Python-ideas November 2013

python-ideas@python.org

78 participants
45 discussions

Specify number of items to allocate for array.array() constructor

by Sven Rahmann

At the moment, the array module of the standard library allows to create arrays of different numeric types and to initialize them from an iterable (eg, another array). What's missing is the possiblity to specify the final size of the array (number of items), especially for large arrays. I'm thinking of suffix arrays (a text indexing data structure) for large texts, eg the human genome and its reverse complement (about 6 billion characters from the alphabet ACGT). The suffix array is a long int array of the same size (8 bytes per number, so it occupies about 48 GB memory). At the moment I am extending an array in chunks of several million items at a time at a time, which is slow and not elegant. The function below also initializes each item in the array to a given value (0 by default). Is there a reason why there the array.array constructor does not allow to simply specify the number of items that should be allocated? (I do not really care about the contents.) Would this be a worthwhile addition to / modification of the array module? My suggestions is to modify array generation in such a way that you could pass an iterator (as now) as second argument, but if you pass a single integer value, it should be treated as the number of items to allocate. Here is my current workaround (which is slow): def filled_array(typecode, n, value=0, bsize=(1<<22)): """returns a new array with given typecode (eg, "l" for long int, as in the array module) with n entries, initialized to the given value (default 0) """ a = array.array(typecode, [value]*bsize) x = array.array(typecode) r = n while r >= bsize: x.extend(a) r -= bsize x.extend([value]*r) return x

7 months

Implicit string literal concatenation considered harmful?

by Guido van Rossum

I just spent a few minutes staring at a bug caused by a missing comma -- I got a mysterious argument count error because instead of foo('a', 'b') I had written foo('a' 'b'). This is a fairly common mistake, and IIRC at Google we even had a lint rule against this (there was also a Python dialect used for some specific purpose where this was explicitly forbidden). Now, with modern compiler technology, we can (and in fact do) evaluate compile-time string literal concatenation with the '+' operator, so there's really no reason to support 'a' 'b' any more. (The reason was always rather flimsy; I copied it from C but the reason why it's needed there doesn't really apply to Python, as it is mostly useful inside macros.) Would it be reasonable to start deprecating this and eventually remove it from the language? -- --Guido van Rossum (python.org/~guido)

2 years, 6 months

Re: [Python-ideas] Update the required C compiler for Windows to a supported version.

by Steve Dower

Stephen J. Turnbull wrote: > Vernon D. Cole writes: > >> I cannot compile a Python extension module with any Microsoft compiler >> I can obtain. > > Your pain is understood, but it's not simple to address it. FWIW, I'm working on making the compiler easily obtainable. The VS 2008 link that was posted is unofficial, and could theoretically disappear at any time (I'm not in control of that), but the Windows SDK for Windows 7 and .NET 3.5 SP1 (http://www.microsoft.com/en-us/download/details.aspx?id=3138) should be around for as long as Windows 7 is supported. The correct compiler (VC9) is included in this SDK, but unfortunately does not install the vcvarsall.bat file that distutils expects. (Though it's pretty simple to add one that will switch on %1 and call the correct vcvars(86|64|...).bat.) The SDK needed for Python 3.3 and 3.4 (VC10) is even worse - there are many files missing. I'm hoping we'll be able to set up some sort of downloadable package/tool that will fix this. While we'd obviously love to move CPython onto our latest compilers, it's simply not possible (for good reason). Python 3.4 is presumably locked to VC10, but hopefully 3.5 will be able to use whichever version is current when that decision is made. > The basic problem is that the ABI changes. Therefore it's going to require > a complete new set of *all* C extensions for Windows, and the duplication > of download links for all those extensions from quite a few different vendors > is likely to confuse a lot of users. Specifically, the CRT changes. The CRT is an interesting mess of data structures that are exposed in header files, which means while you can have multiple CRTs loaded, they cannot touch each other's data structures at all or things will go bad/crash, and there's no nice way to set it up to avoid this (my colleague who currently owns MSVCRT suggested a not-very-nice way to do it, but I don't think it's going to be reliable enough). Python's stable ABI helps, but does not solve this problem. The file APIs are the worst culprits. The layout of FILE* objects can and do change between CRT versions, and file descriptors are simply indices into an array of these objects that is exposed through macros rather than function calls. As a result, you cannot mix either FILE pointers or file descriptors between CRTs. The only safe option is to build with the matching CRT, and for MSVCRT, this means with the matching compiler. It's unfortunate, and the responsible teams are well aware of the limitation, but it's history at this point, so we have no choice but to work with it. Cheers, Steve

6 years, 6 months

while conditional in list comprehension ??

by Wolfgang Maier

Dear all, I guess this is so obvious that someone must have suggested it before: in list comprehensions you can currently exclude items based on the if conditional, e.g.: [n for n in range(1,1000) if n % 4 == 0] Why not extend this filtering by allowing a while statement in addition to if, as in: [n for n in range(1,1000) while n < 400] Trivial effect, I agree, in this example since you could achieve the same by using range(1,400), but I hope you get the point. This intuitively understandable extension would provide a big speed-up for sorted lists where processing all the input is unnecessary. Consider this: some_names=["Adam", "Andrew", "Arthur", "Bob", "Caroline","Lancelot"] # a sorted list of names [n for n in some_names if n.startswith("A")] # certainly gives a list of all names starting with A, but . [n for n in some_names while n.startswith("A")] # would have saved two comparisons Best, Wolfgang

6 years, 7 months

Replacing the if __name__ == "__main__" idiom (was Re: making a module callable)

by Eric Snow

On Fri, Nov 22, 2013 at 1:44 PM, Gregory P. Smith <greg(a)krypto.org> wrote: > It'd be nice to formalize a way to get rid of the __name__ == '__main__' > idiom as well in the long long run. Sure everyone's editor types that for > them now but it's still a wart. Anyways, digressing... ;) This has come up before and is the subject of several PEPS. [1][2] The current idiom doesn't bother me too much as I try not to have files that are both scripts and modules. However, Python doesn't make the distinction all that clear nor does it do much to encourage people to keep the two separate. I'd prefer improvements in both those instead, but haven't had the time for any concrete proposal. FWIW, aside from the idiom there are other complications that arise from a module that also gets loaded in __main__ (run as a script). See PEP 395 [3]. -erc [1] http://www.python.org/dev/peps/pep-0299/ [2] http://www.python.org/dev/peps/pep-3122/ [3] http://www.python.org/dev/peps/pep-0395/ (sort of related)

6 years, 9 months

string codes & substring equality

by spir

Hello, Coming back to python after a long time. My present project is about (yet another) top-down matching / parsing lib. There are 2 issues that, I guess, may be rather easily solved by simple string methods. The core point is that any scanning / parsing process ends up, a the lowest level, constantly comparing either single-char (rather single-code) substrings or constant (literal) substrings of the source string. This is the only operation which, when successful, actually advances in the source. Thus, it is certainly worth having it efficient, or at the minimum not having it needlessly inefficient. I suppose the same functionalities can be highly useful in various other use cases of text processing. Note again that I'm rediscovering Python (with some pleasure :-), thus may miss known solutions -- but I asked on the tutor mailing list. In both cases, I guess ordinary idiomatic Python code actually _creates_ a new string object, as a substring of length 1 or more, which is otherwise useless; for instance: if s[i] == char: # match ok -- object s[i] unneeded if s[i:j] == substr: # match ok -- object s[i:j] unneeded What is actually needed is just to check for equality (or another check about a code, see below). The case of single-code checking appears when (1) a substring happens to hold a single code (meaning it represents a simple or precomposed unicode char) (2) when matching a char from a given set, range, or more complex class (eg in regex [a-zA-Z0-9_-']). In all cases, what we want is tocheck the code: compare it to a constant value, check whether it belongs to a set of value, or lies inside a given range. We need the code --not a single-code string. Ideally, I'd like expressions like: c = s.code(i) # or s.ord(i) or s.ucode(i) [3] # and then one of: if c = code: # match ok if c in codes: # match ok if c >= code1 and c <= code2: # match ok The builtin function ord(char) does not do the job, since it only works for a single-char string. We would again need to create a new string, with ord(s[i]). The right solution apparently is a string method like code(self, i) giving the code at an arbitrary index. I guess this is trivial. I'm surprised it does not exist; maybe some may think this is a symptom there is no strong need for it; instead, I guess people routinely use a typical Python idiom without even noticing it creates a unneeded string object. [2] [3] What do you think? A second need is checking substring equality against constant substrings of arbitrary sizes. This is similar to startswith & endswith, except at any code index in the source string; a generalisation. In C implementation, it would probably delegate to memcomp, with a start pointer set to p_source+i. On the Python side, it may be a string method like sub_equals(self, substr, i). Choose you preferred name ;-). [1] [4] if s.sub_equals(substr, i): # match ok What do you think? (bis) Thank you, Denis [1] I am unsure whether an end index is useful, actually I don't really understand its usage for startswith & endswith neither. [2] Actually, the compiler, if smart enough, may eliminate this object construction and just check the code; does it? Anyway, I think it is not that easy in the cases of ranges & sets. [3] As a side-note, 'ord' is in my view a misnomer, since character codes are not ordinals, with significant order, but nominals, plain numerical codes which only need to be all distinct; they are kinds of id's. For unicode, I call them 'ucodes', an idea I stole somewhere. But I would be happy is the method is called 'ord' anyway, since the term is established in the Python community. [4] Would such a new method make startswith & endswith unneeded?

6 years, 9 months

string.Template v2

by anatoly techtonik

http://docs.python.org/2/library/string.html#template-strings ## Original Idea stdlib lacks the most popular basic variable extension syntax "{{ variable }}" that can be found in Django [1], Jinja2 [2] and other templating engines [3]. ## stdlib Analysis string.Template syntax is ancient (dates back to Python 2.4 from 9 years ago). I haven't seen a template like this for a long time. ## Scope of Enhancement st = 'Hello {{world}}.' world = 'is not enough' t = Template(string, style='brace') t.render(locals()) ## Links 1. https://docs.djangoproject.com/en/dev/topics/templates/#variables 2. http://jinja.pocoo.org/docs/templates/#variables 3. http://mustache.github.io/ ## Feature Creeping # Allow to override {{ }} symbols to make it more generic. # `foo.bar` attribute lookup for 2D (nested) structures. Questions is it has to be supported: `foo.bar` in Django does dictionary lookup first, then attribute lookup `foo.bar` in Jinja2 does attribute lookup first I am not sure which is better. I definitely don't want some method or property on a dict passed to render() method to hide dict value. -- anatoly t.

6 years, 9 months

Refactor Assertions Out of unittest.TestCase

by Daniel Watkins

Hello all, I would like to propose refactoring the assertions out of unittest.TestCase. As code speaks louder than words, you can see my initial work at https://github.com/OddBloke/cpython. The aim of the refactor is: (a) to reduce the amount of repeated code in the assertions, (b) to provide a clearer framework for third-party assertions to follow, and (c) to make it easier to split gnarly assertions in to an easier-to-digest form. My proposed implementation (as seen in the code above) is to move each assertion in to its own class. There will be a shared superclass (called Assert currently; see [0]) implementing the template pattern (look at __call__ on line 69), meaning that each assertion only has to concern itself with its unique aspects: what makes it fail, and how that specific failure should be presented. To maintain the current TestCase interface (all of the tests pass in my branch), the existing assert* methods instantiate the Assert sub-classes on each call with a context that captures self.longMessage, self.failureException, self.maxDiff, and self._diffThreshold from the TestCase instance. Other potential aims include eventually deprecating the assertion methods, providing a framework for custom assertions to hook in to, and providing assertion functions (a la nose.tools) with a default context set. This proposal would help address #18054[1] as the new assertions could just be implemented as separate classes (and not included in the TestCase.assert* API); people who wanted to use them could just instantiate them themselves. I’d love some feedback on this proposal (and the implementation thus far). Cheers, Dan (Odd_Bloke) [0] https://github.com/OddBloke/cpython/blob/master/Lib/unittest/assertions/__i… [1] http://bugs.python.org/issue18054

6 years, 9 months

os.path.join

by anatoly techtonik

>>> os.path.join('/static', '/styles/largestyles.css') '/styles/largestyles.css' Is it only me who thinks that the code above is wrong?

6 years, 9 months

Improving Clarity of re Module

by Alex Seewald

For a match object, m, m.group(0) is the semantics for accessing the entire span of the match. For newcomers to regular expressions who are not familiar with the concept of a 'group', the name group(0) is counter-intuitive. A more natural-language-esque alias to group(0), perhaps 'matchSpan', could reduce the time novices spend from idea to working code. Of course, this convenience would introduce a bit of complexity to the codebase, so it may or may not be worth it to add an alias to group(0). What do people think? -- Alex Seewald

6 years, 10 months

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

Python-ideas November 2013