2010-12-12T00:02:10 *** Azrathud has quit IRC (Quit: Leaving) 2010-12-12T00:13:19 *** yasith has quit IRC (Ping timeout: 240 seconds) 2010-12-12T00:17:59 * antimatroid likes list comprehensions 2010-12-12T00:18:13 def make_map(X, Y): 2010-12-12T00:18:14 return [ [0 for y in range(Y)] for x in range(X) ] 2010-12-12T00:18:14 that's just cool 2010-12-12T00:23:04 Frontier: what are you using as a shift for greek? 2010-12-12T00:33:00 :) 2010-12-12T00:33:06 yeah, list comprehensions are great 2010-12-12T00:34:42 i'm not sure how to then iterate over my lists though :\ 2010-12-12T00:35:53 yeah, you end up having lists of lists 2010-12-12T00:36:26 you can do 'for yrow in map: for y in yrow:' 2010-12-12T00:36:53 or you can do 'for x in range(len(map)):' at the top level to have an index 2010-12-12T00:37:16 yeah got it, but how do i print a list on one line? 2010-12-12T00:37:29 print map 2010-12-12T00:37:35 if i print the list i get brackers, if i iterate through, it adds enter after every element 2010-12-12T00:37:51 put a , at the end of the print statement to suppress the newline 2010-12-12T00:37:58 ah, cool :) 2010-12-12T00:38:00 so print "no newline", 2010-12-12T00:38:11 (i've only ever used python for these map engines) 2010-12-12T00:38:25 i remember you said that 2010-12-12T00:39:20 got it, cheers 2010-12-12T00:47:47 antimatroid: I use Alt Gr as the modifier in my 'normal' layout to access the Greek symbols. Shift is additional to access the upper case letters 2010-12-12T00:49:53 Frontier: cheers, i might look for another app later, i don't even know how to tell the other one to map to greek 2010-12-12T00:50:01 it wants me to push a button to map to ehe 2010-12-12T00:50:14 i'm making an ant map generator atm though 2010-12-12T00:50:24 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-12-12T00:55:26 *** Utkarsh has joined #aichallenge 2010-12-12T00:57:10 can i typecast to int from float? 2010-12-12T01:00:18 *** Meatkat has quit IRC (Quit: Leaving) 2010-12-12T01:02:35 antimatroid: float(x) 2010-12-12T01:03:07 got it, and was the other way oops :) 2010-12-12T01:08:11 Zannick: if you want to iterate over a list with an index, the more "pythonic" way is: for i,x in enumerate(xs) 2010-12-12T01:08:41 ah, right 2010-12-12T01:14:44 *** Rubicon-|-Cross has joined #aichallenge 2010-12-12T01:14:47 i should have the map generator ready in a couple of hours (i'm being lazy about it), hardest is going to be "check map is valid" function when adding in walls 2010-12-12T01:15:04 but i should just be able to scan out with a stack till i find the other base 2010-12-12T01:18:51 *** amriedle has joined #aichallenge 2010-12-12T01:38:57 *** amriedle has quit IRC (Quit: Lost terminal) 2010-12-12T01:40:04 *** bhasker has quit IRC (Quit: bhasker) 2010-12-12T01:49:39 can i do switches in python? 2010-12-12T01:50:11 if/elif is about it 2010-12-12T01:50:37 hmmmm, so much for python being known for neatness :P 2010-12-12T01:50:46 though you can do crazier constructs like dictionary mappings to functions 2010-12-12T01:51:27 do you think ants should know how big the world is when they start? 2010-12-12T02:29:48 *** Cyndre has quit IRC (Read error: Connection reset by peer) 2010-12-12T02:30:03 *** Cyndre has joined #aichallenge 2010-12-12T02:33:55 *** Utkarsh has quit IRC (Ping timeout: 264 seconds) 2010-12-12T02:38:56 *** Utkarsh has joined #aichallenge 2010-12-12T02:42:50 *** reportingsjr has left #aichallenge 2010-12-12T02:43:02 *** Epona has joined #aichallenge 2010-12-12T02:43:07 anyone on? 2010-12-12T02:43:58 yeah, what's up? 2010-12-12T02:44:53 just got pissed off in the python channel 2010-12-12T02:45:45 what are you up to 2010-12-12T02:45:59 writing a map generator for a possible version of the ants game 2010-12-12T02:46:13 ah 2010-12-12T02:46:27 I was thinking of solving my multiprocessing problem 2010-12-12T02:46:34 someone made an interesting idea 2010-12-12T02:46:55 to just code for multiple while statements, in just one while statement 2010-12-12T02:47:20 but 2010-12-12T02:47:27 *** Appleman1234 has quit IRC (Ping timeout: 276 seconds) 2010-12-12T02:47:34 making multiple infinate loops would be nicer 2010-12-12T02:55:31 i think this ant game can actually work really well 2010-12-12T03:00:01 I can't wait to see it? 2010-12-12T03:00:03 ! 2010-12-12T03:00:43 i'm still fine tuning how to make maps 2010-12-12T03:00:47 then i'll get to work on an engine 2010-12-12T03:00:54 i'm going to need someone else to do a gui though 2010-12-12T03:01:05 but hopefully a prototype can be going some time 2010-12-12T03:08:49 print "noWalls: " + noWalls 2010-12-12T03:08:54 how do i do that properly in python? 2010-12-12T03:12:58 print "noWalls:", noWalls 2010-12-12T03:13:07 or print "noWalls: " + str(noWalls) 2010-12-12T03:14:32 *** Palmik has joined #aichallenge 2010-12-12T03:23:57 *** virdo has quit IRC (Ping timeout: 265 seconds) 2010-12-12T03:27:16 *** virdo has joined #aichallenge 2010-12-12T03:36:40 *** Mathnerd314 has quit IRC (Ping timeout: 250 seconds) 2010-12-12T03:38:04 antimatroid: hay... havn't read anything on the forums... just coding up a ant version now. Without locked combat (really don't like that) 2010-12-12T03:38:57 delt0r_: i've almost got a map generator for you 2010-12-12T03:39:11 what format are you using 2010-12-12T03:39:25 have you read about the home hole idea? 2010-12-12T03:39:31 i really like that, and it provides a spawn point 2010-12-12T03:39:40 no 2010-12-12T03:39:50 won't till this version is done. 2010-12-12T03:39:52 you guys should set up a git repo so that you can stay in sync 2010-12-12T03:40:12 simple idea is that once we have something runing we can try different rulesets 2010-12-12T03:40:22 sigh: yea 2010-12-12T03:40:28 http://codepad.org/FiEkYOYZ 2010-12-12T03:40:31 that's the map gen so far 2010-12-12T03:40:31 once i have a basic game i may do 2010-12-12T03:40:50 gah hang on, it timed out when running 2010-12-12T03:41:20 http://codepad.org/Cx2KgcNU 2010-12-12T03:41:34 there is an example map at the bottom of that page (i need to work out why that timed out too) 2010-12-12T03:42:03 a and b represent the home holes, # walls and numbers how much grain can be harvested from that square 2010-12-12T03:45:00 cool 2010-12-12T03:45:10 format is about the same as what i was doing too 2010-12-12T03:45:39 i'm going to run some ideas by you for this first implementation 2010-12-12T03:47:17 an ant can: move, harvest, deposit,give (grains to friend bot), fight on a given turn 2010-12-12T03:47:17 an ant is generated at the home home (by queen bee) if there is at least 1 grain on the home hole square, the hole starts with 5 grains 2010-12-12T03:47:17 an ant can fight an enemy ant that is within x distance, if they land on the same square, they automatically fight to the death 2010-12-12T03:47:26 this is pretty amazing looking, antimatroid 2010-12-12T03:47:35 an ant can only do one of those things in a turn 2010-12-12T03:47:51 and friend ants can be in the same square 2010-12-12T03:48:25 I still don't like overlaping ants... and locked combat... it pulls the dynamics down too much 2010-12-12T03:48:33 you lose chooice 2010-12-12T03:48:52 you can completely remove forced combat 2010-12-12T03:49:24 but i don't know how you handle them landing in the same square and resolving who gets to harvest if not enough is left 2010-12-12T03:49:35 just no overlaps 2010-12-12T03:49:44 ie nothing on the same square 2010-12-12T03:49:50 it's hard to resolve that 2010-12-12T03:49:55 no its not 2010-12-12T03:49:55 and spawning is very hard 2010-12-12T03:50:02 its about 4 lines of code 2010-12-12T03:50:35 it also means people are going to "waste" turns by accident 2010-12-12T03:50:39 I also am still doing the single process per player 2010-12-12T03:50:55 i prefer separate processes 2010-12-12T03:51:05 its ai challenge... stupid ants will lose 2010-12-12T03:51:05 maybe i'll write a second engine, and we can compare :) 2010-12-12T03:51:31 well I still thing when we what 4000 entries to all have a few 100 games... 2010-12-12T03:51:54 having ~16+ process per game is going to be hard work 2010-12-12T03:52:04 and thats not many 2010-12-12T03:52:22 I would prefer lots of ants to a game with only 2 or 3 2010-12-12T03:52:55 getting to work at one game at a time is not the same as working on the main server 2010-12-12T03:53:19 Imagine a tcp version --each process with a separate socket? 2010-12-12T03:54:22 delt0r_: the idea is to have each process fed the game state sequentially, and each ant given a set amount of time to return their move 2010-12-12T03:54:36 then process them as though they were made simultaneously 2010-12-12T03:54:49 yea--and how small can you make that relaibly... and what about tcp? 2010-12-12T03:54:51 and KILL each nonused process, resuming it on its turn 2010-12-12T03:55:11 Still a *lot* of overhead... these are not threads 2010-12-12T03:55:12 i don't know how tcp works sorry :P 2010-12-12T03:55:31 well currently it would requre a new tcp connection per process 2010-12-12T03:55:33 it sucks if all ants have the same info 2010-12-12T03:55:49 may be unavoidable 2010-12-12T03:56:00 ie every new ant would need seconds to connect (tcp takes a 4 way ping for a connection) 2010-12-12T03:56:17 antimatroid: they still only get relative info for where they are 2010-12-12T03:56:38 yes, but you can't stop people from programming the ants to share information 2010-12-12T03:56:43 or can you? 2010-12-12T03:57:05 if it can be done, it doesn't work to say "don't do that" 2010-12-12T03:57:09 but feel free to do the 100 process version... will work fine locally... but good luck supporting that once you get 1000+ entries and everyone is complaining about not getting any games... or debug data 2010-12-12T03:57:21 does it really matter.... 2010-12-12T03:57:27 if they share... 2010-12-12T03:57:38 Think enders game... 2010-12-12T03:57:40 no, my point is one process means we have to go with that 2010-12-12T03:57:44 think cyber ants 2010-12-12T03:57:53 thy will have wireless 2010-12-12T03:57:53 i do think we should use strings now 2010-12-12T03:58:01 but i think it should be private information to the program 2010-12-12T03:58:08 people watching games don't get to see strings 2010-12-12T03:58:08 yes! 2010-12-12T03:58:25 simpler to implement and manage from the organisers point of view. 2010-12-12T03:58:53 okay, well i'm still going to write my own engine, but i'll do it for single process 2010-12-12T03:59:01 and we can see what works best once they're done 2010-12-12T03:59:01 I think a single process per player is by far the best way... from that aspect 2010-12-12T03:59:18 are you going to use python? 2010-12-12T03:59:32 yes 2010-12-12T03:59:40 i don't think communication is needed 2010-12-12T03:59:49 ?? 2010-12-12T04:00:03 if people are handling all their bots at once, they all have each others information, you can just pass them the game state for each ant 2010-12-12T04:00:12 yes 2010-12-12T04:00:16 for all ants* updated on the one map 2010-12-12T04:00:20 so, the bot can only use the methods there? 2010-12-12T04:00:36 Epona: ? 2010-12-12T04:00:57 Im just not sure I understand how this code works yet 2010-12-12T04:01:32 I know the idea of "agents" is nice --buts its very artificial from a application perspective --and i think a lot of overhead from a game engine perspective 2010-12-12T04:01:33 it's not even written, patience 2010-12-12T04:02:17 i still don't like moves being blocked 2010-12-12T04:03:10 its important from visulisation, debug, and simple symatics behavour... 2010-12-12T04:03:24 wehn you have 4 ants from 2 teams what happens? 2010-12-12T04:03:37 fight two weakest or two strongest first 2010-12-12T04:03:41 easyer to avoid the problem alltogether i think 2010-12-12T04:03:51 you still have that problem now 2010-12-12T04:04:03 what if you ants attack an ant with 1 hp left? 2010-12-12T04:04:15 and then there is all the support from post of "WTF" --where players don't understand these rules 2010-12-12T04:04:15 two* ants attack 2010-12-12T04:04:26 the ant dies 2010-12-12T04:04:28 what? 2010-12-12T04:04:40 and you waste 2hp? or does attack not cost you anything? 2010-12-12T04:04:52 the attack cost nothing 2010-12-12T04:05:11 i don't expend bullets with a club ;) 2010-12-12T04:05:20 well it cost movement 2010-12-12T04:06:09 i would process fight before moves, so if you find an enemy in the required range, they are still there when it's processed 2010-12-12T04:06:24 yea 2010-12-12T04:06:47 fight, remove the dead, move, resolve collisons... done 2010-12-12T04:09:04 lets just get the dam things runing 2010-12-12T04:09:19 think it will become clear whats stupid and what not at that point 2010-12-12T04:09:55 *** Appleman1234 has joined #aichallenge 2010-12-12T04:26:11 antimatroid: if you allow stacking... how to ants "see" this? 2010-12-12T04:26:31 what do you mean? 2010-12-12T04:26:51 well I use a simple map format like you do 2010-12-12T04:27:00 so you can't see stacked ants 2010-12-12T04:27:09 for updats that is 2010-12-12T04:27:12 updates 2010-12-12T04:27:31 you have to give people both ant location and hp? 2010-12-12T04:27:47 are you doing that as a-z and A-Z? 2010-12-12T04:27:54 that could work 2010-12-12T04:27:57 just location 2010-12-12T04:28:01 no hp 2010-12-12T04:28:24 prefer to be case insensitive 2010-12-12T04:28:27 the other option is to have them take lines like 2010-12-12T04:28:27 A 1 4 0 0 2010-12-12T04:28:37 ant, player 1, 4 hp, x, y 2010-12-12T04:28:41 then they fill info in 2010-12-12T04:29:07 W 1 1 2010-12-12T04:29:07 R 6 0 0 2010-12-12T04:29:08 etc. 2010-12-12T04:29:19 and you just pass them relevant information 2010-12-12T04:29:50 yea --but if you have a lot of walls... mmm don't know whats best here... 2010-12-12T04:30:05 whats easiest to code? for a player? I think a grid perhaps? 2010-12-12T04:30:36 atm i would be planning a map, but not storing ants on there at all 2010-12-12T04:30:43 they will be their own queue 2010-12-12T04:30:46 or vector 2010-12-12T04:31:03 they are... but "combining" them is not difficult 2010-12-12T04:31:27 yeah, it will probably end up a vector > or something 2010-12-12T04:31:31 stabbing quires on a sparse grid is O(1) 2010-12-12T04:31:50 ants shouldn't be able to see through walls either, around is fine 2010-12-12T04:31:56 it's easy to work out information for that 2010-12-12T04:31:59 thats hard 2010-12-12T04:32:04 no its not 2010-12-12T04:32:07 its hard becuase what is around etc 2010-12-12T04:32:09 code it then 2010-12-12T04:32:15 i will 2010-12-12T04:32:29 and watch the 100 of people claiming that its all wrong with edge cases 2010-12-12T04:32:54 you just add each ants position to a stack, and take a blank new one, paint out till you can't do any more within vision radius, then return the map with those positions updated 2010-12-12T04:33:35 blank new one while looking at walls on original map 2010-12-12T04:33:55 los even on a grid is a pain and somewhat aritarty too (antialisaed vision?) 2010-12-12T04:34:26 ? 2010-12-12T04:34:28 especially with single cell blockers 2010-12-12T04:34:45 i'll write the code and we can see how horrible it is 2010-12-12T04:35:06 def distance(loc1, loc2): 2010-12-12T04:35:07 d1 = abs(loc1[0] - loc2[0]) 2010-12-12T04:35:07 d2 = abs(loc1[1] - loc2[1]) 2010-12-12T04:35:07 return min(d1, dim[0]-d1) + min(d2, dim[1]-d2) 2010-12-12T04:35:07 that's my distance function with wrapped edges, is that right? 2010-12-12T04:35:32 without counting walls 2010-12-12T04:35:45 also the amount of "unkown" info is now much larger and harder too track 2010-12-12T04:42:16 *** tapwater has quit IRC (Quit: tapwater) 2010-12-12T04:43:36 *** Appleman1234 has quit IRC (Ping timeout: 260 seconds) 2010-12-12T04:57:41 ha, interesting idea for turbo charged tron, each square is randomly assigned a number, the number on the square you just landed on is the number of moves you are allowed to make up to, you have to move at least 1 square 2010-12-12T05:05:34 *** Appleman1234 has joined #aichallenge 2010-12-12T05:14:25 delt0r_: i have another idea to simplify, bots can't be on the same square, but have the home squares, whenever an ant harvests something, that automatically means the queen will generate another ant next turn the home base is empty 2010-12-12T05:14:37 or even require harvest to be returned 2010-12-12T05:14:51 yea 2010-12-12T05:15:13 like RTS games... harvesting can be immediate or the peons need to return it 2010-12-12T05:15:38 and queen either has hp 1 or whatever, and if they kill your queen, you lose 2010-12-12T05:15:40 I like the idea of return it... but then the bar for a basic entry is getting set a bit higher... 2010-12-12T05:15:45 yep 2010-12-12T05:15:58 i like it, and it means you need to do an updated version of shortest path back to base 2010-12-12T05:16:01 then there is the idea of protecting the queen 2010-12-12T05:16:04 i'd put in return to begin with 2010-12-12T05:16:19 yeah, home base should be important 2010-12-12T05:16:24 *** yasith has joined #aichallenge 2010-12-12T05:16:33 the idea is to start them a decent distance apart so they can sort of set their own strategies up 2010-12-12T05:16:42 then they'll start finding each other and the battles can behin 2010-12-12T05:16:46 begin* 2010-12-12T05:17:26 i think i might make all maps square, so i can do more symmetries easily, atm their home base pretty much has to be 1 of 3 known locations 2010-12-12T05:18:11 with a square board i can make that more like 7 i think 2010-12-12T05:22:15 *** choas has joined #aichallenge 2010-12-12T05:40:10 *** boegel has joined #aichallenge 2010-12-12T05:48:16 *** boegel has quit IRC (Quit: Leaving) 2010-12-12T06:09:59 *** delt0r___ has joined #aichallenge 2010-12-12T06:12:13 *** delt0r_ has quit IRC (Ping timeout: 276 seconds) 2010-12-12T06:21:23 stupid bugs make me want to cry 2010-12-12T06:21:34 i can't get this is_valid function working :\ 2010-12-12T06:21:44 you seem to be hacking at this game pretty hard 2010-12-12T06:24:03 i want to work it out 2010-12-12T06:24:12 but this is just the map generator 2010-12-12T06:24:42 if only there was a simpler language 2010-12-12T06:24:59 i don't get to choose the language for map gen 2010-12-12T06:25:10 simpler? 2010-12-12T06:26:14 more graphic would be nicer 2010-12-12T06:26:25 what do you mean? 2010-12-12T06:26:36 sigh: feel like reading some python code for me? :P 2010-12-12T06:26:40 like most video games have really powerful engines 2010-12-12T06:26:51 sure, what do you want me to look for? 2010-12-12T06:27:13 http://codepad.org/n7rOdVXG 2010-12-12T06:27:40 that's the code, function is_valid on line 95 fails 2010-12-12T06:28:05 why does the second and third print statement have nothing after it 2010-12-12T06:28:08 i know i should start from 0, but even if i run it like is, it outputs the path it's trying to follow each step, and doesn't follow all directions 2010-12-12T06:28:11 what does it do wrong? 2010-12-12T06:28:13 ive never seen that 2010-12-12T06:28:29 run it 2010-12-12T06:28:29 mapgen.py >> map.txt 2010-12-12T06:28:33 and watch the output 2010-12-12T06:28:44 you will have to cancel the process most likely 2010-12-12T06:29:11 you only follow 3 directions, is that right? 2010-12-12T06:29:22 i = 1,2,3 2010-12-12T06:29:29 it's meant to search out from home1 looking for home2, but it only ever does 2 directions 2010-12-12T06:29:33 and it should do 3 2010-12-12T06:30:18 it should also do all 3 additions to the map before checking the stack for which lock to look from next 2010-12-12T06:30:22 which it doesn't do :\ 2010-12-12T06:30:28 antimatroid? 2010-12-12T06:30:56 Epona: it's giving me an \n character 2010-12-12T06:30:59 eh? how do you know when you;ve visited a node? 2010-12-12T06:31:10 because I have nMap with 0 and 1's 2010-12-12T06:31:13 1 means i have visited it 2010-12-12T06:31:19 oh, that or should be an and too 2010-12-12T06:31:24 that was just me fucking around 2010-12-12T06:31:29 elif map[nLoc[0] ][nLoc[1] ] != '#' and nMap[nLoc[0] ][nLoc[1] ] == 0: 2010-12-12T06:31:37 ah, that makes more sense :) 2010-12-12T06:31:52 it still doesn't work then though :\ 2010-12-12T06:32:45 it wraps all 4 edges properly i'm pretty sure 2010-12-12T06:34:53 ok, i'll try running it 2010-12-12T06:35:40 it will infinite loop, so don't leave it running for ages :) 2010-12-12T06:36:44 doesn't inf loop for me 2010-12-12T06:36:50 ever? 2010-12-12T06:36:55 what if you put in i=0? 2010-12-12T06:37:11 ok, I got an inf loop now 2010-12-12T06:37:40 it can depend on which symmetry it picks as to whether you hit the other home or not 2010-12-12T06:38:28 line 70 2010-12-12T06:38:44 change to: nLoc = list(loc) 2010-12-12T06:38:50 does that copy the reference? 2010-12-12T06:39:17 yes, you aren't creating a new list atm 2010-12-12T06:40:15 i think that did it 2010-12-12T06:40:16 thanks 2010-12-12T06:40:51 *** narnach has joined #aichallenge 2010-12-12T06:40:51 *** narnach is now known as narnach_mb 2010-12-12T06:41:37 if you comment out lines 98 and 111 you will get maps generated 2010-12-12T06:45:22 http://codepad.org/IpXK4CYh 2010-12-12T06:45:26 there is an example generated map 2010-12-12T06:45:31 deltor___: ^^ 2010-12-12T06:46:40 they actually start quite close, a is at the end of line 19, b at the start of 24 2010-12-12T06:46:49 why do you only go though 3 directions? 2010-12-12T06:46:59 i am meant to do 4 2010-12-12T06:47:08 that was just testing 2010-12-12T06:47:22 that link above^^ does 4 2010-12-12T06:47:39 it has a 32x32 grid for you too 2010-12-12T06:49:16 a choice needs to be made 2010-12-12T06:49:16 1. all square grids with more symmetries 2010-12-12T06:49:16 2. 3 symmetries and rectange grids 2010-12-12T06:53:51 delt0r___: when you pass us the game state, can you distinguish a square that is empty from one that is not currently seeable? 2010-12-12T06:54:11 I am not doing "sight" 2010-12-12T06:54:22 but yes you would have to do that 2010-12-12T06:54:57 otherwise all clients will need to dupilicate the los test so i can "know" about what it can't see 2010-12-12T06:57:57 yep :) 2010-12-12T07:02:50 http://codepad.org/VoUkxUsO 2010-12-12T07:02:51 that's a much more dense map 2010-12-12T07:03:34 maybe a little too dense, i think quite a few regions got cut off, but players home bases are guarnteed to be reachable 2010-12-12T07:03:38 from each other 2010-12-12T07:04:48 antimatroid: since you were complaining about lack of switch earlier... here's how I would do the direction stuff: http://codepad.org/cG4Tqm56 2010-12-12T07:05:05 as a bonus it makes it really easy to add more directions (like NE) 2010-12-12T07:06:58 cheers 2010-12-12T07:09:49 i think i might be able to make it so every square is accessible 2010-12-12T07:09:55 i will try that and see how it looks 2010-12-12T07:10:06 every non-wall square that is 2010-12-12T07:10:43 yeah, that's simple with the code you've already got 2010-12-12T07:10:56 might take slightly longer, but doesn't really matter 2010-12-12T07:11:10 other option is to fill blocked off regions before adding in grains 2010-12-12T07:11:12 not much longer 2010-12-12T07:11:43 that way it'd be more like working your way around underground passages? 2010-12-12T07:12:29 if the walls are placed randomly, you aren't going to have many blocked off areas 2010-12-12T07:12:40 unless you are placing a lot more walls than now 2010-12-12T07:13:45 i think dead end are good? 2010-12-12T07:14:08 i'll try the first one and see how it looks 2010-12-12T07:14:36 dead ends are fine, but that doen't have much to do with unreachable regions 2010-12-12T07:16:26 yep 2010-12-12T07:16:41 would just require larger maps for similarly interesting maps 2010-12-12T07:20:14 *** mceier has joined #aichallenge 2010-12-12T07:26:05 bah, i'm getting some weird indent error 2010-12-12T07:26:10 despite the indenting being fine 2010-12-12T07:31:11 where atr 2010-12-12T07:31:20 lemme see all you have I might be able to work it 2010-12-12T07:34:28 keyword: might* 2010-12-12T07:40:19 http://codepad.org/DmnK5T95 2010-12-12T07:40:27 Epona: what do you think of that now? 2010-12-12T07:40:34 sigh: ^^ 2010-12-12T07:41:39 ill look at it now 2010-12-12T07:42:00 there is an example map at the bottom of the page 2010-12-12T07:42:42 seems to have run without errors 2010-12-12T07:42:54 is there a way to make it display that output? 2010-12-12T07:43:10 what do you mean? 2010-12-12T07:43:38 oh I see a def print map(pMap) 2010-12-12T07:43:42 maybe thats it 2010-12-12T07:44:20 hmm 2010-12-12T07:46:10 you can reach every non-wall square from your home base 2010-12-12T07:47:45 how do I get that output thats at the bottom? 2010-12-12T07:48:16 on your computer? 2010-12-12T07:48:16 save file as say mapgen.py 2010-12-12T07:48:16 then from terminal 2010-12-12T07:48:16 mapgen.py 2010-12-12T07:48:16 or 2010-12-12T07:48:16 mapgen.py >> map.txt 2010-12-12T07:48:54 you need to have python installed, otherwise just keep clicking "submit" at the very bottom of the page and the website will generate you a new one 2010-12-12T07:49:09 mfw im doing this from windows ;_; 2010-12-12T07:49:15 so am i :P 2010-12-12T07:49:27 xp <3 2010-12-12T07:49:50 I need to boot up a vm machine 2010-12-12T07:50:38 i'm going to restrict it to always having square dimensions and add more symmetries 2010-12-12T07:52:07 I like it, I have to say 2010-12-12T07:52:09 with the wall 2010-12-12T07:52:11 s 2010-12-12T07:55:09 antimatroid: did you want me to look at something? 2010-12-12T07:56:34 sigh: http://codepad.org/DmnK5T95 2010-12-12T07:56:37 just the map at the bottom 2010-12-12T07:57:10 heh, it's like a maze 2010-12-12T07:57:22 ah, you added the check if all squares are reachable 2010-12-12T07:57:23 yep, with wrapped edges 2010-12-12T07:57:46 btw, your comment for is_valid is out of date 2010-12-12T07:57:59 Epona: if you didn't realise, you're allowed to go off the top and come out on the bottom, same with the sides 2010-12-12T07:58:05 sigh: yeah, i'll fix that now 2010-12-12T07:58:14 i'm adding more symmetries 2010-12-12T07:59:04