of the C++ code with the python code.

Exercise 4 - Python basics

In this tutorial we will use python through jupyter notebooks.

In order to start a session open a terminal and do:

jupyter notebook

This command will open a browser with an interactive python3 session.

Commands are executed cell by cell by pressing SHIFT+ENTER.

Create an empty list v and append integers from [0,15] using a for loop.
Append 3 floats to an empty list.
Create a dictionary with the following keys ('name', 'loss', 'weights'). Initialize the dictionary with the following default values ('neuralnet', 0.12, [10,25,5]). Print to screen the value of weights. Assign to loss the value 10.
Given the list [2,6,3,8,9,11,-5] compute the mean value using a for loop.
Define a list using the expression 2 ** n with n = 10.

Exercise 5 - Functions

Write a function which returns the average from a a list. Test the function using the list v=[2,6,3,8,9,11,-5].
Write the factorial function: f(0) = 1, f(x) = x * f(x-1).

Exercise 6 - Class 1

Given the base class:

class Variable:
    def __init__(self, name):
        self.name = name

    def sample(self, size):
        raise NotImplementedError()

Implement an inherited Normal class which outputs a list of normal samples [mu=0, sigma=1] by overriding the Variable.sample method.

Exercise 7 - Class 2

Construct a class which constructs and evaluates a 1D polynomial model with the following API:

the class constructor must take the polynomial degree as argument.
implement a set_parameters and get_parameters methods to update the parameter list.
provide an execute method to access the polynomial prediction at a specific value in x.

Exercise 8 - Numpy basics

Allocate an array a from the python list [[0.5, -1],[-1, 2]] with dtype float 32bit.
Verify its shape, dimension and create a deep copy b using a flatten shape.
Assign zero to elements in b with even indexes.

Exercise 9 - Numpy performance

Write a code which starting from a specific space dimension N allocates:

a random vector v of real double-precision (numpy.float64) of size N.
a random square matrix A with size NxN.
implement a function which performs the dot product using only python primitives.
measure the execution time of the previous function and compare with NumPy's dot method.
compare the performance results.

Exercise 10 - Matplotlib basics

Write a plotting script (or notebook) using Matplotlib for the function exp(-x) * cos(2*pi*x) for 100 points in x linearly spaced from 0 to 5.

Exercise 11 - Scatter plot

Download the data.dat file using:

wget https://raw.githubusercontent.com/scarrazza/ssfm_python_tutorial/main/data.dat

The file contains 2 columns (x,y) of points.

Load data using numpy and use matplotlib scatter plot for the graphical representation.
Update title with "Charged particles", axis titles with "x-coordinate" e "y-coordinate".
Color points with red.
Store plot to disk using the filename output11.png.

Exercise 12 - Plot functions

Write a python script/notebook with the following steps:

Define a function f(x) = -sin(x*x)/x + 0.01 * x*x.
Generate an array with 100 elements, linearly spaced between [-3, 3].
Write to a file output.dat the values of x and f(x) line by line.
Plot all points.
Bound the x-axis between x=[-3,3].
Add title, axis labels and a line between points, show the equation in the legend.
Store plot to disk as output12.png.

Exercise 13 - Pydicom

Write a python script/notebook which loads the DICOM test file CT_small.dcm through the method pydicom.data.get_testdata_file.

Print to screen the following information: Patient's name and ID, study date and image size.
Using matplotlib's imshow method, plot the DICOM image using the pixel_array attribute.
Store the image to disk as dicom.png.