Math & Collisions library

[davisdude]

I have no idea why I didn't do that earlier...

[Ref]

Not too bright here but doesn't

Code: Select all

function median(tab)
	table.sort(tab)
	if #tab%2 == 0 then
		return (tab[math.floor(#tab/2)]+tab[math.floor(#tab/2+1)])/2
		else
		return tab[#tab/2+.5]
		end
	end

get you to the same place as

Code: Select all

function mlib.get_median( ... )
	name = { ... }
	for i = 1, #name do
		if #name > 2 then
			least = math.min( unpack( name ) )
			largest = math.max( unpack( name ) )
			for ii = 1, #name do
				if name[ii] == least then 
					table.remove( name, ii )
					break
				end
			end
			for ii = 1, #name do
				if name[ii] == largest then
					table.remove( name, ii )
					break
				end
			end
		elseif #name == 2 then
			median = mlib.get_mean( name[1], name[2] )
			return median
		elseif #name == 1 then
			median = name[1]
			return median
		end
	end
end

Just wondering.
Best
Edit: If you want to retain the original order of the table you could add:

Code: Select all

local tab = {unpack(tab)}

as the first line of my function

[davisdude]

Updated GitHub! Functions are now localized, refined, and get_mode returns how many times the number repeats as well.
You can also use straight tables now instead of having to do

Code: Select all

mlib.get_mode( unpack( tab ) )

you can do

Code: Select all

mlib.get_mode( tab )

[Ref]

Updated GitHub! Functions are now localized( tab )[/code]

Might want to avoid unnecessary variable creation by replacing:

Code: Select all

function mlib.get_midpoint( x1, y1, x2, y2 )
	local name = { ( x1 + x2 ) / 2, ( y1 + y2 ) / 2 }
	return name
end

with:

Code: Select all

function mlib.get_midpoint( x1, y1, x2, y2 )
	return  { ( x1 + x2 ) / 2, ( y1 + y2 ) / 2 }
end

No biggy, just a suggestion.
Best

[davisdude]

I read somewhere that it was faster if you assigned variables then called them as opposed to just calling them. I don't know if it's true or not, or if it even really impacts it that much, but...

[Hexenhammer]

I read somewhere that it was faster if you assigned variables then called them as opposed to just calling them. I don't know if it's true or not, or if it even really impacts it that much, but...

Assigning the table to a variable is completely pointless here and only makes your code more verbose. If it has any effect on performance at all it will make things slower.

[davisdude]

I read somewhere that it was faster if you assigned variables then called them as opposed to just calling them. I don't know if it's true or not, or if it even really impacts it that much, but...

Assigning the table to a variable is completely pointless here and only makes your code more verbose. If it has any effect on performance at all it will make things slower.

I guess that does make more sense, but I just thought it would make the code neater anyway...
Fixed!

[davisdude]

Sorry for double posting!

Updated! Added getPolygonArea and removed get_convex_area and get_triangle_area since they are repetitive. Check out the GitHub for the update!

By the way, just in case any of you are curious, I used the Shoelace Formula.

[davisdude]

Sorry for triple posting! (Please don't get mad! )

v. 1.4.1 is now released! You can now get the centroid of a non-intersecting polygon! mlib is also a localized variable now!
I also have a simple little demo if you want to check it out! I may expand upon it/make more if you guys like it! Just an idea...

[davisdude]

The shame! 4 consecutive posts...
Somebody stop the madness!
Anyway, 1.5.0 is out and I'm working on more functions at the moment.
As always, please make sure to vote and comment if you want any particular function/don't like how the arguments work/have any issue.
Enjoy!
(I'm working on a better demo right now, so ignore that bit for now...)