[Code] Grepping entities

thomthom

Again - a second iteration. Of the result, but none the less - an extra iteration.

What's interesting here is what the performance is.

thomthom

I've looked st a couple of methods in Vertex Tools - where I refactored them to use grep in combo with hashes and I'm getting huge improvements!

dkendig

Glad I got frustrated on Wednesday and started throwing stuff at the wall to see what stuck! Nice benchmarks Thomthom, hope my find can help some folks out!

dkendig

Dan, did you try benchmarking your pattern? It might not have much of a performance hit if you have the extra filtering criteria in the grep block:

@unknownuser said:

Don’t forget the extra processing — a map operation — comes “free” if you provide grep with a block

Enumerating Enumerable: Enumerable#grep : Global Nerdy

Once again, it’s Enumerating Enumerable time! This is the latest in my series of articles where I set out to make better documentation for Ruby’s Enumerable module than Ruby-Doc.org’s. In this installment, I cover the grep method. In case you missed any of the previous articles, they’re listed and linked below: all? any? collect / […]

Global Nerdy (www.globalnerdy.com)

Dan Rathbun

@dkendig said:

Dan, did you try benchmarking your pattern?

No, but it can be added to TT's code above if your interested. (As TT said compact() needs to re-iterate the array to remove the nil elements. I tried using next but the block still returns a nil.)

And BTW.. I checked the C source for compact() which uses dup() to create a new array, and then calls compact!(), ie %(#008080)[rb_compact_bang()] on the C-side, so to save time, this pattern might as well use the "bang" edition of the method, and save the time to create a new array.

I'm deep into a Javascript mod for the forums, presently, and not doing much in Ruby.

thomthom

Can you post your result data as well?

thomthom

And what was the comparison code?

dkendig

drawing_ele_test time: 0.00999999046325684, resulted in 1103 elements found
types_test time: 0.00999999046325684, resulted in 1103 elements found
magic_test time: 0.0420000553131104, resulted in 1103 elements found
loop_test time: 0.0190000534057617, resulted in 1103 elements found

that's my results from the huge one-liner that I copy and pasted for a scene with 13108 total entities. 10,000 edges, 1,000 component instances, 1,000 groups, 1,000 faces. The goal was to find anything that could have a material applied to it, so I wanted to filter out edges.

So actually it appears my test was flawed. I'm adding entities to a hash, and looping through the desired types in the "types_test" results. In the "drawing_ele_test" I am just doing a comparison to see if the type matches one of the types I am looking for. They both end up taking the same exact amount of time, so there's probably a cleaner way to write this, but under the hood both of those tests are essentially doing the same thing.

dkendig

Just ran a benchmark in a scene with 10,000 edges, one component instance, one face, and one group. It appears that grepping specific classes you are interested in, is faster than doing a more general grep and then comparing within the block.

sample code soon to follow

dkendig

actually, here's the benchmark code I ran


class Magic;def initialize(&block)
	@block=block;end
	def ===(other)
		@block.call(other)
	end
end

def GetVrayImportantArrayFromGrepDrawingelement(entities)
	return entities.grep(Sketchup;;Drawingelement){|ent| false == (ent.class == Sketchup;;Edge) ? ent ; nil}.compact
end

def GetVrayImportantArrayOfTypesFromGrep(entities,typeHash)
	for curType in typeHash.keys()
		typeHash[curType] = entities.grep(curType )
	end
end

def GetVrayImportantArrayFromGrepMagic(entities)
	return entities.grep(Magic.new {|ent| ent if(ent.class == Sketchup;;ComponentInstance or ent.class == Sketchup;;Group or ent.class == Sketchup;;Face)})
end

def GetVrayImportantArrayFromLoop(entities)
	vrayImportant = Array.new()
	for ent in entities.to_a
		if ent.class == Sketchup;;Face or ent.class == Sketchup;;ComponentInstance or ent.class == Sketchup;;Group
			vrayImportant.push(ent)
		end
	end
	return vrayImportant
end

def RunBenchmarkOnEntities(entities)
	testHash = Hash.new()
	testHash[Sketchup;;Face] = []
	testHash[Sketchup;;ComponentInstance] = []
	testHash[Sketchup;;Group] = []
	
	grep_drawing_ele_result = grep_magic_result = compare_loop_result = nil
	
	grep_drawing_ele_start_time = Time.now.to_f
	grep_drawing_ele_result = GetVrayImportantArrayFromGrepDrawingelement(entities)
	grep_drawing_ele_end_time = Time.now.to_f
	
	grep_types_start_time = Time.now.to_f
	grep_types_result = GetVrayImportantArrayOfTypesFromGrep(entities,testHash)
	grep_types_end_time = Time.now.to_f
	
	grep_magic_start_time = Time.now.to_f
	grep_magic_result = GetVrayImportantArrayFromGrepMagic(entities)
	grep_magic_end_time = Time.now.to_f 
	
	compare_loop_start_time = Time.now.to_f
	compare_loop_result = GetVrayImportantArrayFromLoop(entities)
	compare_loop_end_time = Time.now.to_f
	
	puts "drawing_ele_test time; #{grep_drawing_ele_end_time - grep_drawing_ele_start_time}, resulted in #{grep_drawing_ele_result.size} elements found"
	
	puts "types_test time; #{grep_types_end_time - grep_types_start_time}, resulted in #{testHash.values.flatten.size} elements found"
	
	puts "magic_test time; #{grep_magic_end_time - grep_magic_start_time}, resulted in #{compare_loop_result.size} elements found"
	
	puts "loop_test time; #{compare_loop_end_time - compare_loop_start_time}, resulted in #{compare_loop_result.size} elements found"
	
	return nil
end

RunBenchmarkOnEntities(Sketchup.active_model.entities)