Suggestion: add pysnooper.snoop decorator to add failing function's variables' types/values/updates to error message context

[mlaugharn]

pysnooper.snoop() is a decorator that helps automate printf style debugging: example

import pysnooper

@pysnooper.snoop()
def number_to_bits(number):
    if number:
        bits = []
        while number:
            number, remainder = divmod(number, 2)
            bits.insert(0, remainder)
        return bits
    else:
        return [0]

number_to_bits(6)

I think this will help focus/improve GPT-4's debugging ability - https://github.com/cool-RR/PySnooper

There's also torchsnooper for even better snoop insight into pytorch

[Tes-star]

Cool idea. As it increases the number of tokens it increases costs and reduces the maximal length of processed code as far as I know. Could be an additional option.

[AswanthManoj]

Does vectorizing the Python code make it easier for the LLM to understand the syntax and semantics of the code, and potentially provide better suggestions for error fixes?

[biobootloader]

thanks for the tip, I will look into PySnooper!

[biobootloader]

Does vectorizing the Python code make it easier for the LLM to understand the syntax and semantics of the code, and potentially provide better suggestions for error fixes?

GPT-4 is pretty good at understanding syntax, loops, etc. It's possible it could help, feel free to experiment and let us know!

[jakob1379]

I mean, you could simply cut out the first row with timestamps to reduce the number of token. But I would suggest making a test in all 3 cases to compare the performance