I am having troubles testing if the recent changes in https://github.com/catalystneuro/roiextractors/pull/343 work well with the data that I got from Clandinin. The problem is that the data takes too long too load (around 20 minutes). I profiled and the bottleneck is xml parsing.
This PR uses xpath from lxml instead of the standard parser from the library. There is a speed up of around ~100 times
The firs run in the screenshot is using this PR and the other one is using the base branch here:
I am having troubles testing if the recent changes in https://github.com/catalystneuro/roiextractors/pull/343 work well with the data that I got from Clandinin. The problem is that the data takes too long too load (around 20 minutes). I profiled and the bottleneck is xml parsing.
This PR uses xpath from lxml instead of the standard parser from the library. There is a speed up of around ~100 times
The firs run in the screenshot is using this PR and the other one is using the base branch here: