Procedural programming in python

Topics

• Flow control, part 1
  • If
  • For
    • range() function
• Some hacky hack time
• Exercises

Review of Data Types

Flow control

Flow control refers how to programs do loops, conditional execution, and order of functional operations. Let's start with conditionals, or the venerable if statement.

Let's start with a simple list of instructors for these classes.

If

If statements can be use to execute some lines or block of code if a particular condition is satisfied. E.g. Let's print something based on the entries in the list.

Usually we want conditional logic on both sides of a binary condition, e.g. some action when True and some when False

There is a special do nothing word: pass that skips over some arm of a conditional, e.g.

Note: what have you noticed in this session about quotes? What is the difference between ' and "?

Another simple example:

It is always good practice to handle all cases explicity. Conditional fall through is a common source of bugs.

Sometimes we wish to test multiple conditions. Use if, elif, and else.

Conditionals can take and and or and not. E.g.

For

For loops are the standard loop, though while is also common. For has the general form:

for items in list:
    do stuff

For loops and collections like tuples, lists and dictionaries are natural friends.

You can combine loops and conditionals:

Dictionaries can use the keys method for iterating.

range()

Since for operates over lists, it is common to want to do something like:

NOTE: C-like
for (i = 0; i < 3; ++i) {
    print(i);
}

The Python equivalent is:

for i in [0, 1, 2]:
    do something with i

What happens when the range you want to sample is big, e.g.

NOTE: C-like
for (i = 0; i < 1000000000; ++i) {
    print(i);
}

That would be a real pain in the rear to have to write out the entire list from 1 to 1000000000.

Enter, the range() function. E.g. range(3) is [0, 1, 2]

Notice that Python (in the newest versions, e.g. 3+) has an object type that is a range. This saves memory and speeds up calculations vs. an explicit representation of a range as a list - but it can be automagically converted to a list on the fly by Python. To show the contents as a list we can use the type case like with the tuple above.

Sometimes, in older Python docs, you will see xrange. This used the range object back in Python 2 and range returned an actual list. Beware of this!

Remember earlier with slicing, the syntax :3 meant [0, 1, 2]? Well, the same upper bound philosophy applies here.

This would probably be better written as

But in all, it isn't very Pythonesque to use indexes like that (unless you have another reason in the loop) and you would opt instead for the instructor in instructors form.

More often, you are doing something with the numbers that requires them to be integers, e.g. math.

For loops can be nested

Note: for more on formatting strings, see: https://pyformat.info

You can exit loops early if a condition is met:

You can skip stuff in a loop with continue

There is a unique language feature call for...else

You can iterate over letters in a string

Exercise

Objective: Replace the bash magic bits for downloading the Pronto data and uncompressing it with Python code. Since the download is big, check if the zip file exists first before downloading it again. Then load it into a pandas dataframe.

Notes:

• The os package has tools for checking if a file exists: os.path.exists

  import os
  filename = 'pronto.csv'
  if os.path.exists(filename):
    print("wahoo!")

• Use the requests package to get the file given a url (got this from the requests docs)

  import requests
  url = 'https://s3.amazonaws.com/pronto-data/open_data_year_two.zip'
  req = requests.get(url)
  assert req.status_code == 200 # if the download failed, this line will generate an error
  with open(filename, 'wb') as f:
    f.write(req.content)

• Use the zipfile package to decompress the file while reading it into pandas

  import pandas as pd
  import zipfile
  csv_filename = '2016_trip_data.csv'
  zf = zipfile.ZipFile(filename)
  data = pd.read_csv(zf.open(csv_filename))

Now, use your code from above for the following URLs and filenames

What pieces of the data structures and flow control that we talked about earlier can you use?

Functions

For loops let you repeat some code for every item in a list. Functions are similar in that they run the same lines of code for new values of some variable. They are different in that functions are not limited to looping over items.

Functions are a critical part of writing easy to read, reusable code.

Create a function like:

def function_name (parameters):
    """
    optional docstring
    """
    function expressions
    return [variable]

Note: Sometimes I use the word argument in place of parameter.

Here is a simple example. It prints a string that was passed in and returns nothing.

To call the function, use:

print_string("Dave is awesome!")

Note: The function has to be defined before you can call it!

If you don't provide an argument or too many, you get an error.

Parameters (or arguments) in Python are all passed by reference. This means that if you modify the parameters in the function, they are modified outside of the function.

See the following example:

def change_list(my_list):
   """This changes a passed list into this function"""
   my_list.append('four');
   print('list inside the function: ', my_list)
   return

my_list = [1, 2, 3];
print('list before the function: ', my_list)
change_list(my_list);
print('list after the function: ', my_list)

Variables have scope: global and local

In a function, new variables that you create are not saved when the function returns - these are local variables. Variables defined outside of the function can be accessed but not changed - these are global variables, Note there is a way to do this with the global keyword. Generally, the use of global variables is not encouraged, instead use parameters.

my_global_1 = 'bad idea'
my_global_2 = 'another bad one'
my_global_3 = 'better idea'

def my_function():
    print(my_global)
    my_global_2 = 'broke your global, man!'
    global my_global_3
    my_global_3 = 'still a better idea'
    return

my_function()
print(my_global_2)
print(my_global_3)

In general, you want to use parameters to provide data to a function and return a result with the return. E.g.

def sum(x, y):
    my_sum = x + y
    return my_sum

If you are going to return multiple objects, what data structure that we talked about can be used? Give and example below.

Parameters have four different types:

Play around with the above function.

Functions can contain any code that you put anywhere else including:

• if...elif...else
• for...else
• while
• other function calls

How would you functionalize the above code for downloading, unzipping, and making a dataframe?

Once you have some code that is functionalized and not going to change, you can move it to a file that ends in .py, check it into version control, import it into your notebook and use it!

Homework: Save your functions to pronto_utils.py. Import the functions and use them to rewrite HW1. This will be laid out in the homework repo for HW2. Check the website.