consider various edge cases when converting osu mania maps

[orwenn22]

this pr does the following

• some maps might have an health drain value of 0, but when importing the map in fluxis, it will be set to 8. i made sure it is set to 1 instead in such cases
• in osu mania, bpm changes actually change the scroll speed. while we can barely see the difference on "regular" maps, some sv maps abuse this to get huge scroll multipliers. this pr makes it so on each bpm changes, a scroll velocity is added to match osu!stable's behaviour, and the current bpm is also taken into account for each effect points (useful reference)
• explicitly ignore p notes (which, in most cases, are creating using LNs with a start time of NaN and an n time of where the mapper want to place the p notes). there exists more complex types of pnotes, but i haven't done anything to handle these.

i tried to explain things as clearly as possible in comments to make sure anyone who doesn't know about these edge cases can understand why i did things this way, but this might feel over-commented idk

i obviously haven't tested every sv map in existence, but here are some of the most cursed maps i know converted using these changes

Hito Mania (doesn't start on the current version of the game because of p notes and cursed bpm)

hito.mp4

Digibuprofen (doesn't start on the current version of the game due to insanely high bpm) (yes this is what 1:52 IGT is supposed to look like)

digi.mp4

dont say "i can sample that" for 24 hours challenge (on the current version of the game, the bpm-only effects at 1:02 IGT don't do anything)

icansamplethat.mp4

stresstest (on the current version of the game most of the teleport effects don't work, some parts are really hard to play due to low scroll speed)

stress.mp4

closes #72 (hopefully)