In Re: Fill & Download for Free

Caveat: "Time saving" is a loose term -- especially for a large and/or growing codebase shared by many developers, you want to be careful that time saved by one person in writing code (e.g. with clever one-liners) doesn't grossly compromise time spent by other people in reading it, maintaining it, or otherwise working with it. This answer addresses the question in terms of broader productivity gains across a software team, as opposed to tips and tricks to hack up throwaway Python code as quickly as possible.For idiomatic Python, get comfortable with list comprehensions, generators, decorators, and context managers.-- List comprehensions are often the most idiomatic way of creating and transforming lists.-- Generators are most often used for defining iterators cleanly, usually with less memory overhead than alternative solutions.-- Decorators are great for common setup/cleanup code, like authentication, caching, or timing. Some built-in decorators like @classmethod and @property are very useful when working with classes. See also: What are common uses of Python decorators?-- Context managers are also handy for setup/cleanup code, for example in opening and closing files or tags (Quora uses these to make sure tags in HTML generated from WebNode are always closed). Another cool use case is setting some state for the duration of a code block, like flags for bypassing cache or for pipelining networked instructions.Familiarize yourself with all the methods of the most common data structures: lists, sets, dictionaries. Python has a lot of useful built-in functionality for working with these data structures, like slicing for lists or mathematical set operations with syntactic sugar for easy use on sets.Run code snippets at an iPython shell. Take full advantage of the fact that Python has a great interactive shell -- test out your code snippets quickly and fiddle with syntax or logic 'til they work, inspect your objects and functions, time function calls, etc. Make sure to use iPython and not the standard Python shell, as iPython is much more powerful and has nice features like tab-completion and %history.Use a linter like PyFlakes, PyLint, or PyChecker (comparison here: http://stackoverflow.com/questions/1428872/pylint-pychecker-or-pyflakes). Better yet, hook your linter into your editor so it checks as you go, similar to the code analysis and automatic error or warning highlighting that Java IDEs do.Familiarize yourself with Python's standard libraries, like urllib2, time and datetime, random, itertools, functools, re, collections, etc. See also: http://stackoverflow.com/questions/1453952/most-useful-python-modules-from-the-standard-libraryUse print statements and pdb to debug. print is particularly effective to use in Python since you don't need to compile your code again to start printing useful debug output. As for pdb, just dump this wherever you need to start investigating your running program: import pdb; pdb.set_trace(); and step through/in, continue, or print values as you please. See also: http://docs.python.org/library/pdb.htmlAdopt strong naming conventions to make your code more self-documenting. For example, if you have user ids and user objects in your codebase, always use *_user_id to refer to the former and *_user to refer to the latter, so it's always clear whether you're working with an int/long that is the id or an object that is the full user object. (This is an example that is particularly salient given Python's lack of static type annotations.) And unless you are using _ for i18n, use _ to indicate values to be thrown away.

Forget C, use C++, and learn STL.Note: I used STL and the C++ Standard Library interchangeably in this answer (though I probably shouldn't have). All the below structures and algorithms are in the C++ Standard Library, so it's what you need to learn.All of the following are must-knows:Data StructuresYou should know how to use all these + all their related functions + the time complexity of all these functions + (probably) how these functions workvectors (dynamic arrays) - know why push_back is amortized constant, not just that it is somaps and sets (red-black BSTs usually)unordered_maps and unordered_sets (hash tables)stacks/queues/deques/listspriority_queues (though I'd argue in an interview you should just use a map since it can do anything a priority_queue can do and more)You should also know what iterators are and how to use them with the above data structures and below algorithms, because you need to when using STL.Graphs are also important to know how to implement. For interviews, I usually use a vector of vectors for an adjacency-list (or even adjacency-matrix) representation just because it's fast to implement and usually is sufficient enough for you to write your algorithm, but alternatives are also feasible and some people may prefer building it from scratch using classes/structs/pointers. In many cases you can just forgo building the graph completely during an interview and assume you have it given, and in some cases you shouldn't build the graph at all, but rather have your algorithm dynamically generate the needed parts as it executes.AlgorithmsAside from all the related functions for the data structures above, the algorithm library is extremely important, and you should probably know every function in it as well as how it's implemented and its time/space complexity (because all of these are fair game for interview questions and you might be asked to implement them from scratch, though probably without as much generalization as STL). Here are the most important ones, though, in my opinion:sortstable_sortnth_elementlower_boundupper_boundmerge/inplace_mergenext_permutation/prev_permutationrotatereverseFamiliarizing yourself with the various math functions via the cmath library is probably also a good idea, just in case you need them, though if you need a tan function in an interview you can probably just tell your interviewer "I'm going to assume I have a function to calculate tan(x)" and it'd be fine.Pointers, Classes/Structs, and OOP ConceptsYou should know these at least enough to make a linked list and trie (which are the most useful data structure that is not implemented in STL) in about 5-7 minutes, complete with an insert (or search) function (and yes you may be asked to do this in an interview). You should be very comfortable with them - I doubt pointers and classes will be explicitly asked about (you might be asked about OOP Concepts though), but you may very well have to use them.Why You Should Know All ThisRead this answer for the full version: Jimmy Saade's answer to How do I get a job at Facebook or Google in 6 months?But basically, there are a lot of questions which rely on you knowing/having/being able to use these data structures/algorithms directly, without implementing them - especially sort and vectors, maps, and unordered_maps. Your solution to the asked question may require a hash table or heap to achieve a decent time complexity, or require that you sort the input - are you really going to implement these in the span of a 45 minute interview? No, and neither are you expected to. You're expected to make it easy on yourself and use STL, because the rest of the algorithm you're coding up is already going to be complex enough, plus you're never going to that in real life, and it shows you know not to re-make stuff that's already at your disposal.The exception to this is if you're explicitly asked to implement something already in STL. For example, if you're asked to implement the next_permutation function, you obviously shouldn't use the STL function, though you could (and probably should) note that there is such a function in STL. But basically the interviewer is looking for you to figure out the algorithm and code it yourself.The theory version of this question is answered here: Jimmy Saade's answer to What should I know from CLRS 3rd edition book if my aim is to get into Google? and so they kind of go together in that what's listed in this answer is an implementation of the concepts listed in the other.

Lip service, huh? How are these lip service?1.The highly successful 2012 campaign to get the FBI to re-define rape to include the rape of men, oral and anal rape, and any kind of penetration instead of just vaginal penetration of women. The Feminist Majority Foundation also campaigned heavily to do something about the refusal to process rape kits, so that rape could be prosecuted in a timely manner.Feminist Majority FoundationAttorney General Eric Holder Announces Revisions to the Uniform Crime Report’s Definition of Rape2. Ruth Bader Ginsberg's successful arguments before the Supreme Court in the 1970s to extend Social Security and survivor's benefits to calculating the income of working women and men, so that widowed men and children were able to collect based on the income of the wife if she dies, not just the husband's income if he does.FindLaw | Cases and Codes