View on GitHub

Python-best-practices-course

Material for the course "Python best practices", Scientific Software Center, Heidelberg University

PEP recommendations

What is a PEP?

Find out about the purpose of PEPs (Python Enhancement Proposals) on the pep site. Short answer: These are documents helping to organize the evolution of the Python language in the community.

You can find a list of PEPs here.

What is PEP 8?

PEP 8 is a meta-PEP = a PEP about other PEPs. PEP 8 specifies Python styling do’s and dont’s.

PEP 8 summary

  1. Line length
    Don’t write lines that are very long. Reason: Often you would use multiple windows next to each other and that makes it hard to read code (and possibly lead to introduction of bugs).
    PEP 8 recommends lines no longer than 79 characters, but 90 is also used quite often.

  2. Alignment of code
    Make use of indentation when using continuation lines:

     foo = long_function_name(var_one, var_two,
                              var_three, var_four)
    

    instead of

     foo = long_function_name(var_one, var_two,
         var_three, var_four)
    

    You may also use hanging indents

     foo = long_function_name(
         var_one, var_two,
         var_three, var_four)
    
  3. Break line before binary operator

     income = (gross_wages
       + taxable_interest
       + (dividends - qualified_dividends)
       - ira_deduction
       - student_loan_interest)
    

    instead of

     income = (gross_wages +
       taxable_interest +
       (dividends - qualified_dividends) -
       ira_deduction -
       student_loan_interest)
    
  4. Blank lines:
    • two blank lines after import statements
    • two blank lines between functions and classes
    • one blank line between methods of classes

            import os
      
      
            def greeting():
                print("Hello world!")
      
      
            def goodbye():
                print("See ya!")    
      

      instead of

            import os
      
            def greeting():
                print("Hello world!")
      
            def goodbye():
                print("See ya!")
      

      and

            def something():
                pass
      
      
            class MyClass:
      
                def my_method():
                    pass
      
                def some_other():
                    pass
      

    instead of

         def something():
             pass
    
         class MyClass:
             def my_method():
                 pass
             def some_other():
                 pass
    
  5. Encoding: The source file should always use UTF-8 encoding, and preferably without non-ASCII characters. It also makes your code more general if you use English words for identifiers (function names etc.).
  6. Imports should be on separate lines:

     import os
     import sys
    

    and not

     import os, sys
    
  7. String quotes: Do either use ' or " throughout your project - do not mix. For triple-quoted strings, always use ", so """ instead of '''.
  8. Whitespaces:
    • no whitespace in parantheses - do {item: 1}, don’t { item: 1 }
    • whitespace after comma - do a, b, don’t a,b, but not if followed by a parenthesis - do (a, b,) don’t (a, b, )
    • no whitespace before parenthesis - do calc_this(x) don’t calc_this (x)
    • no trailing whitespaces
    • whitespace around operators (ie. =, +, +=, ==, is, and) except for keywords - do (x=circle), don’t (x = circle)
  9. Keep comments up-to-date when changing the code. Comments should be complete sentences with a capitalized first word (unless it is an identifier). Use English language preferably.
  10. Include docstrings - see below (PEP 257).
  11. Naming conventions:
    • package and module names: short all-lowercase, preferably no underscores (do analysis.py, don’t Analysis.py; do mypackage, don’t my_package)
    • class names: use CamelCase (do MyClass, don’t myclass, Myclass, MYCLASS)
    • functions and variable names: use lowercase with underscore to increase readibility (do my_function(), don’t myfunction(), MyFunction(); do accuracy_param, don’t accuracyparam, accuracy_PARAM)
    • constants: all capital with separating underscore (do MIN_THRESHOLD, don’t Min_Threshold)
    • never use l, O, I letter names because of their similarity with 1 and 0
  12. Underscores:
    • avoid name clashes using a trailing underscore or synonym (class_)
    • single leading underscore for methods that should only be used “internally” (should not be made available to the package or class interface)
    • use two leading underscores to invoke name mangling for attributes that should not be used in subclasses of the parent class (__only_parent)
    • double leading and trailing underscores for “magic” objects (dunder methods) - __init__, __str__

Task 1a: GitHub assignment PEP right or wrong?

Please checkout the assignment repo and mark the answers in the file chapter1/PEP_right_or_wrong.md using markdown. Commit your answer to a new branch and open a pull request. If the autograding/CI checks succeed, you have marked the correct answers and can merge the pull request with main to submit your assignment.

Task 1b: GitHub assignment PEP fixme’s

Please checkout the assignment repo and correct the files chapter1/example1.py, chapter1/example2.py and chapter3/example1.py for stylistic and other errors that contradict PEP 8 recommendations. Commit your changes to a new branch and open a pull request. If the autograding/CI checks succeed, you have resolved all issues and can merge the pull request with main to submit your assignment.

What is PEP 257?

PEP 257 specifies styling of Python docstrings. Docstrings are the descriptive string literals that you should include directly after the definition of a function, class or method. This string becomes the __doc__ special attribute of that object.

For example:

def area_circ(r_in ):
    """Calculate the area of a circle with given radius.

    :Input: The radius of the circle (float, >=0).
    :Returns: The area of the circle (float)."""

Rules for docstrings:

PEP 20

The Zen of Python. Type import this in your Python console.