Assault of batteries clear solution for IP Network: Route Summarization by veky - python coding challenges

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

Assault of batteries solution in Clear category for IP Network: Route Summarization by veky

def checkio(data):
    from ipaddress import IPv4Address, IPv4Network
    ipmin, *_, ipmax = sorted(map(IPv4Address, data))
    spread = (int(ipmin) ^ int(ipmax)).bit_length()
    return str(IPv4Network(ipmin).supernet(spread))

March 13, 2016

Comments:

ljy95135 on April 19, 2016, 3:39 p.m.
eye-opening. Seems that as long as knew a module or method exists before coding can save lots of time.(like ipaddress and bit_length) BTW, how do i know IPveAddress is sortable and can be converted to type int, try it at command line?

veky on April 19, 2016, 8:31 p.m.
<blockquote> eye-opening. </blockquote> The title is suggestive. ;-) <blockquote> Seems that as long as knew a module or method exists before coding can save lots of time.(like ipaddress and bit_length) </blockquote> Of course. That's why they are there. <blockquote> BTW, how do i know IPveAddress is sortable and can be converted to type int, try it at command line? </blockquote> You can always do that (or, if you're feeling more exploratory, use `dir` builtin). Or more simply, read the docs. <blockquote> <a href="https://docs.python.org/3/library/ipaddress.html#comparison-operators">Comparison operators</a> Address objects can be compared with the usual set of comparison operators. <a href="https://docs.python.org/3/library/ipaddress.html#conversion-to-strings-and-integers">Conversion to Strings and Integers</a> To interoperate with networking interfaces such as the socket module, addresses must be converted to strings or integers. This is handled using the str() and int() builtin functions. </blockquote>

ljy95135 on April 20, 2016, 1:01 a.m.
Thank you, i get it. I am learning Python by reading Learning Python, and when something is beyond what i have learned i turn to the docs which contains all details and sometimes hard to absorb all. But the more the information, better to use when need.

veky on April 20, 2016, 8:14 p.m.
You only have to know (or dare to hope:) that there is a module to do what you want. Then you go to docs.python.org, enter "IP address" into Quick Search field, and see that the module you're looking for is called ipaddress. Then you import it, and read the docs. :-)

pohmelie on Aug. 28, 2016, 10:15 p.m.
Awesome solution! 👏 Funny, that python have this module at all. Sometimes I think that standard library was made as: — WOW, MAN, WE NEED THIS IN STDLIB RIGHT NOW! — YEAH, DUDE! But this can't be removed or refactored cause legacy. Since now only small part of stdlib is used as mainstream and everyone use pip.

veky on Aug. 29, 2016, 4:32 a.m.
<blockquote> But this can’t be removed or refactored cause legacy. Since now only small part of stdlib is used as mainstream and everyone use pip. </blockquote> I must say I don't understand you. Every version of Python brings great changes to stdlib. Look at "What's new" documents, you'll be surprised. Even ipaddress is new since 3.3. Many things are improved, bugs are fixed, and even features deprecated constantly. And stdlib APIs have <a href="https://www.python.org/dev/peps/pep-0411/">"provisional" phase</a> carefully designed to minimize the effect of freezing the bad interface. It is not a completely solved problem, but it surely isn't as bad as you make it seem. There is also an instructive notion of <a href="https://en.wikipedia.org/wiki/Long_tail">a long tail</a>. Yes, few libraries are used very often, and many libraries are used rarely, but total usage is non-negligible in almost every case (one of rare examples is <a href="https://docs.python.org/3/library/formatter.html">formatter</a>, but as I <a href="https://checkio.org/mission/roman-numerals/publications/veky/python-3/a-derelict-battery/">showed here on CiO</a>, it also has its very charming strengths.;) You need <a href="https://docs.python.org/3/library/sndhdr.html">sndhdr</a> maybe once or twice in a lifetime, but when you need it, you're very grateful it's in stdlib. Without those "never used" packages, Python would be a much poorer language. (Compare with Joel's <a href="http://www.joelonsoftware.com/articles/fog0000000020.html">"80/20 myth" article</a>.) "Everyone uses pip" is a very biased statement. You probably work in web technology, right? When you can control your own environment on the server you have admin rights on, it's easy to go medieval on your `requirements.txt`. :-) But when your software is distributed on machines you've never seen, by the people you've never met, situation is very different. Also, if you work in education, it's very important to teach the best practiced in a self-contained installation. Otherwise you just never get out of <a href="http://stackoverflow.com/questions/2817869/error-unable-to-find-vcvarsall-bat">vcvarsall.bat</a> hell. Trust me, I know. :-)

pohmelie on Aug. 29, 2016, 9:19 a.m.
<blockquote> Every version of Python brings great changes to stdlib </blockquote> I didn't say, that it does not, of course. Don't get me wrong, stdlib is awesome and strong side of python. But I don't understand some modules and why they are in stdlib. They purpose is to solve tiny task, when stdlib should be (I belive) like a good base for solving more different tasks. And in this way `struct` module is a tool for this task (I mean dealing with ip address), I think. But `struct` is stuck with poor api. Look at `construct` and what can be done on base of it (I mean some cleaning and a little more features). Another thing with legacy is that you can't drop old features. Just look at three ways of formatting since 3.6. The last one (`f"{x + 1}"`) is just what the world need™. But we can't drop old formatting and some people will still use percents, huh. I understand, that if I don't want to use percents I can do it, but other people will, and I will see their code and become sad. My opinion is that some things should be dropped, for clearing/simplifying future codebase. <blockquote> You probably work in web technology, right? </blockquote> Not at all. <blockquote> But when your software is distributed on machines you’ve never seen, by the people you’ve never met, situation is very different. </blockquote> What can go wrong? I don't get it, give me example. Even if you have no internet connection you can build self contained binaries with `pyinstaller` or provide wheel files. I really don't get those who use `urllib` from stdlib and they know about `requests`, but want «no third dependencies». This is pretty weird. <blockquote> Also, if you work in education, it’s very important to teach the best practiced in a self-contained installation. Otherwise you just never get out of vcvarsall.bat hell. </blockquote> wheels solve problem with binaries and if package author will provide wheels for different platforms, than everything will be simple as `pip install`. Of course some «bad guys» did not, but for linux this is not a problem. For windows there is <a href="http://www.lfd.uci.edu/">http://www.lfd.uci.edu/</a>~gohlke/pythonlibs

veky on July 17, 2022, 9:39 a.m.
I was sure I replied to this at the time (6 years ago:), yet I don't see my reply here. :-/ So I'll try to reconstruct main parts of it. <blockquote> I don’t understand some modules and why they are in stdlib. They purpose is to solve tiny task, when stdlib should be (I belive) like a good base for solving more different tasks. </blockquote> Why exactly? Official rationale for inclusion into stdlib is "people need this often, and when they do it themselves, they usually implement it badly, forgeting corner cases and introducing bugs that need a lot of time to surface". I think IP address handling (especially IPv6, but also IPv4) fits that description. There are a lot of cultural-social conventions of what certain addresses mean, and how to do certain things (especially related to submasks), that it's hard to imagine anyone wanting to "just make it work" thinking about them for a long time. <blockquote> And in this way struct module is a tool for this task (I mean dealing with ip address), I think. </blockquote> Then you might be amused to know that `ipaddress` doesn't use `struct` at all. The only things it could use from there are already available as int methods .tobytes and .frombytes. <blockquote> But struct is stuck with poor api. Look at construct and what can be done on base of it (I mean some cleaning and a little more features). </blockquote> First, I think people are overemphasizing the freedom modules have when they are not in the stdlib. Yes, if you're some obscure module nobody uses, then you can change your API every day. But if you're even moderately popular (let alone as frequently used as modules in stdlib), you in fact have the same restrictions on evolving. It's not whether you're "official", it's the number of users you have. Second, why do these libraries always have to do some weird stuff that's not Pythonic at all, just to save a few characters in the usage? As far as I see, `construct` overloads `rtruediv` on their field names just so people could write "width" / Int8ub, instead of, idontknow, a normal field annotation such as `width: Unsigned(8)`? This is for me in the same bucket as `attr.ib` and stuff. <blockquote> Another thing with legacy is that you can’t drop old features. Just look at three ways of formatting since 3.6. The last one (f"{x + 1}") is just what the world need™. But we can’t drop old formatting and some people will still use percents, huh. I understand, that if I don’t want to use percents I can do it, but other people will, and I will see their code and become sad. </blockquote> Now I've become sad, because it's obvious that we've done a bad job explaining what various formatting techniques are for and when they should be used, letting people like you think that it is just some kind of Perlian Tim Toady mess. It isn't. It's just that the world is complicated, and we're solving many similar, yet distinct, problems. Yes, if you have a formatting message and names of some objects at the same time, and you want to make a nice string out of it (e.g. for printing), and then forget about it, `f"..."` is nice. But what if you want to translate that message? What if you want user to pick the level of detail? What if you want to have both machine-readable and human-readable output? Suddenly you see that yes, having the message separate from the names of objects going into it makes a lot of sense. "Ok", you say, "that's what .format is for, I guess. I still don't see why I need these weird % signs when I have my nice curly braces?" That requires us to go deeper in the complicated world of formatting minilanguages. Python wasn't the language that invented them, they existed a long time ago. See <a href="https://www.strfti.me">https://www.strfti.me</a> for example. The thing is, people already know how to say "give me dot-separated day, month and year in that order", and it would be very counterproductive if Python just reinvented the wheel, having something other than well-known %-codes. Another example is the regex minilanguage. It's not backward compatibility with previous versions of Python, it's side compatibility with a consensus between many mainstream languages. <blockquote> My opinion is that some things should be dropped, for clearing/simplifying future codebase. </blockquote> And many of them are dropped. Just with regard to stdlib, see PEP 594. Yes, now that we don't have a BDFL any more, it all goes a lot slower, but the things are happening. <blockquote> What can go wrong? I don’t get it, give me example. Even if you have no internet connection you can build self contained binaries with pyinstaller or provide wheel files. </blockquote> Ok, you made me curious. In the past I never actually were able to make pyinstaller work, but I said, maybe the things are better now. They aren't. <pre class='brush: python'>OSError: Python library not found: libpython3.10.so, libpython3.10m.so.1.0, libpython3.10m.so, libpython3.10mu.so.1.0, libpython3.10.so.1.0 This means your Python installation does not come with proper shared library files. This usually happens due to missing development package, or unsuitable build parameters of the Python installation. * On Debian/Ubuntu, you need to install Python development packages: * apt-get install python3-dev * apt-get install python-dev </pre> Of course, when I try the first one, I get the message that it is already the newest version (3.8.2-0ubuntu2). When I try the second one, it tries to install a bunch of 2.7 libraries, which 1) I don't want on my system out of principle, 2) I really doubt are what will solve the problem, since the app is written in Python 3. <blockquote> I really don’t get those who use urllib from stdlib and they know about requests, but want «no third dependencies». This is pretty weird. </blockquote> I don't know whether you cound me among those you don't get, but I think requests is too bloated, and that there's nothing wrong with urllib. Certificate management is a pain, dependencies are complicated, and requests isn't in the stdlib precisely because they reserve the right to make ideological choices over stability. Besides, Kenneth Rietz is not a nice person IMO. <blockquote> wheels solve problem with binaries and if package author will provide wheels for different platforms, than everything will be simple as pip install. Of course some «bad guys» did not, but for linux this is not a problem. For windows there is <a href="http://www.lfd.uci.edu/">http://www.lfd.uci.edu/</a>~gohlke/pythonlibs </blockquote> Yes, things are a bit better now than 5 years ago, but still there is a lot of pain in this space. See above for my experience with pyinstaller. I guess my attempt to produce a wheel would be equally unsuccessful.

oduvan on July 18, 2022, 8:53 p.m.
<blockquote> I was sure I replied to this at the time (6 years ago:) </blockquote> I wish I have your memory or the level of confidence (any of those two would work :) )

veky on July 19, 2022, 5:46 a.m.
As I <a href="https://py.checkio.org/mission/ghosts-age/publications/veky/python-3/finite/share/1eb17c0f9af8f148018c1b205a8a8b5a/#comment-18172">said before</a>, my memory is much more complete than true. :-)) And having my level of confidence is really not that hard. The real question is why you think it would be useful. :-)

oduvan on July 20, 2022, 10:07 a.m.
not every dream should be useful ;-P

oduvan on Jan. 16, 2017, 12:52 p.m.
Video Code Review <a href="https://www.youtube.com/watch?v=196f9o5KPqU">https://www.youtube.com/watch?v=196f9o5KPqU</a>

Dropbox

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.