Accelerate Euclidean distance by numba

[dachengx]

Before you submit this PR: make sure to put all operations-related information in a wiki-note, a PR should be about code and is publicly accessible

What does the code in this PR do / what does it improve?

Nothing is changed numerically.

Can you briefly describe how it works?

Can you give a minimal working example (or illustrate with a figure)?

import os
from memory_profiler import memory_usage
import numpy as np
import numba
import matplotlib.pyplot as plt
import strax
import cutax


home = os.path.expanduser('~')
st = cutax.xenonnt_offline(xedocs_version='global_v16')
run_id = '050775'
st.storage = [strax.DataDirectory(os.path.join(home, f'{run_id}/input'), readonly=True)]

PeakletClassificationSOM0 = st.get_single_plugin(run_id, 'peaklet_classification')
peaklets = st.get_array(run_id, 'peaklets', chunk_number={'peaklets': 0}, progress_bar=False)

def compute():
    PeakletClassificationSOM0.compute(peaklets)

mem = memory_usage(proc=compute, interval=0.1, timestamps=True)
mem = np.array(mem)

fig, ax = plt.subplots(1, 1, figsize=(4, 3))
ax.plot(mem[:, 1] - mem[0, 1], mem[:, 0] / 1e3, color='black', lw=0.5)
ax.set_xlabel('Time [s]')
ax.set_ylabel('Memory usage [GB]')

fig.savefig('peaklet_classification.svg', transparent=True, dpi=800)
plt.close(fig)

The improvement is visible:

Please include the following if applicable:

• Update the docstring(s)
• Update the documentation
• Tests to check the (new) code is working as desired.
• Does it solve one of the open issues on github?

Notes on testing

• Until the automated tests pass, please mark the PR as a draft.
• On the XENONnT fork we test with database access, on private forks there is no database access for security considerations.

All italic comments can be removed from this template.

[coveralls]

coverage: 89.501% (+0.04%) from 89.463%
when pulling 9438f56 on numba_euclidean_dist
into 9813b30 on master.

[LuisSanchez25]

Looks good to me, let me just do some quick tests to make sure we get the same numbers for both methods

[LuisSanchez25]

I have a quick question since I don't seem to be able to get the function to run locally since numba seems to reject using the assert function, did you not run into this error? The error goes away when using @numba.jit(nopython=True, cache=True) instead

[dachengx]

I have a quick question since I don't seem to be able to get the function to run locally since numba seems to reject using the assert function, did you not run into this error? The error goes away when using @numba.jit(nopython=True, cache=True) instead

My gut feeling is that the discrepancy is from different numba versions. But anyway, I moved the asssert out.

[LuisSanchez25]

Ok then other than that I think everything looks good!