Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Reelya]

Yeah, I already accepted the limitation of convex polygons. I'm just rejecting any input that isn't convex. It's really confusing to visualize how to determine which vertices from polygon A should be placed between which ones from polygon B, and how I should be altering the coordinates to account for the new vertices I shoved in between.

Unless I horribly, horribly misunderstood what a Minkowski sum is, or what your task is, you don't need to put the angles in the same list. You just need to add the vertices.

Well you're not wrong about what a Minkowski sum is, what you haven't looked into is the linear time algorithm for generating the Minkowski set for two convex polygons, which is what Gatleos is trying to code.

@Gatleos: having thought about it, what I think the algorithm is doing is you take each edge of both polygons, then you always pick the next one with a slope/angle which is the closest possible one to the last one, but in a clockwise direction. Then append that onto your line, and the end result is that you get back where you started, with the completed Minkowski polygon for two convex polys. Each edge from your starting collection should be used only once in the final output. Just look at the next segment from each poly, and pick the one that's closest to the current segment's angle. Once you've exhausted all points, you have your poly. To get the offsets, just find any point where segments from each poly meet, then add their offsets, then you can propagate these around the edges.

[Shadowlord]

So when you said some of your polygons were clockwise and some were anti-spinward counterclockwise, I'm wondering: Is there a reason for that?

[Reelya]

That's just a data entry issue. You can do the cross-product check (in 3d space) to know to wind the polys the other way if needed.

[miauw62]

You messed up the comparison in said for loop.

Yeah whoops. Not sure how I didn't notice that... I just ended up switching it out for forEach loops. Still getting used to javascript.

Code: [Select]

for(var i = 0; i++; i >= pipes.length - 1)
Just looks wrong to me. For loops in JavaScript aren't any different to in c/c++ (except for "var"). The comparison is wrong, and it's in the wrong spot.

The comparison is wrong, yeah, but how is it in the wrong spot?

[TheBiggerFish]

for(initializing;check this condition;what happens at the end)
You switched up the last two.

[MagmaMcFry]

Here's how to fix it:

Code: [Select]

for(var i = 0; i < pipes.length; ++i)


[miauw62]

Eh, whoops. Haven't coded in a while >.< Already using forEach() anyway. (Still thanks for reminding me of the correct syntax though)

Now facing a more confounding problem, namely, this:



E:

the solution is pretty obvious again, i was using absolute coordinates in a function that called for (x, y, width, height) (but i wanted to share the bug anyway because it looks funny)

[Avis-Mergulus]

Spoiler (click to show/hide)

Python problem. Suppose I have a dictionary defined as such:

Code: (Python) [Select]

table = {}
table["Key"] = ("String",[(True,"Alice"),(False,"Charlie"),(True,"Bob")])
What's the best or most efficient way to check if a given name exists or not in a tuple of the list table["Key"][1], then remove the tuple that contains it?

I would do this:

Code: [Select]

table['Key'][1] = [tuple for tuple in table['Key'][1] if name not in tuple]


Hmm... that doesn't look too efficient (as the real list might be much larger.) It also needs to know which names were found. I should've mentioned I'm actually trying to find a number of names, which might make some kind of set intersection more ideal.
I tried this:

Code: (Python) [Select]

table = {}
table["Key"] = ("String",[(True,"Alice"),(False,"Charlie"),(True,"Bob")])

found=[]
not_found=[]
names = ["Alice","Bob","Dave"]
for tuple in table["Key"][1]:
if tuple[1] in names:
found.append(tuple)
else:
not_found.append(tuple)
table["Key"][1] = not_foundUnfortunately:

Code: [Select]

Traceback (most recent call last):
  File "<pyshell#7>", line 1, in <module>
    table["Key"][1] = not_found
TypeError: 'tuple' object does not support item assignmentNot sure what the deal is there. It seems the outer-most tuple doesn't want me swapping out its list for another. Swapping the outer tuple out for a list works, and should be functionally identical. Or I could replace the entire tuple:

Code: [Select]

table["Key"] = (table["Key"][0],not_found)Not sure which is better.

Your problem here is that, just as the traceback days, tuples don't support item assignment, or anything else for that matter. They're not modifiable. So maybe just use a list?

[Gatleos]

So when you said some of your polygons were clockwise and some were anti-spinward counterclockwise, I'm wondering: Is there a reason for that?

Yeah, it's a data-entry thing. These polygons are data imported from Tiled and used for collision mapping. I'm getting the Minkowski Sum of an (arbitrarily-sized) AABB and a completely arbitrary polygon pulled from the collision map, so that I can perform predictive collision testing.

[Bumber]

Your problem here is that, just as the traceback days, tuples don't support item assignment, or anything else for that matter. They're not modifiable. So maybe just use a list?

I ended up just leaving the dictionary unaltered and adding each found name to a set, which I could then check names against for membership without worrying about tuples. Sets use hashing, so that part should be efficient. It looks something like this now:

Code: (Python) [Select]

found = []
not_found = []
intersect = set()
for tuple in table['Key'][1]:
if tuple[1] in names:
found.append(tuple)
intersect.add(tuple[1])
else:
not_found.append(tuple)
for name in names:
if name not in intersect:
#name exists only in names
for tuple in found:
#tuples with name in names
for tuple in not_found:
#remaining tuples

[Mephisto]

I don't know what you're doing with that code but would something like this help?

Code: [Select]

found = filter(lambda x: x[1] in names, table['key'][1])
not_found = set(table['key'][1]).difference(found)


[Bumber]

I don't know what you're doing with that code but would something like this help?

Code: [Select]

found = filter(lambda x: x[1] in names, table['key'][1])
not_found = set(table['key'][1]).difference(found)


That definitely looks better. I'll have to see what I can do with it. It's for this (down in the output_merge function.)

[Elephant Parade]

Working on a Python script for a number-heavy forum game I'm planning to run. It crashes, though, and I can't quite figure out why; any help would be appreciated!

Code: (getstats.py) [Select]

### IMPORTS ###

#currently none

### FUNCTIONS ###

def getstats(list):
    HPF = int(list[1] * (list[0] + 5))
    SPF = int(list[2] * (list[0] + 5))
    ATKF = int(list[3] * (list[0] + 5))
    DEFF = int(list[4] * (list[0] + 5))
    SPAF = int(list[5] * (list[0] + 5))
    SPDF = int(list[6] * (list[0] + 5))
    SPEF = int(list[7] * (list[0] + 5))

    print("HP: " + str(HPF))
    print("SP: " + str(SPF))
    print("ATK: " + str(ATKF))
    print("DEF: " + str(DEFF))
    print("SPA: " + str(SPAF))
    print("SPD: " + str(SPDF))
    print("SPE: " + str(SPEF))

### OTHER DEFINITIONS ###

loop = 1 #this is what makes the loop work

### THE LOOP ###

while loop != 0:

    Name = str(input("NAME: "))
    info = list(input("LV, HP, SP, ATK, DEF, SPA, SPD, SPE: "))
    print()
    print(Name)
    getstats(list)   
    print()


Code: [Select]

Traceback (most recent call last):
  File "C:/Users/.../stats-list.py", line 36, in <module>
    getstats(list)
  File "C:/Users/.../stats-list.py", line 8, in getstats
    HPF = int(list[1] * (list[0] + 5))
TypeError: 'type' object is not subscriptable
The old version didn't use a list; instead, you had to type bases in one at a time. That was dumb, so I changed it to accept a list, but I apparently messed it up somehow?

That, or I'm entering the list wrong. I looked it up online, and (if I read correctly) you need to enter integers one at a time, each separated from the last by a single space. Is that correct?

[TheBiggerFish]

Which forum game?

[Elephant Parade]

Which forum game?

Working on a Python script for a number-heavy forum game I'm planning to run.

It doesn't exist have a thread yet, but it's going to be a tile-based Pokemon strategy game; it draws pretty heavily from FEF (and Pokemon, obviously), but it should still be unique enough to be its own thing.

edit: fixed silly phrasing