Chapter 32 - Python Code Analysis¶
Python code analysis can be a heavy subject, but it can be very helpful in making your programs better. There are several Python code analyzers that you can use to check your code and see if they conform to standards. pylint is probably the most popular. It’s very configurable, customizable and pluggable too. It also checks your code to see if it conforms to PEP8, the official style guide of Python Core and it looks for programming errors too.
Note that pylint checks your code against most, but not all of PEP8’s standards. We will spend a little time learning about another code analysis package that is called pyflakes.
Getting Started with pylint¶
The pylint package isn’t included with Python, so you will need to go to the Python Package Index (PyPI) or the package’s website to download it. You can use the following command to do all the work for you:
pip install pylint
If everything goes as planned, you should now have pylint installed and we’ll be ready to continue.
Analyzing Your Code¶
Once pylint is installed, you can run it on the command line without any arguments to see what options it accepts. If that doesn’t work, you can type out the full path like this:
c:\Python34\Scripts\pylint
Now we need some code to analyze. Here’s a piece of code that has four errors in it. Save this to a file named crummy_code.py:
import sys
class CarClass:
""""""
def __init__(self, color, make, model, year):
"""Constructor"""
self.color = color
self.make = make
self.model = model
self.year = year
if "Windows" in platform.platform():
print("You're using Windows!")
self.weight = self.getWeight(1, 2, 3)
def getWeight(this):
""""""
return "2000 lbs"
Can you spot the errors without running the code? Let’s see if pylint can find the issues!
pylint crummy_code.py
When you run this command, you will see a lot of output sent to your screen. Here’s a partial example:
c:\py101>c:\Python34\Scripts\pylint crummy_code.py
No config file found, using default configuration
************* Module crummy_code
C: 2, 0: Trailing whitespace (trailing-whitespace)
C: 5, 0: Trailing whitespace (trailing-whitespace)
C: 12, 0: Trailing whitespace (trailing-whitespace)
C: 15, 0: Trailing whitespace (trailing-whitespace)
C: 17, 0: Trailing whitespace (trailing-whitespace)
C: 1, 0: Missing module docstring (missing-docstring)
C: 3, 0: Empty class docstring (empty-docstring)
C: 3, 0: Old-style class defined. (old-style-class)
E: 13,24: Undefined variable 'platform' (undefined-variable)
E: 16,36: Too many positional arguments for function call (too-many-function-args)
C: 18, 4: Invalid method name "getWeight" (invalid-name)
C: 18, 4: Empty method docstring (empty-docstring)
E: 18, 4: Method should have "self" as first argument (no-self-argument)
R: 18, 4: Method could be a function (no-self-use)
R: 3, 0: Too few public methods (1/2) (too-few-public-methods)
W: 1, 0: Unused import sys (unused-import)
Let’s take a moment to break this down a bit. First we need to figure out what the letters designate: C is for convention, R is refactor, W is warning and E is error. pylint found 3 errors, 4 convention issues, 2 lines that might be worth refactoring and 1 warning. The 3 errors plus the warning were what I was looking for. We should try to make this crummy code better and reduce the number of issues. We’ll fix the imports and change the getWeight function to get_weight since camelCase isn’t allowed for method names. We also need to fix the call to get_weight so it passes the right number of arguments and fix it so it has “self” as the first argument. Here’s the new code:
# crummy_code_fixed.py
import platform
class CarClass:
""""""
def __init__(self, color, make, model, year):
"""Constructor"""
self.color = color
self.make = make
self.model = model
self.year = year
if "Windows" in platform.platform():
print("You're using Windows!")
self.weight = self.get_weight(3)
def get_weight(self, this):
""""""
return "2000 lbs"
Let’s run this new code against pylint and see how much we’ve improved the results. For brevity, we’ll just show the first section again:
c:\py101>c:\Python34\Scripts\pylint crummy_code_fixed.py
No config file found, using default configuration
************* Module crummy_code_fixed
C: 1,0: Missing docstring
C: 4,0:CarClass: Empty docstring
C: 21,4:CarClass.get_weight: Empty docstring
W: 21,25:CarClass.get_weight: Unused argument 'this'
R: 21,4:CarClass.get_weight: Method could be a function
R: 4,0:CarClass: Too few public methods (1/2)
That helped a lot! If we added docstrings, we could halve the number of issues. Now we’re ready to take a look at pyflakes!
Getting Started with pyflakes¶
The pyflakes project is a part of something known as the Divmod Project. Pyflakes doesn’t actually execute the code it checks much like pylint doesn’t execute the code it analyzes. You can install pyflakes using pip, easy_install or from source.
We will start by running pyflakes against the original version of the same piece of code that we used with pylint. Here it is again:
import sys
class CarClass:
""""""
def __init__(self, color, make, model, year):
"""Constructor"""
self.color = color
self.make = make
self.model = model
self.year = year
if "Windows" in platform.platform():
print("You're using Windows!")
self.weight = self.getWeight(1, 2, 3)
def getWeight(this):
""""""
return "2000 lbs"
As was noted in the previous section, this silly piece of code has 4 issues, 3 of which would stop the program from running. Let’s see what pyflakes can find! Try running the following command and you’ll see the following output:
c:\py101>c:\Python34\Scripts\pyflakes.exe crummy_code.py
crummy_code.py:1: 'sys' imported but unused
crummy_code.py:13: undefined name 'platform'
While pyflakes was super fast at returning this output, it didn’t find all the errors. The getWeight method call is passing too many arguments and the getWeight method itself is defined incorrectly as it doesn’t have a self argument. Well, you can actually call the first argument anything you want, but by convention it’s usually called self. If you fixed your code according to what pyflakes told you, your code still wouldn’t work.
Wrapping Up¶
The next step would be to try running pylint and pyflakes against some of your own code or against a Python package like SQLAlchemy and seeing what you get for output. You can learn a lot about your own code using these tools. pylint is integrated with many popular Python IDEs, such as Wingware, Editra, and PyDev. You may find some of the warnings from pylint to be annoying or not even really applicable. There are ways to suppress such things as deprecation warnings through command line options. Or you can use the -generate-rcfile to create an example config file that will help you control pylint. Note that pylint and pyflakes does not import your code, so you don’t have to worry about undesirable side effects.