In this lecture, you will learn how to
The easiest way to get started coding in Python is by running it in the cloud.
(That is, by using a remote server that already has Python installed.)
One option that’s both free and reliable is Google Colab.
Colab also has the advantage of providing GPUs, which we will make use of in more advanced lectures.
Tutorials on how to get started with Google Colab can be found by web and video searches.
Most of our lectures include a “Launch notebook” button (with a play icon) on the top right connects you to an executable version on Colab.
Local installs are preferable if you have access to a suitable machine and plan to do a substantial amount of Python programming.
At the same time, local installs require more work than a cloud option like Colab.
The rest of this lecture runs you through the some details associated with local installs.
The core Python package is easy to install but not what you should choose for these lectures.
These lectures require the entire scientific programming ecosystem, which
Hence the best approach for our purposes is to install a Python distribution that contains
The best such distribution is Anaconda Python.
Anaconda is
Anaconda also comes with a package management system to organize your code libraries.
All of what follows assumes that you adopt this recommendation!
Anaconda supplies a tool called conda
to manage and upgrade your Anaconda packages.
One conda
command you should execute regularly is the one that updates the whole Anaconda distribution.
As a practice run, please execute the following
conda update anaconda
For more information on conda, type conda help in a terminal.
Jupyter notebooks are one of the many possible ways to interact with Python and the scientific libraries.
They use a browser-based interface to Python with
Because of these features, Jupyter is now a major player in the scientific computing ecosystem.
Here’s an image showing execution of some code (borrowed from here) in a Jupyter notebook
While Jupyter isn’t the only way to code in Python, it’s great for when you wish to
These lectures are designed for executing in Jupyter notebooks.
Once you have installed Anaconda, you can start the Jupyter notebook.
Either
jupyter notebook
If you use the second option, you will see something like this
The output tells us the notebook is running at http://localhost:8888/
localhost
is the name of the local machine8888
refers to port number 8888 on your computerThus, the Jupyter kernel is listening for Python commands on port 8888 of our local machine.
Hopefully, your default browser has also opened up with a web page that looks something like this
What you see here is called the Jupyter dashboard.
If you look at the URL at the top, it should be localhost:8888
or similar, matching the message above.
Assuming all this has worked OK, you can now click on New
at the top right and select Python 3
or similar.
Here’s what shows up on our machine:
The notebook displays an active cell, into which you can type Python commands.
Notice that, in the previous figure, the cell is surrounded by a green border.
This means that the cell is in edit mode.
In this mode, whatever you type will appear in the cell with the flashing cursor.
When you’re ready to execute the code in a cell, hit Shift-Enter
instead of the usual Enter
.
Note
There are also menu and button options for running code in a cell that you can find by exploring.
The next thing to understand about the Jupyter notebook is that it uses a modal editing system.
This means that the effect of typing at the keyboard depends on which mode you are in.
The two modes are
b
adds a new cell below the current oneTo switch to
Esc
key or Ctrl-M
Enter
or click in a cellThe modal behavior of the Jupyter notebook is very efficient when you get used to it.
Let’s run a test program.
Here’s an arbitrary program we can use: http://matplotlib.org/3.1.1/gallery/pie_and_polar_charts/polar_bar.html.
On that page, you’ll see the following code
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
plt.rcParams['figure.figsize'] = (10,6)
# Fixing random state for reproducibility
np.random.seed(19680801)
# Compute pie slices
N = 20
θ = np.linspace(0.0, 2 * np.pi, N, endpoint=False)
radii = 10 * np.random.rand(N)
width = np.pi / 4 * np.random.rand(N)
colors = plt.cm.viridis(radii / 10.)
ax = plt.subplot(111, projection='polar')
ax.bar(θ, radii, width=width, bottom=0.0, color=colors, alpha=0.5)
plt.show()
Don’t worry about the details for now — let’s just run it and see what happens.
The easiest way to run this code is to copy and paste it into a cell in the notebook.
Hopefully you will get a similar plot.
Here are a few more tips on working with Jupyter notebooks.
In the previous program, we executed the line import numpy as np
After this import command, functions in NumPy can be accessed with np.function_name
type syntax.
np.random.randn(3)
.We can explore these attributes of np
using the Tab
key.
For example, here we type np.random.r
and hit Tab
Jupyter offers several possible completions for you to choose from.
In this way, the Tab key helps remind you of what’s available and also saves you typing.
In addition to executing code, the Jupyter notebook allows you to embed text, equations, figures and even videos in the page.
For example, we can enter a mixture of plain text and LaTeX instead of code.
Next we Esc
to enter command mode and then type m
to indicate that we
are writing Markdown, a mark-up language similar to (but simpler than) LaTeX.
(You can also use your mouse to select Markdown
from the Code
drop-down box just below the list of menu items)
Now we Shift+Enter
to produce this
Debugging is the process of identifying and removing errors from a program.
You will spend a lot of time debugging code, so it is important to learn how to do it effectively.
If you are using a newer version of Jupyter, you should see a bug icon on the right end of the toolbar.
Clicking this icon will enable the Jupyter debugger.
Note
You may also need to open the Debugger Panel (View -> Debugger Panel).
You can set breakpoints by clicking on the line number of the cell you want to debug.
When you run the cell, the debugger will stop at the breakpoint.
You can then step through the code line by line using the buttons on the “Next” button on the CALLSTACK toolbar (located in the right hand window).
You can explore more functionality of the debugger in the Jupyter documentation.
Notebook files are just text files structured in JSON and typically ending with .ipynb
.
You can share them in the usual way that you share files — or by using web services such as nbviewer.
The notebooks you see on that site are static html representations.
To run one, download it as an ipynb
file by clicking on the download icon at the top right.
Save it somewhere, navigate to it from the Jupyter dashboard and then run as discussed above.
Note
If you are interested in sharing notebooks containing interactive content, you might want to check out Binder.
To collaborate with other people on notebooks, you might want to take a look at
To keep the code private and to use the familiar JupyterLab and Notebook interface, look into the JupyterLab Real-Time Collaboration extension.
QuantEcon has its own site for sharing Jupyter notebooks related to economics – QuantEcon Notes.
Notebooks submitted to QuantEcon Notes can be shared with a link, and are open to comments and votes by the community.
Most of the libraries we need come in Anaconda.
Other libraries can be installed with pip
or conda
.
One library we’ll be using is QuantEcon.py.
You can install QuantEcon.py by starting Jupyter and typing
!conda install quantecon
into a cell.
Alternatively, you can type the following into a terminal
conda install quantecon
More instructions can be found on the library page.
To upgrade to the latest version, which you should do regularly, use
conda upgrade quantecon
Another library we will be using is interpolation.py.
This can be installed by typing in Jupyter
!conda install -c conda-forge interpolation
So far we’ve focused on executing Python code entered into a Jupyter notebook cell.
Traditionally most Python code has been run in a different way.
Code is first saved in a text file on a local machine
By convention, these text files have a .py
extension.
We can create an example of such a file as follows:
%%writefile foo.py
print("foobar")
This writes the line print("foobar")
into a file called foo.py
in the local directory.
Here %%writefile
is an example of a cell magic.
If you come across code saved in a *.py
file, you’ll need to consider the
following questions:
JupyterLab is an integrated development environment built on top of Jupyter notebooks.
With JupyterLab you can edit and run *.py
files as well as Jupyter notebooks.
To start JupyterLab, search for it in the applications menu or type jupyter-lab
in a terminal.
Now you should be able to open, edit and run the file foo.py
created above by opening it in JupyterLab.
Read the docs or search for a recent YouTube video to find more information.
One can also edit files using a text editor and then run them from within Jupyter notebooks.
A text editor is an application that is specifically designed to work with text files — such as Python programs.
Nothing beats the power and efficiency of a good text editor for working with program text.
A good text editor will provide
Right now, an extremely popular text editor for coding is VS Code.
VS Code is easy to use out of the box and has many high quality extensions.
Alternatively, if you want an outstanding free text editor and don’t mind a seemingly vertical learning curve plus long days of pain and suffering while all your neural pathways are rewired, try Vim.
If Jupyter is still running, quit by using Ctrl-C
at the terminal where
you started it.
Now launch again, but this time using jupyter notebook --no-browser
.
This should start the kernel without launching the browser.
Note also the startup message: It should give you a URL such as http://localhost:8888
where the notebook is running.
Now
http://localhost:8888
) in the address bar at the top.You should now be able to run a standard Jupyter notebook session.
This is an alternative way to start the notebook that can also be handy.
This can also work when you accidentally close the webpage as long as the kernel is still running.