First we take Manhattan... creative solution for Moore Neighbourhood by veky - python coding challenges

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

First we take Manhattan... solution in Creative category for Moore Neighbourhood by veky

count_neighbours = lambda grid, I, J: sum(cell
    for i, row in enumerate(grid)
    for j, cell in enumerate(row)
    if max(abs(i-I), abs(j-J)) == 1)

Oct. 1, 2014

Comments:

nickie on Oct. 6, 2014, 6:56 p.m.
OK, it's reasonably "clear" for your standards, but still I'd put it in "creative". Unless we're waiting for Berlin... :-)

veky on Oct. 6, 2014, 7:31 p.m.
Your criteria is not the same as mine. But here I thought you'd surely criticize me for algorithmic complexity. :-) P.S. of course we're waiting for <a href="https://www.youtube.com/watch?v=RAoSgpaMupE">Berlin</a>. ;-]

oduvan on Nov. 22, 2014, 5:13 p.m.
The video behind the link is blocked. What is the story about Berlin? :)

veky on Nov. 23, 2014, 6:32 p.m.
Nothing special, just a local song. But you can play "Take my breath away" instead if you want. :-)

LostxinthexMusic on Oct. 15, 2014, 11:11 p.m.
Why make it a lambda?

veky on Oct. 16, 2014, 8:19 a.m.
Why not? It's an expression. lambda is the most natural way of writing expressions with arguments. It's as if you wrote "why make it a sum?". Of course, I could have made an explicit loop with a variable initialized to 0 and added to in every pass, but why? sum builtin is the most natural way of writing such construct. Or, "why make it enumerate?". Of course, I could have iterated through the indices with range of len, and fetch items with indexing, but enumerate is the most Pythonic way of iterating through indices and items together. Or even, why make it abs? I could have made an if-else expression denoting the same thing. But abs is obviously clearer. Your giving me -1 most probably shows that in this particular aspect of Python, but not in all the others, superstitions are clouding your judgement. * "Most probably", because it is possible that you just misunderstood PEP8. In that case, things can be fixed. But superstitions are hard to fix. :-)

LostxinthexMusic on Oct. 17, 2014, 1:21 a.m.
It's rather hard to misunderstand this bit of PEP8: <blockquote> Always use a def statement instead of an assignment statement that binds a lambda expression directly to an identifier. Yes: def f(x): return 2*x No: f = lambda x: 2*x The first form means that the name of the resulting function object is specifically 'f' instead of the generic '<lambda>'. This is more useful for tracebacks and string representations in general. The use of the assignment statement eliminates the sole benefit a lambda expression can offer over an explicit def statement (i.e. that it can be embedded inside a larger expression) </blockquote> There is absolutely no benefit to naming a lambda other than making your code (technically) a one-liner, which, when trying to write a "clear" solution, is pointless. I'd really like to know how you think I'm misunderstanding PEP8.

veky on Oct. 17, 2014, 5:43 a.m.
Ok, there are two things to consider. <h1>Capturing natural levels of abstraction</h1> First, look at this: <pre class='brush: python'>product = 1 for item in sequence: product *= item </pre> vs. this: <pre class='brush: python'>product, index = 1, 0 while index < len(sequence): product *= sequence[index] index += 1 </pre> I assume you agree that first is better for calculating the product of a sequence than the second. Now read carefully the question: is the real reason for that, the fact that it's one line shorter? I hope you realize that it isn't. It is the fact that you're capturing the algorithm on its abstraction level: you need product of all items in sequence, which doesn't speak about index at all. Yes, the second solution is more general: you can increment index in some passes and in some passes not, or you can increment it by 2, or whatever. You can even increment it by sequence[index]. :-D There is simply no way to write such things in the first approach. But that's the point: In Python, you don't automatically write the most general thing (as opposed to Java for example;). It gives you generality at the cost of failing to capture the essence of your algorithm, since it's mixed with the bureaucracy of keeping the generality level too high. Now let's see def vs lambda. I hope you realize the parallel. def gives you much that you don't need. It gives you a Python function, which is really a powerful and general object. It does something to the world. It has its own scope, modifiable with global and nonlocal modifiers. It is introspective: it knows its own name, and inserts itself into stack frames and consequently tracebacks when there is an exception. It is capable of handling the exceptions, or raising them. It has a docstring (and knows it too). When inserted into a class, it has powerful descriptors that make it do the right thing no matter how it's called. It can control its flow; it can not finish at all, it can return early, or it can return nothing (implicit None). And so on. lambda, on the other hand, has nothing of that. It's just an "expression with holes": when called, it inserts some arguments into the holes, and evaluates the resulting expression. Mathematically, it's a closed form instead of a def which is an algorithm. I hope you agree that having a closed formula is (often) better than having an algorithm for solution. (-b(+-)sqrt(b^2-4ac))/2a is much better than "normalize, complete the square, equalize, extract roots, get rid of constants at the left side". :-) <h1>Misunderstanding PEP8</h1> Now, considering all that, let's see what's wrong with your understanding of PEP8. You quoted the text correctly. What you failed to see is the context. Please read the first sentence of PEP8 in full. <blockquote> This document gives coding conventions for the Python code comprising the standard library in the main Python distribution. </blockquote> Its primary concern is writing code that PSF stands behind, the code other people will maintain. That shifts the perspective greatly: capturing natural level of abstraction is dwarfed by the considerations such as flexible introspection and debugging. When an unexpected exception occurs somewhere, it's crucial to pinpoint as precisely as possible where it happened, so it can be fixed quickly. That's what defs do: they are a way of giving introspectable names to framed pieces of code. Even when it's just one expression, it is useful to name not only the expression (which lambda does), but the process of calculating it, so that if something goes wrong with that process, it can be known as soon as possible. In other words, if any of these is true: <ul> <li>you're writing code for the Python's standard library</li> <li>your expression is so complicated that the act of calculating it deserves a name</li> <li>your code is extremely introspective (for example, you naturally import inspect:) and a stack frame is natural unit of thought</li> </ul> then the advice you quoted stands. It's just a tradeoff: a little bureaucracy for greater maintainability and flexibility. But if none of these are true, why would you complicate your code? Simple is better than complex. And expression beats algorithm almost any time. :-)

oduvan on Nov. 22, 2014, 5:18 p.m.
If I get this code for a production solution, I would ask to refactor it.

oduvan on Nov. 22, 2014, 5:23 p.m.
And I believe in the creative section you will get more points for your idea :)

veky on Nov. 23, 2014, 6:38 p.m.
<blockquote> If I get this code for a production solution, I would ask to refactor it. </blockquote> And here you'd be completely right, for a whole different reason. Namely, this solution completely unnecessarily traverses whole grid instead of looking just at the neighbourhood. <blockquote> And I believe in the creative section you will get more points for your idea :) </blockquote> That's the problem, I don't think I would (except from you probably:). Since Clear was made the default category, many people don't even look at others. And that's a big problem since people post in Clear without much thinking. But yes, I have enough points, and I don't really care about these categories, and people looking for my code know where to find it;-), so ok. Category changed.

oduvan on Nov. 24, 2014, 7:15 p.m.
<blockquote> Since Clear was made the default category, many people don't even look at others. </blockquote> Agree with you here. We are still looking for a way to show user some interesting creative solutions. Do you have ideas for us? <blockquote> Category changed. </blockquote> Thank you :)

veky on Nov. 25, 2014, 5:19 a.m.
<blockquote> show user some interesting creative solutions. Do you have ideas for us? </blockquote> Easy, make Creative category the default. :-D Joking aside, you know my view. Categories (as currently implemented) are unnecessary and misleading. They should just be removed. Or, if you insist on them, they should be reworked so readers vote (with their thumbs) about the category. You can even have different thumbscales, so one can give a piece of code e.g. 1 thumb for readability, and 3 thumbs for creativity. (Thumbs for speed probably don't make sense, this is something CheckiO should automatically judge if it's needed.) <blockquote> <blockquote> Category changed. </blockquote> Thank you :) </blockquote> Can I get my thumbs back now? ;-P

oduvan on Nov. 25, 2014, 12:47 p.m.
<blockquote> they should be reworked so readers vote (with their thumbs) about the category. You can even have different thumbscales, so one can give a piece of code e.g. 1 thumb for readability, and 3 thumbs for creativity. </blockquote> That makes sense, but what about commenting? If your solution is focused only on creativity, you get -5 for readability (which makes you feel bad) and +100500 comments which explains to you why it is bad (useless comments from your perspective).

veky on Nov. 25, 2014, 1:13 p.m.
I don't see how this is much different to how it is now. :-P Now people complain it's unreadable if I put it in Clear. If I put it in Creative, they complain it uses too much memory, is too slow, is too long[:-)], has a ragged right edge, or whatever. At least this way it's predictable. :-D Well, I am probably wrong example for such discussions for obvious reasons, but I assume that if someone gets more points for creativity than negative points for readability, I guess they shouldn't feel bad. And of course, comments should only be needed if sum of thumbs given is negative. :-)

hanpari on Nov. 25, 2014, 2:37 p.m.
Nobody can protest seeing your solution in Creative. Creative was made because of you :)

veky on Nov. 25, 2014, 4:42 p.m.
And what, Speedy was made because of nickie? And Clear was made because oduvan hates lambdas? And Uncategorized was made because bryukh can never decide where to put his solution? :-D As I said, I am probably wrong example. :-) But I'm trying to speak generally.

oduvan on Nov. 25, 2014, 12:38 p.m.
<blockquote> Can I get my thumbs back now? ;-P </blockquote> You have all my thumbs :)

zlomovsky on Feb. 28, 2015, 8:33 p.m.
Looks like I've done <a href="http://www.checkio.org/mission/count-neighbours/publications/zlomovsky/python-3/lambda-91/">shortest lambda</a> (91 characters). :)

veky on March 1, 2015, 3:58 a.m.
<a href="http://www.checkio.org/mission/count-neighbours/publications/veky/python-3/wet/">Not any more</a>. :-F

Czarinov on Aug. 15, 2015, 5:23 p.m.
Hey Veky, just wanted to say thanks. Due to analyzing your solution I learn more and more about latest Python fashions... for example using comprehension lists: <a href="http://www.checkio.org/mission/count-neighbours/publications/Czarinov/python-3/quite-comprehensive/">quite comprehensive</a>. Will you clear this out please: 1) how do you not get out of bounds? 2) how do you not get into negative coords? Thanks.

veky on Aug. 15, 2015, 10:48 p.m.
Are you sure you're commenting the right solution? This one doesn't go "around (row,col)" and see which locations are in the grid. It goes through all the positions in the grid, and sees which ones are around (row,col). So it's trivial to always have correct indices: that's what I'm looping through. :-)

Czarinov on Aug. 16, 2015, 4:31 p.m.
Oh, yes, silly me, it's understandable it can't get outta bounds... Again, being tired is not great for learning programming :D...

veky on Aug. 16, 2015, 5:21 p.m.
<blockquote> it’s understandable it can’t get outta bounds… </blockquote> Of course, the problem then is detailed in the first half of <a href="http://www.checkio.org/mission/count-neighbours/publications/veky/python-3/first-we-take-manhattan/share/b3af2a9208dcca30fcfe65b0098695ac/#comment-21015">this comment</a>. Don't do that in real code if you value speed (or sanity:).

Czarinov on Aug. 16, 2015, 8:56 p.m.
Yes, funny thing I actually read it and gave it a thought, so I apologize...

veky on Aug. 16, 2015, 9:22 p.m.
BTW I did laugh very loud on "latest Python fashions"... please see the date on <a href="https://www.python.org/dev/peps/pep-0202/">PEP 202</a>, which introduced list comprehensions. Great fashion, 15 years old. :-DD

Czarinov on Aug. 16, 2015, 9:29 p.m.
:D, damn it!...

veky on Aug. 16, 2015, 10 p.m.
Yes, people usually don't realize how old Python is. See <a href="https://www.pgcon.org/2012/schedule/attachments/240_snakes-and-elephants.pdf,">https://www.pgcon.org/2012/schedule/attachments/240_snakes-and-elephants.pdf,</a> slide 5. :-)

xaoc94 on Jan. 7, 2016, 1:13 p.m.
Thx for really creative solution! I have learned a lot!

veky on Jan. 7, 2016, 1:58 p.m.
Did you get the title pun, too? :-)

xaoc94 on Jan. 7, 2016, 7:56 p.m.
Can you explain? I don't know English enough to understand it. Is it idiom or rather a kind of award from checkio ;) (Hope my guess will be entertaining).

veky on Jan. 7, 2016, 8 p.m.
No, I just meant if you understood what the Manhattan in the title is referring to.

Flying_Eye on May 4, 2016, 9:57 p.m.
Nice functional programming style solution

veky on May 5, 2016, 5:50 a.m.
I like your wording. :-) Just a week ago, I was explaining to people that Python is not a functional language, but can imitate the style of functional programs without much problems. :-)

U.V on March 15, 2023, 7:09 p.m.
I find this old solution interesting B-).

veky on March 15, 2023, 7:31 p.m.
You're not the only one. :-D

GitHub

0 %

PyCharm The Python IDE for Professional Developers. Enjoy productive Python, web and scientific development with PyCharm. Download it now!