Factored clear solution for Weak Point by veky - python coding challenges

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

Factored solution in Clear category for Weak Point by veky

def weak_point(matrix):
    def ms(rows):
        sums = tuple(map(sum, rows))
        return sums.index(min(sums))
    return ms(matrix), ms(zip(*matrix))

July 9, 2014

Comments:

james.verran on July 10, 2014, 11 p.m.
Why the tuple()?

veky on July 11, 2014, 6:04 a.m.
Without it, sums would be an iterator, and min(sums) would exhaust it, so I couldn't give it to sums.index. Moreover, iterator doesn't even have .index method. I could have avoided that by using enumerate, then min with key=operator.itemgetter(1), and then [0] on it, but I think this is more readable.

james.verran on July 11, 2014, 7:17 p.m.
I'm still confused... Doesn't map return a list? I tried it without the tuple() and it appears to work.

veky on July 11, 2014, 7:44 p.m.
Ah, I see. You use Python 2. I use the real Python. ;-]

lsdda on Aug. 3, 2014, 10:15 p.m.
What's the advantage for changing it in Py3?

veky on Aug. 4, 2014, 11:08 p.m.
Ah... as if you asked "what's the advantage of living in 21st century instead of 19th one?" :-P There's just too many fundamental changes. For just one instance, text is finally text (string of characters), not some weird ASCII-encoded bunch of bytes where you never know what exactly \x83 means. For another, int is really a representation of mathematical integer, not of some C-memory-imposed view of a machine word on 32bit architecture. Also, 2/4 is 0.5, not 0. :-P And so on... the language is just better. It has broken up with its C heritage and taken its philosophy seriously. For more information, you can watch <a href="https://www.youtube.com/watch?v=f_6vDi7ywuA">https://www.youtube.com/watch?v=f_6vDi7ywuA</a>

mfurones on Jan. 6, 2015, 2:50 p.m.
great explain

veky on Jan. 6, 2015, 8:01 p.m.
Thanks. :-)

eugene128736871 on May 11, 2015, 12:12 a.m.
What’s the advantage for changing list to iterator in py3?

veky on May 11, 2015, 8:46 a.m.
In fact, most of the same kind of advantage as arbitrary int or str: not imposing any technical limitation of implementation onto an abstract data structure. Many times, what you need is a bunch of "items", but not in whole: just one by one. For example, `sum` might take a list, a tuple, a set or even a dict, and in fact <a href="http://www.checkio.org/forum/post/2975/greeting-2/">many other collections</a>, but there is no need to access more than one item at a time. That means you can forget about "the walls of a container" and just feed it items. For some others, like `any`, it isn't even necessary to have a finite bunch of items, as long as one of them is true. <hr /> On the other hand, often you need to produce a bunch of items one by one, and you don't want to commit to any particular container. You might need all the permutations of "too", but do they form a set? A list? A tuple or something else? Does the producing algorithm really need to care about that? It's a technical detail, it doesn't affect the algorithm. (It does affect the result, since it matters whether you want duplication when you have two equal letters, and it matters whether you impose a lexicographic order derived from the starting data onto the permutations - but the producing algorithm still does the same thing.) And also here, many times the bunch of data you need is not really finite. You need Fibonacci numbers. They surely don't form a Python list, set, or anything value-like, but you can generate them one by one. Stopping condition is something the caller needs to care about, not generator of Fibonacci. <hr /> So, you might need to produce generically an amorphous and potentially infinite bunch of data, and you might need to consume generically an amorphous and potentially infinite bunch of data. Value-freezing that bunch into a list or whatever on one end, and iter-unfreezing it on the other is wasteful, slow, and might not even make sense when the bunch is infinite. Much better is to represent those bunches on their natural level of abstraction. <pre class='brush: python'>def Fibonacci(start=0, next=1): while True: yield start start, next = next, start + next >>> any(str(f).endswith(4*"0") for f in Fibonacci()) True </pre> We see that there is a Fibonacci number ending with four zeros. To conclude that fact, we didn't need to think about what's the limit we need to compute up to. It didn't matter. It's just a technological limitation of implementation, not a feature of the algorithm. (There is an important feature of the algorithm: semidecidability. If there were no Fibonacci number ending with four zeros, the above algorithm would never conclude that fact. But that's just a fundamental limitation: there is no way around the Halting problem in general. What is there, is a way to make semidecidable algorithms as easy to code as decidable ones.)

eugene128736871 on May 11, 2015, 10:52 a.m.
thank you. What's best way to cache iterator sequence? And should i do it?

veky on May 11, 2015, 11:38 a.m.
The whole point of my article is that it depends on what are you intending to do, and Python will not do it for you implicitly. :-) Also implied is the advice that you shouldn't do it yourself without good reason, and surely not "by default". For this concrete solution, I wanted something having `.index` method, and what `min` can iterate through. Tuple was the cheapest. But in many other instances, I make the whole program just juggling the items "in the air", and value-freeze only at the end, depending on what the CiO testing system is expecting.

james.verran on July 11, 2014, 7:17 p.m.
I'm still confused... Doesn't map return a list? I tried it without the tuple() and it appears to work.

Winandfx on Jan. 23, 2015, 5:32 p.m.
completely forgot about zip functionality and....) <pre class='brush: python'>map(lambda *a: a, *matrix) </pre>

veky on Jan. 24, 2015, 9:28 a.m.
Yes, zip is a special case of map. :-) Or rather, map is generalized to more than 2 arguments via zip.

AlexGiesbrecht on Feb. 26, 2015, 2:48 p.m.
This is SO clean and efficient compared to mine... Gotta learn 'pythonese' a lot still

veky on Feb. 27, 2015, 4:11 a.m.
Of course, your life would be much easier then. :-) And don't trust people who say all languages are the same, you should "just learn to program". That's like saying "you should just learn to talk, all human languages are the same". Language truly shapes our thoughts, and it matters. "Pythonese" for Pythonese is Pythonic, BTW. ;-D (Yes, it's an adjective sometimes used as noun, but so is Pythonese.:)

vananabu on Aug. 24, 2015, 1:52 p.m.
your solution is laconic

veky on Aug. 24, 2015, 8:05 p.m.
In fact, Pythonic word is "concise". But that's nitpicking. :-)

vananabu on Aug. 25, 2015, 1:25 p.m.
It is not bad :) There is an adage in my native language - "conciseness is a sister of talent".

veky on Aug. 25, 2015, 2:12 p.m.
<blockquote> It is not bad :) </blockquote> <a href="http://paddy3118.blogspot.com/2007/11/python-concise.html">Here</a> is the language nitpick if you're interested. <blockquote> in my native language - "conciseness is a sister of talent". </blockquote> Wow. May I ask what language is it? :-)

vananabu on Aug. 25, 2015, 7:56 p.m.
I am from Ukraine, but my native language is a russian. what does the word 'nitpick' mean? 'google' translate this word strangely, because it sounds like a bad language feature

veky on Aug. 26, 2015, 4:27 a.m.
dictionary.com says: "nitpick: to be excessively concerned with [...] inconsequential details." Does this help? :-) I just wanted to say it's really irrelevant what the word is, but it's also interesting that there are people who would write a blog post about it. :-) <hr /> Ah yes, Russian has very charming proverbs. :-)

vananabu on Aug. 26, 2015, 7:29 a.m.
yes, it is rich language

Ice Base

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.