Speedup peak spreading

[Dronablo]

I tried recognize_song on different inputs and py-spy shows that most of the time was spent not on waiting response from Shazam itself, but on waiting for do_peak_spreading(). My timings was like '1.5 seconds in preak_spreading for 3 seconds of input or 6 seconds in preak_spreading for 10 seconds of input`.

So, here is a numpy implementation:

• It do exactly the same: doesn't change any logic or class variable types.
• It works ~10 times faster.
• As most of [my] numpy code it's pretty less easy readable/understandable than pure python version.

[lgtm-com[bot]]

This pull request introduces 1 alert and fixes 1 when merging ff1080f592b53115a2456b078a6b0e08bce78f68 into 3f3c7b9da2e487f4fc78335b18d75a642f973621 - view on LGTM.com

new alerts:

• 1 for Module is imported with 'import' and 'import from'

fixed alerts:

• 1 for Unnecessary pass

[dotX12]

@Dronablo, Please post tests before and after implementing the numpy library.

[Dronablo]

@Dronablo, Please post tests before and after implementing the numpy library.

Done

[lgtm-com[bot]]

This pull request introduces 1 alert when merging 4f9e4935b364857ced5e8a5b56ca144135080415 into 3f3c7b9da2e487f4fc78335b18d75a642f973621 - view on LGTM.com

new alerts:

• 1 for Module is imported with 'import' and 'import from'

[lgtm-com[bot]]

This pull request introduces 1 alert when merging a9578035f6ffd5980f02fc0f8e660a6362f5bc7b into 3f3c7b9da2e487f4fc78335b18d75a642f973621 - view on LGTM.com

new alerts:

• 1 for Module is imported with 'import' and 'import from'

[dotX12]

test:

async def tasks():
    a = [main(), main(), main(), main(), main(), main(), main()]
    res = await asyncio.gather(*a)
    return res

before: 6.474631299999601 after: 3.4023000999986834

realy good

[dotX12]

@Dronablo, Thx!

[dotX12]

https://github.com/dotX12/ShazamIO/releases/tag/0.3.0.0 Released!