DOX and XOX creative solution for Xs and Os Referee by veky - python coding challenges

Enable Javascript in your browser and then refresh this page, for a much enhanced experience.

DOX and XOX solution in Creative category for Xs and Os Referee by veky

checkio=lambda r,j="".join:"DOX"[sum(m*3in[j(r[::e])[::4]for e
in(1,-1)]+r+list(map(j,zip(*r)))for m in"XOX")]

Sept. 12, 2013

Comments:

qria on Sept. 14, 2013, 7:36 p.m.
>>> r = ['012','345','678'] >>> ["".join(r)[::e]for e in (4, -4)]+r+list(map("".join,zip(*r))) ['048', '840', '012', '345', '678', '036', '147', '258'] # :'(

qria on Sept. 14, 2013, 7:47 p.m.
Maybe change ["".join(r)[::e]for e in (4, -4)] to ["".join(r)[::4]] +["".join(r)[2:-1:2]] ? Though it'll break your structure :\

veky on Sept. 14, 2013, 9:50 p.m.
<pre class='brush: python'>"".join(r[::e])[::4] for e in (-1,1) </pre> is more symmetric. But yeah, structure bye-bye...

yan123_703b617968a74969bca2fba on Oct. 23, 2013, 12:54 p.m.
or just change ["".join(r)[::e]for e in (4, -4)] to something like ["".join(a)[2-e:7+e:2+e]for e in(0, 2)]

veky on Oct. 23, 2013, 2:07 p.m.
Yes, but I think the above (what I wrote in response to the same post) is still easier. :-)

veky on Sept. 14, 2013, 9:48 p.m.
Damn you're right. :-/ Ah, not the first time the tests were not covering everything...

JackSchneider on Sept. 18, 2013, 8:05 p.m.
look briljant but is unreadable. If one of my employees did this to me he must look for an other job. or deliver a good explanation why this good work.

veky on Sept. 19, 2013, 9:16 p.m.
I completely agree with you. I would also like you to note the fact that I am not your employee. ;-) [But you can try paying me, say, a $100, and I promise you won't be disappointed by the code I write for this task.;]

JackSchneider on Sept. 19, 2013, 9:30 p.m.
Ik wil be higher when you explain your solution in a way that ik can make an even briljantsolution in a readable structure.

veky on Sept. 19, 2013, 9:40 p.m.
It's not so hard. But as you can see below, it has a bug. So it really has no explanation except "test cases not covering all possibilities". :-)

veky on Feb. 26, 2014, 1:31 p.m.
It's fixed now. If you're still interested, let me know. ;-)

nptwz on Sept. 19, 2013, 3:55 p.m.
unreadable

veky on Sept. 19, 2013, 9:13 p.m.
Really? A reason for -3? Unreadable? Not even a mention that it actually has a bug (down in the comments)... one of my rare solutions with a concrete bug, and it gets -3 for unreadability. Oh the irony. :-D

Asday on April 26, 2014, 1:49 p.m.
The bug is that it throws when both players win, right? This is only a valid checker for legal games of O/X.

veky on April 26, 2014, 2:48 p.m.
No. It's a long story. :-) It had a bug, but it was fixed without resetting comments. Don't ask me how, I can't tell you. ;-P

Asday on April 26, 2014, 3:22 p.m.
I'm right, though, right? If both players win, it sum()s to 3, which throws an index error.

veky on April 26, 2014, 3:30 p.m.
The only problem is that both players can't win. Have you ever played this game for real? :-P This highlights a bigger problem in your thinking. Try to answer this: what does checkio(["ABC", "DEF", "GHI"]) do? What does checkio(["This", "is", "a", "nice", "task"]) do? What does checkio([]) do? Even more, Python (nonannotated) function argument names don't have types. Names don't have types, ever. Values have types. What does checkio(3415) do? Yep, it "throws" (it's "raises" in Python:). It's the most sensible thing it could do. :-)

Asday on April 26, 2014, 5:33 p.m.
I know they can't, but in the little tester thing below the code editor for this problem, you could randomise illegal games. I saw one with O winning twice, having placed 6 pieces, to X's 3.

veky on April 26, 2014, 6:16 p.m.
That's the problem with tester, of course. [I have long since given up on trying to fix tests here, it's hopeless.:]

TarasBasiuk on Oct. 31, 2013, 6:17 a.m.
unreadable

veky on Oct. 31, 2013, 8:52 a.m.
LOL. :-) I really can't say anything else.

oduvan on Oct. 31, 2013, 2:02 p.m.
unreadable

veky on Nov. 1, 2013, 6:33 a.m.
You're doing it wrong. You're supposed to give it at least a -1, preferably -5. :-D

chros on Nov. 7, 2013, 4:14 p.m.
"it was hard to write, it should be hard to understand" ;-)

veky on Nov. 7, 2013, 5:14 p.m.
Unfortunately, the converse doesn't hold. This wasn't hard to write. :-D

rhcoder on Nov. 14, 2013, 9:46 a.m.
can u explain the code please?????

veky on Nov. 14, 2013, 1:18 p.m.
checkio=lambda r:blah is kinda like def checkio(r):return blah, but shorter. :-) So, the thing left to explain is blah. It is a letter of string "DOX", whose index is 0, 1 or 2, whose value is calculated as sum(bleh for m in "XOX"): bleh is a condition, so it is 1 (True) or 0 (False), and the sum is 1 if m=='O' is successful (since there is 1 "O" in "XOX", and 2 if m=='X' is successful (since there are 2 "X" in "XOX"), or just 0 if none are succesful. So, we must explain the condition bleh. It is "m * 3 in first+second+third", where m*3 is threefold repetition of m ("XXX" or "OOO"), and first, second and third are lists iterating through verticals, horizontals and diagonals (for example, second is simply r, since it is a list of three strings containing what's written in rows). Clear so far?

lsdda on Nov. 20, 2013, 1:31 p.m.
This solution, though difficult to interpret, is actually really inspiring. I was thinking "I'm sure there's no way to write an answer within 10 lines for this task" while doing the quiz, but... All I can say is "Hello world"...

veky on Nov. 20, 2013, 2:12 p.m.
Thanks. Be sure to check out other solutions of mine. ;-)

AndrewSkripko on Nov. 28, 2013, 5:33 p.m.
great solution!

marshall.zheng on Dec. 8, 2013, 12:57 p.m.
wow! really awesome solution! eagerly to learn more about python!

JuliusArcanum on Dec. 8, 2013, 4:32 p.m.
Gee? That's all? This two little lines return the winner of the game? Good! Thanks for sharing (and for having explained this code below.)

seandor on Feb. 7, 2014, 2:32 a.m.
not pythonic.

veky on Feb. 7, 2014, 2:47 p.m.
Compared to yours, it even might be. :-P

......... on Feb. 26, 2014, 10:40 p.m.
a nice-looking guy

veky on Feb. 26, 2014, 11:31 p.m.
Guy? A strange name for a one-and-half-liner. :-D

Asday on April 26, 2014, 1:48 p.m.
Alright, let's look at this, then. <pre class='brush: python'>["".join(r[::e])[::4] for e in (1, -1)] </pre> This gives you the diagonals. How? The list comp expands to: <pre class='brush: python'>"".join(r[::1])[::4] "".join(r[::-1])[::4] </pre> This is a use of slice notation, that not errbody knows about. An empty slice, (the first colon), refers to the entire list. The third entry, (past the second colon), is the stepping. 1 is the list as normal. -1 is the list backwards. 4 is every fourth item. The first expansion of the list comp arranges the rows as normal, clumps them into one string, then gets the zeroth, the fourth, and the eighth elements, which is the topleft to bottom right diagonal. The second expansion reverses the rows, thus flipping the diagonals, and does the same. The next bit, the "+r", appends the rows to the diagonals list. We're slowly building up an AoE, a view of the game result, which is a bunch of 3strings that, if they are the same, count as a win. Next up, we want the columns. <pre class='brush: python'>zip(*r) </pre> Unpacks the results into three strings, and gives them to zip. Zip takes these three iterables, and returns a list of tuples which are: <pre class='brush: python'>[(1st item of the 1st iterable, 2nd item of the 1st iterable...), 2nd item of the 1st iterable, 2nd item of the 2nd iterable...), .... ] </pre> In effect, zip(*r) is transposing the input. All we need to do to this now is get the output, for example <pre class='brush: python'>[('O', 'X', 'X'), ('.', 'X', 'O'), ('X', '.', 'O')] </pre> Back into the original form <pre class='brush: python'>['OXX', '.XO', 'X.O'] </pre> Which is what the map() is for. It performs "".join(x) for every x in the iterable given to it, in this case, the results of zip(). Next up, he turns that into a list. I'm not sure why. Maybe a different version of python has a different output for map(). NEXT up, he extends the list which already has the diagonals and horizontals with the newly created verticals list. So now we understand: <pre class='brush: python'>[j(r[::e])[::4]for e in(1,-1)]+r+list(map(j,zip(*r)) </pre> We'll call this `AoE` from now on to shorten things. In doing so, we get: <pre class='brush: python'>"DOX"[sum(m*3 in AoE for m in "XOX")] </pre> There's a bit of shorthand here we can expand to: <pre class='brush: python'>"DOX"[sum(m in AoE for m in ("XXX", "OOO", "XXX")] </pre> Let's take the comprehension and expand that again. <pre class='brush: python'>"XXX" in AoE "OOO" in AoE "XXX" in AoE </pre> So these check for wins by each player, and evaluate to True or False. Boolean values can be used in maths, and act as 1 == True, and 0 == False, so summing them works just fine. If there is a win for O, `"OOO" in AoE` will return True, and (hopefully) `"XXX" in AoE` will return False, these will be summed, equal 1, and make the lambda return the 1th item of "DOX", which is "O". If there is a win for X, both of the `"XXX" in AoE` will return True, `"OOO" in AoE` will return False, these will be summed, and equal 2, returning the 2th item of "DOX", which is "X". If there isn't a win, nothing evaluates to True, so <pre class='brush: python'>sum(False, False, False) </pre> happens, returns 0, and makes the lambda return the 0th item of "DOX", which is D. Cool.

veky on April 26, 2014, 3:52 p.m.
<blockquote> Cool. </blockquote> So is your comment (and it confirms my theory "the shorter the code, the longer the explanations people write":). However, it has some mistakes and weird phrasings, so I'll try to correct them below. <blockquote> slice notation, that not errbody knows about. </blockquote> Yep, it's also a use of "".join, That Not Everybody Knows About. It also uses list comprehensions, TNEKA. And so on. Python really shouldn't have "obscure features", at least not in the core language. Guessing what is widely or less widely known is prone to errors, even with BDFL does it. ;-] <blockquote> empty slice </blockquote> That's a really strange way of naming it. "Default" would be better. "empty first two arguments" is even more correct. :-) BTW I really don't see what AoE means. Surely my code is less complicated than Age of Empires - though not by much. ;-DD <blockquote> a bunch of 3strings that, if they are the same, count as a win. </blockquote> 3strings don't have to be same. Characters in them do. <blockquote> Maybe a different version of python has a different output for map(). </blockquote> It's the only true version of Python. ;-) (Python2 really should get renamed to milk snake.:)) We had that discussion about range. We'd have to have range and xrange, zip and izip, keys and iterkeys... meh. Much better to have One Way To Do It. <blockquote> Let's take the comprehension and expand that again. </blockquote> There's no comprehension here. There are called generator expressions, and are really cute. They solve the same problem as in the above paragraph. ;-) <blockquote> act as 1 == True, and 0 == False </blockquote> While correct, it's a bit like saying that 2.0 "acts as 2.0 == 2". It is equal to 2. 1 is equal to True. This is not magic, just a subclass with a different __repr__. <blockquote> sum(False, False, False) </blockquote> There is no such thing. sum still (despite some people's wishes:) doesn't take 3 arguments. :-) You probably wanted to say a list <pre class='brush: python'>sum([False] * 3) </pre> or a generator expression <pre class='brush: python'>sum(False for _ in range(3)) </pre>

Asday on April 26, 2014, 5:40 p.m.
<blockquote> BTW I really don't see what AoE means. Surely my code is less complicated than Age of Empires - though not by much. ;-DD </blockquote> Area of Effect. I've always used that for describing shapes in a multidimensional array, I dunno. Guess it's habit. <blockquote> 3strings don't have to be same. Characters in them do. </blockquote> Right you are, that's what I meant, but spoke differently. That was silly. <blockquote> There's no comprehension here. There are called generator expressions, and are really cute. They solve the same problem as in the above paragraph. ;-) </blockquote> Oh cool, I actually thought generator expressions were another form of comprehension. They're the same syntax, etc., minus the different brackets. <blockquote> While correct, it's a bit like saying that 2.0 "acts as 2.0 == 2". It is equal to 2. 1 is equal to True. This is not magic, just a subclass with a different repr. </blockquote> Yeah, after posting, I thought "int(True) == 1" would have been better wording, but owell. <blockquote> There is no such thing. sum still (despite some people's wishes:) doesn't take 3 arguments. :-) You probably wanted to say a list </blockquote> Aah, I didn't actually realise. I dunno what function I was thinking of. I was fairly sure there was a function which took n arguments, then some more as named arguments. Could have sworn it was sorted(), but apparently not.

veky on April 26, 2014, 6:31 p.m.
<blockquote> Oh cool, I actually thought generator expressions were another form of comprehension. They're the same syntax, etc., minus the different brackets. </blockquote> Yes, but one big difference that usually surprises people: delayed execution. They are functions, like lambdas, which means their code isn't executed when they are constructed, but when they are called ("iterated over" would be more correct). With the values of closed-over variables at that time (same as other functions). <blockquote> Yeah, after posting, I thought "int(True) == 1" would have been better wording, but owell. </blockquote> "True == 1" would be even more better wording. ;-] Not "acts as", but "is". :-) <blockquote> I was fairly sure there was a function which took n arguments, then some more as named arguments. Could have sworn it was sorted(), but apparently not. </blockquote> You probably think of min and max. But that is probably a design error. Very convenient, but still subtly wrong.

Asday on April 26, 2014, 9:10 p.m.
<blockquote> Yes, but one big difference that usually surprises people: delayed execution. They are functions, like lambdas, which means their code isn't executed when they are constructed, but when they are called ("iterated over" would be more correct). With the values of closed-over variables at that time (same as other functions). </blockquote> Yeah, I knew that, just didn't realise they had different terminology. I only learned about them recently, though. I REALLY wanna use them for stuff, but I have very little that needs them. <blockquote> You probably think of min and max. But that is probably a design error. Very convenient, but still subtly wrong. </blockquote> Yep! Those are the ones. Thanks for your replies, man. :D

veky on April 26, 2014, 9:13 p.m.
<blockquote> I REALLY wanna use them for stuff, but I have very little that needs them. </blockquote> Once your eyes are opened, you'll be amazed. ;-)

canatan on May 8, 2014, 7:03 p.m.
is it possible to cut down further or already ultimate?

veky on May 8, 2014, 8:16 p.m.
I guess this is a rhetorical question. ;-)

xiongbiao on May 20, 2014, 9:19 a.m.
So.......

veky on May 20, 2014, 12:56 p.m.
yes?

LorenzoPanichi on July 23, 2014, 4:05 p.m.
Man, you're so good!! That line is so beautiful!! I mean, I'm not a proper Python programmer, I didn't even know of all of the language features you used, but the way you create the diagonals is a true lesson!

veky on July 23, 2014, 10:21 p.m.
Tnx. Read other solutions of mine here, you might like them. ;-)

maximusderango on July 29, 2014, 3:26 a.m.
This is quite impressive. I spent some time picking apart the code and learned a lot about python in the process. Thanks for the contribution.

veky on July 29, 2014, 10:36 a.m.
That was exactly my intention. Try doing that with other my solutions here. :-)

Czarinov on Aug. 18, 2015, 8:01 p.m.
mind blown... PS I have yet to pick your solution apart. I tried reading comments, but I just need to explore all the parts by myself to understand. By the way, I had an idea to how check for diagonals, by shifting first and 3rd rows by +-1 position and checking for rows in transposed version. Not sure if needed, but it's always good to expand one's imagination. EDIT: I have a feeling you got that implemented and it actually gave me a subliminal idea just looking at your code :D

veky on Aug. 18, 2015, 8:30 p.m.
<blockquote> By the way, I had an idea to how check for diagonals, by shifting first and 3rd rows by ±1 position and checking for rows in transposed version. </blockquote> [One of these italic rows should be columns, right?] Not bad, but really too long. :-D

Czarinov on Aug. 18, 2015, 9:05 p.m.
<blockquote> [One of these italic rows should be columns, right?] </blockquote> :D, yes, probably, you got the idea.. Man, I need this sorted out before I go any further, would you help? I tried googling, but didn't really find the answer to this: <pre class='brush: python'> def checkio(x=None): a = [1,1,1,1,1] b = a for x in a: if x == 1: b.remove(x) return 0 </pre> I try deleting all 1 from b, but it starts jumping over x in a, as if I was really trying to delete from a... So instead of [] I end up with [1,1]. And ofc a gets edited too... It's like b and a are connected and it drives me mad.

veky on Aug. 18, 2015, 9:20 p.m.
They are not connected, they are the same. You said yourself, `b = a` - what did you expect? The sooner you forget about C and its value semantics, the better. :-P <a href="http://nedbatchelder.com/text/names.html">http://nedbatchelder.com/text/names.html</a> If you want a copy, just say so. <pre class='brush: python'>b = a.copy() </pre>

Czarinov on Aug. 18, 2015, 9:26 p.m.
Holy cows!... triple thanks. You don't even know how much weight you lifted off my mind.

veky on Aug. 18, 2015, 9:31 p.m.
It's fascinating how much easier life would be if people could just unlearn the wrong concepts from the blind alleys of computer science. :-)

Czarinov on Aug. 18, 2015, 10:37 p.m.
Wrong? Or different? :) That was a great link you posted. That was some awesome stuff to learn.

veky on Aug. 19, 2015, 6:41 a.m.
Yes, you could defend the view that they are just different. But please read <a href="http://www.jerf.org/iri/post/2942">this</a>. I would argue that when a concept has so many bad consequences, it is wrong. The time when our programs were running in controlled environments, with data whose bounds and extent we knew exactly, is long gone. We need dynamic objects to make sense of the universe. Everything else, including value semantics, is just holding us down with no real benefit (except speed, but that's overvalued in most cases anyway). Another important point: consistency. BCPL is one of these rare languages that really tried to have a consistent value semantics. It was very hard then, and it's much harder now. C gave up regarding arrays (except when they are a part of a struct:-P) and functions, and C++ just made a big mess of it, where you never know whether you're dealing with a value or a reference. See any of "move constructor" discussions in modern C++ for an example of how this can go horribly wrong. Java tried to regain sanity, but still had to bow down to false god of value semantics for primitive types, in the name of performance. [C# just copied Java in an ugly way, adding some irrelevant syntactic sugar, like MS usually does.] The language just makes it too easy to do the wrong thing, see <a href="http://programmers.stackexchange.com/questions/203970/when-to-use-primitive-vs-class-in-java">this</a> for example. Python realized that if you want people to do the right thing, you cannot make any compromises. Naming an object shouldn't change it, nor do anything with its value. It's orthogonal. Whether that object is an int, str, list, or open file. (Suppose I start referring to you as Eddie in my private conversations with other people. Do you feel changed in any way? :-) Do you care, even? Note that you didn't have to know this up until now. I didn't request a blueprint of you to make my private copy of you and call it Eddie. That would be horribly inefficient, not to mention technically impossible, and violating many human rights.:)

Czarinov on Aug. 19, 2015, 8:06 p.m.
Thanks man, that's a great answer. ;) <blockquote> Solved all tasks a long time ago. Now I'm just teasing people with reviews and comments. ;-) </blockquote> Ahaha :D, you devil!

veky on Aug. 19, 2015, 8:59 p.m.
Well, that's not completely true anymore since CiO got some new tasks. Thanks for reminding me. :-]

Czarinov on Aug. 19, 2015, 10:27 p.m.
Oh man, finally I picked your solution apart and got to understand it. Suddenly it's not some random letters as it seemed at first. Cheers.

Flying_Eye on May 4, 2016, 10:44 p.m.
With spaces (PEP8 you know) it would be much better :-) Despite that, great solution!

veky on May 5, 2016, 5:46 a.m.
I don't think it's worth it to PEP8 this one. You can try if you want, but my guess is it won't get significantly more readable. :-) But I <a href="https://checkio.org/mission/x-o-referee/publications/veky/python-3/trick-or-treat/?ordering=most_voted">do have</a> a readable solution for this task, if you want to see it. :-)

Flying_Eye on May 6, 2016, 5:45 a.m.
I don't know, for me something like this <pre class='brush: python'>checkio = lambda r, j="".join: ( "DOX"[sum(m * 3 in ([j(r[::e])[::4] for e in (1, -1)] + r + list(map(j, zip(*r)))) for m in "XOX")] ) </pre> is somewhat readable :-) Though formatted like this it looks not nice :-(

veky on May 8, 2016, 12:43 p.m.
Yes, when you format Python like LISP, you get worst of both worlds. :-P

george.blackthorn on Aug. 6, 2016, 6:58 p.m.
Thanks for this solution. It helped me to learn or understand better some Python techniques. And not to forget, I like your way of thinking :-)

veky on Aug. 7, 2016, 3:03 p.m.
I'm glad to hear that.

przemyslaw.daniel on Nov. 3, 2017, 8:37 p.m.
Hell has frozen over. I can see 'Your solution failed...'. Do you know what happens with all task points in that case?

veky on Nov. 3, 2017, 8:48 p.m.
First, it's not like I care. I have too many points here already. :-D Second, I think it's bogus. I have manually resubmitted now, and it works without problems.

bsquare on Aug. 29, 2019, 10:11 a.m.
Outch ... This is one I would call not readable ^^

veky on Aug. 29, 2019, 11:08 a.m.
Me too. Look at the category. ;-)

List, list and list again

0 %

You ClassRoom Demo is Ready With the new tool CheckiO ClassRoom, you will be able to analyze your students' progress, have your own leaderboard, change missions order on the map, and so on.