debugging.md

Debugging Python Wallaroo Applications

As you're developing Python Wallaroo applications, there will come a time where you will need to do some debugging during your development process. Debugging can be as simple as inserting a print statement in your code (we even do it in our examples) or a bit more involved by using a debugger to get backtraces and such. In this section we'll cover how to debug using print, an interactive debugger, and a remote debugger. It will help if you've gone through the Python API Introduction, Setting Up Your Environment for Wallaroo, and Writing Your Own Wallaroo Python Application sections prior to continuing so you're aware of the components that make up a Python Wallaroo application and how they interact with each other.

Debugging Using print and repr

The simplest way to do some debugging would be to include a print statement in your code to analyze data. You can see several examples of this in our reverse [example](https://github.com/WallarooLabs/wallaroo/tree/{{ book.wallaroo_version }}/examples/python/reverse/reverse.py) where we have print statements in our Encoder, Decoder, and Reverse classes. Here's an example in our compute function in our Reverse class:

def compute(self, data):
    print "compute", data
    return data[::-1]

In this example, we'll be printing the incoming data before we reverse it.

Using print is a very useful way to get logged output that you can return to and analyze after your program has completed running. However, there are some downsides when using print to debug: you'll need to add a print statement everywhere you predict you might need it, you can't get a "state of the world" look at your application, etc.

Using print in Python is not without risks. If you try to print a Unicode string, and your locale does not support Unicode, you may encounter an error like the following:

>>> u = u'\ua000abcd\u07b4'
>>> print u
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
UnicodeEncodeError: 'ascii' codec can't encode character u'\ua000' in position 0: ordinal not in range(128)

With a streaming input application, you can't always be certain about the contents of the data you try to print. However, you can effectively avoid this sort of error by printing the byte representation of your data instead, using print repr(data). In that case, the same code that resulted in an error before will provide a useful printout of the contents of the data object:

>>> u = u'\ua000abcd\u07b4'
>>> print repr(u)
u'\ua000abcd\u07b4'

Debugging Using PDB

If you need a more robust tool to do debugging, the Python standard library provides pdb. pdb is an interactive debugger which gives you the option to set breakpoints, inspect the stack frames, and other features expected from an interactive debugger. A quick example of using pdb in your application would be importing the pdb module and then calling its set_trace() function. The import pdb command goes in the module you pass to --application-module when running machida. Usage example:

import pdb
# ...
def application_setup(arg):
    pdb.set_trace()
        # ...

The above will insert a breakpoint in the current stack frame and allow you to inspect the application_setup function. pdb comes with a nice set of features so if you're interested in using it go have a look at the official documentation.

Other Debugging Options

Debugging using print or pdb is great because they're available on all platforms. However, we realize you might have your own debugging process. We encourage you to use the tools that best fit your needs and we are available to help get you set up if needed.