[Improvement] Faster _ecg_findpeaks_elgendi

[purpl3F0x]

Description

Optimize _ecg_findpeaks_elgendi in those areas of the code:

1. blocks could be just a bool array
2. Detection of start/end of block by only comparing prev vs cur (blocks[i-1], blocks[i]) value. (about 2x faster) (even better when combined with "1")
3. Use enumerate(zip(...)) for iterating the array, (gives a small performance -but measurable- improvement)
4. Move some thresholds outisde of the loop, not to get calculated at each iteration

Checklist

Here are some things to check before creating the PR. If you encounter any issues, do let us know :)

• I have read the CONTRIBUTING file.
• My PR is targeted at the dev branch (and not towards the master branch).
• I ran the CODE CHECKS on the files I added or modified and fixed the errors.
• I have added the newly added features to News.rst (if applicable)

[DominiqueMakowski]

For some reason it fails the tests

[purpl3F0x]

It seems to be passing this time., worried-I did some copy-pasta typo on the variables. but seems ok now 🤷

EDIT: it seemes an unrelated .tar file was broken during the setup

[DominiqueMakowski]

Out of curiosity, do you have a benchmark do show how much time we shave wit the version?

[purpl3F0x]

Out of curiosity, do you have a benchmark do show how much time we shave wit the version?

I did some measurments, mainly on individual parts of the code. Don't have the excact numbers in hand, but:

• the comparison optimization is 2x faster on average&worst case since it gets narrowed down from up to 4 to up to 2.
• On using the bool array, without taking into account the reduced memory footprint + not traversing the loop for finding it's max value (why was this needed??)
  

[DominiqueMakowski]

Nice job, thanks!