visual studio code debug python with arguments

In this article, I show how to quickly and efficiently debug Python scripts using VS Code. Alternatively, we could use some fancy magic commands like %debug to open the interactive debugger window within Jupyter Notebook.Īnd the last approach is very similar to how we can debug Python scripts.

We could also put many print statements into the code just to see if everything is going as intended. In a Notebook, we could split the code into more cells and execute them sequentially to inspect each and every step. So what happens when we have a bug somewhere out there in the code? The code might work returning something (either what we wanted or not), or might not work at all and throw some error mid-execution. The entire “putting something to production” can be a topic for a series of articles or a book, so I will not go down that path, at least for now. Instead, the code from a Notebook is refactored and stored in one or multiple Python scripts, which are then scheduled to be executed. And while we can productionize a Notebook in a creative way (for example, a cronjob running a Notebook every morning), this is definitely not the way to go and is pretty much against all the best practices of the data science/software engineering craft. If you do not trust me, you can check out Jeremy Howard’s arguments in this video.Ĭoming back to the main thought, data scientists encounter bugs at all the stages of their work - from EDA and playing around with the data or some initial modeling in a Notebook to actually putting some model into production. And don’t get me wrong, the Notebooks are a great tool for data scientists. Many data scientists (especially the ones just starting their adventure) work primarily with Jupyter Notebooks. The process is by no means restricted to data science, which merely adopted it from software engineering. Debugging is the process of detecting and removing bugs (errors or unintended behavior) from any kind of code. Well… At least that is the idea, and to be honest, most often the code does not work (well or at all) on the first attempt. Most frequently, we do our best to write good and clean code, which does exactly what we want it to do.

Learn how to efficiently debug your scripts in 10 mins!

Visual studio code debug python with arguments