python编码规范

转载以记之。
PEP: 8
Title: Style Guide for Python Code
Version: $Revision$
Last-Modified: $Date$
Author:
guido@python.org
(Guido van Rossum),
   
barry@python.org
(Barry Warsaw)
Status: Active
Type: Informational
Created: 05-Jul-2001
Post-History: 05-Jul-2001
Introduction
    This document gives coding conventions for the Python code comprising the
    standard library in the main Python distribution.  Please see the
    companion informational PEP describing style guidelines for the C code in
    the C implementation of Python[1].
    This document was adapted from Guido's original Python Style Guide
    essay[2], with some additions from Barry's style guide[5].  Where there's
    conflict, Guido's style rules for the purposes of this PEP.  This PEP may
    still be incomplete (in fact, it may never be finished ).
A Foolish Consistency is the Hobgoblin of Little Minds
    One of Guido's key insights is that code is read much more often than it
    is written.  The guidelines provided here are intended to improve the
    readability of code and make it consistent across the wide spectrum of
    Python code.  As PEP 20 [6] says, "Readability counts".
    A style guide is about consistency.  Consistency with this style guide is
    important.  Consistency within a project is more important. Consistency
    within one module or function is most important.
    But most importantly: know when to be inconsistent -- sometimes the style
    guide just doesn't apply.  When in doubt, use your best judgment.  Look
    at other examples and decide what looks best.  And don't hesitate to ask!
    Two good reasons to break a particular rule:
    (1) When applying the rule would make the code less readable, even for
        someone who is used to reading code that follows the rules.
    (2) To be consistent with surrounding code that also breaks it (maybe for
        historic reasons) -- although this is also an opportunity to clean up
        someone else's mess (in true XP style).
Code lay-out
  Indentation
    Use 4 spaces per indentation level.
    For really old code that you don't want to mess up, you can continue to
    use 8-space tabs.
  Tabs or Spaces?
    Never mix tabs and spaces.
    The most popular way of indenting Python is with spaces only.  The
    second-most popular way is with tabs only.  Code indented with a mixture
    of tabs and spaces should be converted to using spaces exclusively.  When
    invoking the Python command line interpreter with the -t option, it issues
    warnings about code that illegally mixes tabs and spaces.  When using -tt
    these warnings become errors.  These options are highly recommended!
    For new projects, spaces-only are strongly recommended over tabs.  Most
    editors have features that make this easy to do.
  Maximum Line Length
    Limit all lines to a maximum of 79 characters.
    There are still many devices around that are limited to 80 character
    lines; plus, limiting windows to 80 characters makes it possible to have
    several windows side-by-side.  The default wrapping on such devices looks
    ugly.  Therefore, please limit all lines to a maximum of 79 characters.
    For flowing long blocks of text (docstrings or comments), limiting the
    length to 72 characters is recommended.
    The preferred way of wrapping long lines is by using Python's implied line
    continuation inside parentheses, brackets and braces.  If necessary, you
    can add an extra pair of parentheses around an expression, but sometimes
    using a backslash looks better.  Make sure to indent the continued line
    appropriately.  Some examples:
    class Rectangle(Blob):
        def __init__(self, width, height,
                     color='black', emphasis=None, highlight=0):
            if width == 0 and height == 0 and \
               color == 'red' and emphasis == 'strong' or \
               highlight > 100:
                raise ValueError("sorry, you lose")
            if width == 0 and height == 0 and (color == 'red' or
                                               emphasis is None):
                raise ValueError("I don't think so")
            Blob.__init__(self, width, height,
                          color, emphasis, highlight)
  Blank Lines
    Separate top-level function and class definitions with two blank lines.
    Method definitions inside a class are separated by a single blank line.
    Extra blank lines may be used (sparingly) to separate groups of related
    functions.  Blank lines may be omitted between a bunch of related
    one-liners (e.g. a set of dummy implementations).
    Use blank lines in functions, sparingly, to indicate logical sections.
    Python accepts the control-L (i.e. ^L) form feed character as whitespace;
    Many tools treat these characters as page separators, so you may use them
    to separate pages of related sections of your file.
  Encodings (PEP 263)
    Code in the core Python distribution should aways use the ASCII or Latin-1
    encoding (a.k.a. ISO-8859-1).
    Files using ASCII should not have a coding cookie.  Latin-1 should only be
    used when a comment or docstring needs to mention an author name that
    requires Latin-1; otherwise, using \x escapes is the preferred way to
    include non-ASCII data in string literals.
Imports
    - Imports should usually be on separate lines, e.g.:
        Yes: import os
             import sys
        No:  import sys, os
      it's okay to say this though:
        from subprocess import Popen, PIPE
    - Imports are always put at the top of the file, just after any module
      comments and docstrings, and before module globals and constants.
      Imports should be grouped in the following order:
      1. standard library imports
      2. related third party imports
      3. local application/library specific imports
      You should put a blank line between each group of imports.
      Put any relevant __all__ specification after the imports.
    - Relative imports for intra-package imports are highly discouraged.
      Always use the absolute package path for all imports.
      Even now that PEP 328 [7] is fully implemented in Python 2.5,
      its style of explicit relative imports is actively discouraged;
      absolute imports are more portable and usually more readable.
    - When importing a class from a class-containing module, it's usually okay
      to spell this
        from myclass import MyClass
        from foo.bar.yourclass import YourClass
      If this spelling causes local name clashes, then spell them
        import myclass
        import foo.bar.yourclass
      and use "myclass.MyClass" and "foo.bar.yourclass.YourClass"
Whitespace in Expressions and Statements
  Pet Peeves
    Avoid extraneous whitespace in the following situations:
    - Immediately inside parentheses, brackets or braces.
      Yes: spam(ham[1], {eggs: 2})
      No:  spam( ham[ 1 ], { eggs: 2 } )
    - Immediately before a comma, semicolon, or colon:
      Yes: if x == 4: print x, y; x, y = y, x
      No:  if x == 4 : print x , y ; x , y = y , x
    - Immediately before the open parenthesis that starts the argument
      list of a function call:
      Yes: spam(1)
      No:  spam (1)
    - Immediately before the open parenthesis that starts an indexing or
      slicing:
      Yes: dict['key'] = list[index]
      No:  dict ['key'] = list [index]
    - More than one space around an assignment (or other) operator to
      align it with another.
      Yes:
          x = 1
          y = 2
          long_variable = 3
      No:
          x             = 1
          y             = 2
          long_variable = 3
  Other Recommendations
    - Always surround these binary operators with a single space on
      either side: assignment (=), augmented assignment (+=, -= etc.),
      comparisons (==, , !=, , =, in, not in, is, is not),
      Booleans (and, or, not).
    - Use spaces around arithmetic operators:
      Yes:
          i = i + 1
          submitted += 1
          x = x * 2 - 1
          hypot2 = x * x + y * y
          c = (a + b) * (a - b)
      No:
          i=i+1
          submitted +=1
          x = x*2 - 1
          hypot2 = x*x + y*y
          c = (a+b) * (a-b)
    - Don't use spaces around the '=' sign when used to indicate a
      keyword argument or a default parameter value.
      Yes:
          def complex(real, imag=0.0):
              return magic(r=real, i=imag)
      No:
          def complex(real, imag = 0.0):
              return magic(r = real, i = imag)
    - Compound statements (multiple statements on the same line) are
      generally discouraged.
      Yes:
          if foo == 'blah':
              do_blah_thing()
          do_one()
          do_two()
          do_three()
      Rather not:
          if foo == 'blah': do_blah_thing()
          do_one(); do_two(); do_three()
    - While sometimes it's okay to put an if/for/while with a small
      body on the same line, never do this for multi-clause
      statements.  Also avoid folding such long lines!
      Rather not:
          if foo == 'blah': do_blah_thing()
          for x in lst: total += x
          while t
      Definitely not:
          if foo == 'blah': do_blah_thing()
          else: do_non_blah_thing()
          try: something()
          finally: cleanup()
          do_one(); do_two(); do_three(long, argument,
                                       list, like, this)
          if foo == 'blah': one(); two(); three()
Comments
    Comments that contradict the code are worse than no comments.  Always make
    a priority of keeping the comments up-to-date when the code changes!
    Comments should be complete sentences.  If a comment is a phrase or
    sentence, its first word should be capitalized, unless it is an identifier
    that begins with a lower case letter (never alter the case of
    identifiers!).
    If a comment is short, the period at the end can be omitted.  Block
    comments generally consist of one or more paragraphs built out of complete
    sentences, and each sentence should end in a period.
    You should use two spaces after a sentence-ending period.
    When writing English, Strunk and White apply.
    Python coders from non-English speaking countries: please write
    your comments in English, unless you are 120% sure that the code
    will never be read by people who don't speak your language.
  Block Comments
    Block comments generally apply to some (or all) code that follows them,
    and are indented to the same level as that code.  Each line of a block
    comment starts with a # and a single space (unless it is indented text
    inside the comment).
    Paragraphs inside a block comment are separated by a line containing a
    single #.
  Inline Comments
    Use inline comments sparingly.
    An inline comment is a comment on the same line as a statement.  Inline
    comments should be separated by at least two spaces from the statement.
    They should start with a # and a single space.
    Inline comments are unnecessary and in fact distracting if they state
    the obvious.  Don't do this:
        x = x + 1                 # Increment x
    But sometimes, this is useful:
        x = x + 1                 # Compensate for border
Documentation Strings
    Conventions for writing good documentation strings (a.k.a. "docstrings")
    are immortalized in PEP 257 [3].
    - Write docstrings for all public modules, functions, classes, and
      methods.  Docstrings are not necessary for non-public methods, but you
      should have a comment that describes what the method does.  This comment
      should appear after the "def" line.
    - PEP 257 describes good docstring conventions.  Note that most
      importantly, the """ that ends a multiline docstring should be on a line
      by itself, and preferably preceded by a blank line, e.g.:
      """Return a foobang
      Optional plotz says to frobnicate the bizbaz first.
      """
    - For one liner docstrings, it's okay to keep the closing """ on the same
      line.
Version Bookkeeping
    If you have to have Subversion, CVS, or RCS crud in your source file, do
    it as follows.
        __version__ = "$Revision$"
        # $Source$
    These lines should be included after the module's docstring, before any
    other code, separated by a blank line above and below.
Naming Conventions
    The naming conventions of Python's library are a bit of a mess, so we'll
    never get this completely consistent -- nevertheless, here are the
    currently recommended naming standards.  New modules and packages
    (including third party frameworks) should be written to these standards,
    but where an existing library has a different style, internal consistency
    is preferred.
  Descriptive: Naming Styles
    There are a lot of different naming styles.  It helps to be able to
    recognize what naming style is being used, independently from what they
    are used for.
    The following naming styles are commonly distinguished:
    - b (single lowercase letter)
    - B (single uppercase letter)
    - lowercase
    - lower_case_with_underscores
    - UPPERCASE
    - UPPER_CASE_WITH_UNDERSCORES
    - CapitalizedWords (or CapWords, or CamelCase -- so named because
      of the bumpy look of its letters[4]).  This is also sometimes known as
      StudlyCaps.
      Note: When using abbreviations in CapWords, capitalize all the letters
      of the abbreviation.  Thus HTTPServerError is better than
      HttpServerError.
    - mixedCase (differs from CapitalizedWords by initial lowercase
      character!)
    - Capitalized_Words_With_Underscores (ugly!)
    There's also the style of using a short unique prefix to group related
    names together.  This is not used much in Python, but it is mentioned for
    completeness.  For example, the os.stat() function returns a tuple whose
    items traditionally have names like st_mode, st_size, st_mtime and so on.
    (This is done to emphasize the correspondence with the fields of the
    POSIX system call struct, which helps programmers familiar with that.)
    The X11 library uses a leading X for all its public functions.  In Python,
    this style is generally deemed unnecessary because attribute and method
    names are prefixed with an object, and function names are prefixed with a
    module name.
    In addition, the following special forms using leading or trailing
    underscores are recognized (these can generally be combined with any case
    convention):
    - _single_leading_underscore: weak "internal use" indicator.  E.g. "from M
      import *" does not import objects whose name starts with an underscore.
    - single_trailing_underscore_: used by convention to avoid conflicts with
      Python keyword, e.g.
      Tkinter.Toplevel(master, class_='ClassName')
    - __double_leading_underscore: when naming a class attribute, invokes name
      mangling (inside class FooBar, __boo becomes _FooBar__boo; see below).
    - __double_leading_and_trailing_underscore__: "magic" objects or
      attributes that live in user-controlled namespaces.  E.g. __init__,
      __import__ or __file__.  Never invent such names; only use them
      as documented.
  Prescriptive: Naming Conventions
    Names to Avoid
      Never use the characters `l' (lowercase letter el), `O' (uppercase
      letter oh), or `I' (uppercase letter eye) as single character variable
      names.
      In some fonts, these characters are indistinguishable from the numerals
      one and zero.  When tempted to use `l', use `L' instead.
    Module Names
      Modules should have short, lowercase names, without underscores.
      Since module names are mapped to file names, and some file systems are
      case insensitive and truncate long names, it is important that module
      names be chosen to be fairly short -- this won't be a problem on Unix,
      but it may be a problem when the code is transported to older Mac or
      Windows versions, or DOS.
      When an extension module written in C or C++ has an accompanying Python
      module that provides a higher level (e.g. more object oriented)
      interface, the C/C++ module has a leading underscore (e.g. _socket).
      Like modules, Python packages should have short, all-lowercase names,
      without underscores.
    Class Names
      Almost without exception, class names use the CapWords convention.
      Classes for internal use have a leading underscore in addition.
    Exception Names
      Because exceptions should be classes, the class naming convention
      applies here.  However, you should use the suffix "Error" on your
      exception names (if the exception actually is an error).
    Global Variable Names
      (Let's hope that these variables are meant for use inside one module
      only.)  The conventions are about the same as those for functions.
      Modules that are designed for use via "from M import *" should use the
      __all__ mechanism to prevent exporting globals, or use the the older
      convention of prefixing such globals with an underscore (which you might
      want to do to indicate these globals are "module non-public").
    Function Names
      Function names should be lowercase, with words separated by underscores
      as necessary to improve readability.
      mixedCase is allowed only in contexts where that's already the
      prevailing style (e.g. threading.py), to retain backwards compatibility.
    Function and method arguments
      Always use 'self' for the first argument to instance methods.
      Always use 'cls' for the first argument to class methods.
      If a function argument's name clashes with a reserved keyword, it is
      generally better to append a single trailing underscore rather than use
      an abbreviation or spelling corruption.  Thus "print_" is better than
      "prnt".  (Perhaps better is to avoid such clashes by using a synonym.)
    Method Names and Instance Variables
      Use the function naming rules: lowercase with words separated by
      underscores as necessary to improve readability.
      Use one leading underscore only for non-public methods and instance
      variables.
      To avoid name clashes with subclasses, use two leading underscores to
      invoke Python's name mangling rules.
      Python mangles these names with the class name: if class Foo has an
      attribute named __a, it cannot be accessed by Foo.__a.  (An insistent
      user could still gain access by calling Foo._Foo__a.)  Generally, double
      leading underscores should be used only to avoid name conflicts with
      attributes in classes designed to be subclassed.
      Note: there is some controversy about the use of __names (see below).
    Designing for inheritance
      Always decide whether a class's methods and instance variables
      (collectively: "attributes") should be public or non-public.  If in
      doubt, choose non-public; it's easier to make it public later than to
      make a public attribute non-public.
      Public attributes are those that you expect unrelated clients of your
      class to use, with your commitment to avoid backward incompatible
      changes.  Non-public attributes are those that are not intended to be
      used by third parties; you make no guarantees that non-public attributes
      won't change or even be removed.
      We don't use the term "private" here, since no attribute is really
      private in Python (without a generally unnecessary amount of work).
      Another category of attributes are those that are part of the "subclass
      API" (often called "protected" in other languages).  Some classes are
      designed to be inherited from, either to extend or modify aspects of the
      class's behavior.  When designing such a class, take care to make
      explicit decisions about which attributes are public, which are part of
      the subclass API, and which are truly only to be used by your base
      class.
      With this in mind, here are the Pythonic guidelines:
      - Public attributes should have no leading underscores.
      - If your public attribute name collides with a reserved keyword, append
        a single trailing underscore to your attribute name.  This is
        preferable to an abbreviation or corrupted spelling.  (However,
        notwithstanding this rule, 'cls' is the preferred spelling for any
        variable or argument which is known to be a class, especially the
        first argument to a class method.)
        Note 1: See the argument name recommendation above for class methods.
      - For simple public data attributes, it is best to expose just the
        attribute name, without complicated accessor/mutator methods.  Keep in
        mind that Python provides an easy path to future enhancement, should
        you find that a simple data attribute needs to grow functional
        behavior.  In that case, use properties to hide functional
        implementation behind simple data attribute access syntax.
        Note 1: Properties only work on new-style classes.
        Note 2: Try to keep the functional behavior side-effect free, although
        side-effects such as caching are generally fine.
        Note 3: Avoid using properties for computationally expensive
        operations; the attribute notation makes the caller believe
        that access is (relatively) cheap.
      - If your class is intended to be subclassed, and you have attributes
        that you do not want subclasses to use, consider naming them with
        double leading underscores and no trailing underscores.  This invokes
        Python's name mangling algorithm, where the name of the class is
        mangled into the attribute name.  This helps avoid attribute name
        collisions should subclasses inadvertently contain attributes with the
        same name.
        Note 1: Note that only the simple class name is used in the mangled
        name, so if a subclass chooses both the same class name and attribute
        name, you can still get name collisions.
        Note 2: Name mangling can make certain uses, such as debugging and
        __getattr__(), less convenient.  However the name mangling algorithm
        is well documented and easy to perform manually.
        Note 3: Not everyone likes name mangling.  Try to balance the
        need to avoid accidental name clashes with potential use by
        advanced callers.
Programming Recommendations
    - Code should be written in a way that does not disadvantage other
      implementations of Python (PyPy, Jython, IronPython, Pyrex, Psyco,
      and such).
      For example, do not rely on CPython's efficient implementation of
      in-place string concatenation for statements in the form a+=b or a=a+b.
      Those statements run more slowly in Jython.  In performance sensitive
      parts of the library, the ''.join() form should be used instead.  This
      will assure that concatenation occurs in linear time across various
      implementations.
    - Comparisons to singletons like None should always be done with
      'is' or 'is not', never the equality operators.
      Also, beware of writing "if x" when you really mean "if x is not None"
      -- e.g. when testing whether a variable or argument that defaults to
      None was set to some other value.  The other value might have a type
      (such as a container) that could be false in a boolean context!
    - Use class-based exceptions.
      String exceptions in new code are strongly discouraged, as they will
      eventually (in Python 2.5) be deprecated and then (in Python 3000 or
      perhaps sooner) removed.
      Modules or packages should define their own domain-specific base
      exception class, which should be subclassed from the built-in Exception
      class.  Always include a class docstring.  E.g.:
        class MessageError(Exception):
            """Base class for errors in the email package."""
      Class naming conventions apply here, although you should add the suffix
      "Error" to your exception classes, if the exception is an error.
      Non-error exceptions need no special suffix.
    - When raising an exception, use "raise ValueError('message')" instead of
      the older form "raise ValueError, 'message'".
      The paren-using form is preferred because when the exception arguments
      are long or include string formatting, you don't need to use line
      continuation characters thanks to the containing parentheses.  The older
      form will be removed in Python 3000.
    - Use string methods instead of the string module.
      String methods are always much faster and share the same API with
      unicode strings.  Override this rule if backward compatibility with
      Pythons older than 2.0 is required.
    - Use ''.startswith() and ''.endswith() instead of string slicing to check
      for prefixes or suffixes.
      startswith() and endswith() are cleaner and less error prone.  For
      example:
        Yes: if foo.startswith('bar'):
        No:  if foo[:3] == 'bar':
      The exception is if your code must work with Python 1.5.2 (but let's
      hope not!).
    - Object type comparisons should always use isinstance() instead
      of comparing types directly.
        Yes: if isinstance(obj, int):
        No:  if type(obj) is type(1):
      When checking if an object is a string, keep in mind that it might be a
      unicode string too!  In Python 2.3, str and unicode have a common base
      class, basestring, so you can do:
        if isinstance(obj, basestring):
      In Python 2.2, the types module has the StringTypes type defined for
      that purpose, e.g.:
        from types import StringTypes
        if isinstance(obj, StringTypes):
      In Python 2.0 and 2.1, you should do:
        from types import StringType, UnicodeType
        if isinstance(obj, StringType) or \
           isinstance(obj, UnicodeType) :
    - For sequences, (strings, lists, tuples), use the fact that empty
      sequences are false.
      Yes: if not seq:
           if seq:
      No: if len(seq)
          if not len(seq)
    - Don't write string literals that rely on significant trailing
      whitespace.  Such trailing whitespace is visually indistinguishable and
      some editors (or more recently, reindent.py) will trim them.
    - Don't compare boolean values to True or False using ==
        Yes:   if greeting:
        No:    if greeting == True:
        Worse: if greeting is True:
References
    [1] PEP 7, Style Guide for C Code, van Rossum
    [2]
http://www.python.org/doc/essays/styleguide.html
    [3] PEP 257, Docstring Conventions, Goodger, van Rossum
    [4]
http://www.wikipedia.com/wiki/CamelCase
    [5] Barry's GNU Mailman style guide
        
http://barry.warsaw.us/software/STYLEGUIDE.txt
    [6] PEP 20, The Zen of Python
    [7] PEP 328, Imports: Multi-Line and Absolute/Relative
Copyright
    This document has been placed in the public domain.

Local Variables:
mode: indented-text
indent-tabs-mode: nil
End: