Pyheatmagic: Profile and view your Python code as a heat map | Hacker News Reader

Pyheatmagic: Profile and view your Python code as a heat map

github.com

98 pointsjonbaer15 comments

It would probably make more sense to link to the library [0] directly instead of the IPython module (or update the title accordingly)

[0] https://github.com/csurfer/pyheat

Profiling Python code is such a hot mess. Since Linesman is not available anymore I actually don't know a good tool to do development profiling except use cProfile and a lot of time I don't have. Production is worse. Only thing that (sort of works) is Elastic APM (RIP Opbeat which was phenomenal).

Anyone have a better option out there?

Does line_profiler not work anymore?

https://github.com/pyutils/line_profiler

I suppose it does. I miss call graphs thou.

Austin: https://github.com/p403n1x87/austin

This looks really promising.

Austin can make heat maps in VS Code :) https://marketplace.visualstudio.com/items?itemName=p403n1x8...

I've always used snakeviz with the stdlib profiler https://jiffyclub.github.io/snakeviz/

In prod, the pyinstrument profiler has worked well for me https://pyinstrument.readthedocs.io/en/latest/guide.html#pro...

py-spy with speedscope output is what I use in both prod and dev.

I haven't tried it yet, but Scalene looks interesting: https://github.com/plasma-umass/scalene.

That looks promising and I’ll give it a try. I recently used birdseye a lot to profile complex code blocks.

Btw, I took the liberty to post your project to my Python News aggregator over here: https://news.python.sc/item/99783ac6-50bc-42ea-9a48-8e66b538...

It is not explained anywhere what the heat map is of… is it the age of the line or the number of commits? Is this configurable or what?

I used this tool a few month ago, as a way to quickly profile some slow running functions in one of my notebooks. The color is mapped to the cumulative run time of the line.

This project wraps py-heat https://github.com/csurfer/pyheat in a convenient to use ipython magic command. Pyheat presents a heat map of profiler output: i.e. the output of running your program and measuring the time to execute each line of python code. Pyheat in turn uses pprofile https://github.com/vpelletier/pprofile.

I think it warrants an explaination of what py-heat is rather than assuming that the reader knows.

Reply on news.ycombinator.com