I feel a bit frustrated in the lack of debugging tools available to us. It seems like many of my errors are the result of not being able to see the dimensions of variables and, as far as I can tell, not being able to step through the script. It seems like the print function is the only debugging tool. I understand that it is important to do some of these calculations “by hand” instead of feeding them into a library function, but it almost seems like doing this nitty gritty stuff with limited tools is not the most efficient way to learn.
It looks like there are more tools available to jupyter notebooks that are running on the local machine. Do people create clones of the classwork onto a local jupyter notebook as a way to debug? Or use some other tools?
Hi, @Jake_Corbin !
I know it can be frustrating sometimes. I personally prefer doing some pre-debugging in jupyter before writing my scripts. During all these years I find that’s the fastest way of writing code. Apart from that I can only suggest you to try some custom functions to automate that process playing with the .shape attribute.
Thanks for the advice @alvaroramajo
could you share what your pre-debugging workflow looks like?
Sure! I normally have vscode with the script I’m trying to write in python and a temporary jupyter notebook side by side. When in doubt about how a function works or what object it returns, try some dummy snippet in jupyter first (docstrings are really useful) until it is correct before copying it into python.
It really depends on the project you’re working on and its dependencies and that’s a very personal suggestion
What tools do you like the most?