Python code minifier (Python recipe) by Dan McDougall

ActiveState Code (http://code.activestate.com/recipes/576704/)

#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
#       pyminifier.py
#
#       Copyright 2009 Dan McDougall <YouKnowWho@YouKnowWhat.com>
#
#       This program is free software; you can redistribute it and/or modify
#       it under the terms of the GNU General Public License as published by
#       the Free Software Foundation; Version 3 of the License
#
#       This program is distributed in the hope that it will be useful,
#       but WITHOUT ANY WARRANTY; without even the implied warranty of
#       MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#       GNU General Public License for more details.
#
#       You should have received a copy of the GNU General Public License
#       along with this program; if not, the license can be downloaded here:
#
#       http://www.gnu.org/licenses/gpl.html

# Meta
__version__ = '1.4.1'
__license__ = "GNU General Public License (GPL) Version 3"
__version_info__ = (1, 4, 1)
__author__ = 'Dan McDougall <YouKnowWho@YouKnowWhat.com>'

"""
**Python Minifier:**  Reduces the size of (minifies) Python code for use on
embedded platforms.

Performs the following:
     - Removes docstrings.
     - Removes comments.
     - Minimizes code indentation.
     - Joins multiline pairs of parentheses, braces, and brackets (and removes extraneous whitespace within).
     - Preserves shebangs and encoding info (e.g. "# -- coding: utf-8 --").

Various examples and edge cases are sprinkled throughout the pyminifier code so
that it can be tested by minifying itself.  The way to test is thus:

.. code-block:: bash

    $ python pyminifier.py pyminifier.py > minified_pyminifier.py
    $ python minified_pyminifier.py pyminifier.py > this_should_be_identical.py
    $ diff minified_pyminifier.py this_should_be_identical.py
    $

If you get an error executing minified_pyminifier.py or
'this_should_be_identical.py' isn't identical to minified_pyminifier.py then
something is broken.
"""

import sys, re, cStringIO, tokenize
from optparse import OptionParser

# Compile our regular expressions for speed
multiline_quoted_string = re.compile(r'(\'\'\'|\"\"\")')
not_quoted_string = re.compile(r'(\".*\'\'\'.*\"|\'.*\"\"\".*\')')
trailing_newlines = re.compile(r'\n\n')
shebang = re.compile('^#\!.*$')
encoding = re.compile(".*coding[:=]\s*([-\w.]+)")
multiline_indicator = re.compile('\\\\(\s*#.*)?\n')
# The above also removes trailing comments: "test = 'blah \ # comment here"

# These aren't used but they're a pretty good reference:
double_quoted_string = re.compile(r'((?<!\\)".*?(?<!\\)")')
single_quoted_string = re.compile(r"((?<!\\)'.*?(?<!\\)')")
single_line_single_quoted_string = re.compile(r"((?<!\\)'''.*?(?<!\\)''')")
single_line_double_quoted_string = re.compile(r"((?<!\\)'''.*?(?<!\\)''')")

def remove_comments_and_docstrings(source):
    """
    Returns 'source' minus comments and docstrings.

    **Note**: Uses Python's built-in tokenize module to great effect.

    Example:

    .. code-block:: python

        def noop(): # This is a comment
            '''
            Does nothing.
            '''
            pass # Don't do anything

    Will become:

    .. code-block:: python

        def noop():
            pass
    """
    io_obj = cStringIO.StringIO(source)
    out = ""
    prev_toktype = tokenize.INDENT
    last_lineno = -1
    last_col = 0
    for tok in tokenize.generate_tokens(io_obj.readline):
        token_type = tok[0]
        token_string = tok[1]
        start_line, start_col = tok[2]
        end_line, end_col = tok[3]
        ltext = tok[4]
        # The following two conditionals preserve indentation.
        # This is necessary because we're not using tokenize.untokenize()
        # (because it spits out code with copious amounts of oddly-placed
        # whitespace).
        if start_line > last_lineno:
            last_col = 0
        if start_col > last_col:
            out += (" " * (start_col - last_col))
        # Remove comments:
        if token_type == tokenize.COMMENT:
            pass
        # This series of conditionals removes docstrings:
        elif token_type == tokenize.STRING:
            if prev_toktype != tokenize.INDENT:
        # This is likely a docstring; double-check we're not inside an operator:
                if prev_toktype != tokenize.NEWLINE:
                    # Note regarding NEWLINE vs NL: The tokenize module
                    # differentiates between newlines that start a new statement
                    # and newlines inside of operators such as parens, brackes,
                    # and curly braces.  Newlines inside of operators are
                    # NEWLINE and newlines that start new code are NL.
                    # Catch whole-module docstrings:
                    if start_col > 0:
                        # Unlabelled indentation means we're inside an operator
                        out += token_string
                    # Note regarding the INDENT token: The tokenize module does
                    # not label indentation inside of an operator (parens,
                    # brackets, and curly braces) as actual indentation.
                    # For example:
                    # def foo():
                    #     "The spaces before this docstring are tokenize.INDENT"
                    #     test = [
                    #         "The spaces before this string do not get a token"
                    #     ]
        else:
            out += token_string
        prev_toktype = token_type
        last_col = end_col
        last_lineno = end_line
    return out

def reduce_operators(source):
    """
    Remove spaces between operators in 'source' and returns the result.

    Example:

    .. code-block:: python

        def foo(foo, bar, blah):
            test = "This is a %s" % foo

    Will become:

    .. code-block:: python

        def foo(foo,bar,blah):
            test="This is a %s"%foo
    """
    io_obj = cStringIO.StringIO(source)
    remove_columns = []
    out = ""
    out_line = ""
    prev_toktype = tokenize.INDENT
    prev_tok = None
    last_lineno = -1
    last_col = 0
    lshift = 1
    for tok in tokenize.generate_tokens(io_obj.readline):
        token_type = tok[0]
        token_string = tok[1]
        start_line, start_col = tok[2]
        end_line, end_col = tok[3]
        ltext = tok[4]
        if start_line > last_lineno:
            last_col = 0
        if start_col > last_col:
            out_line += (" " * (start_col - last_col))
        if token_type == tokenize.OP:
            # Operators that begin a line such as @ or open parens should be
            # left alone
            start_of_line_types = [ # These indicate we're starting a new line
                tokenize.NEWLINE, tokenize.DEDENT, tokenize.INDENT]
            if prev_toktype not in start_of_line_types:
                # This is just a regular operator; remove spaces
                remove_columns.append(start_col) # Before OP
                remove_columns.append(end_col+1) # After OP
        if token_string.endswith('\n'):
            out_line += token_string
            if remove_columns:
                for col in remove_columns:
                    col = col - lshift
                    try:
            # This was really handy for debugging (looks nice, worth saving):
                        #print out_line + (" " * col) + "^"
                        # The above points to the character we're looking at
                        if out_line[col] == " ": # Only if it is a space
                            out_line = out_line[:col] + out_line[col+1:]
                            lshift += 1 # To re-align future changes on this line
                    except IndexError: # Reached and end of line, no biggie
                        pass
            out += out_line
            remove_columns = []
            out_line = ""
            lshift = 1
        else:
            out_line += token_string
        prev_toktype = token_type
        prev_token = tok
        last_col = end_col
        last_lineno = end_line
    # This makes sure to capture the last line if it doesn't end in a newline:
    out += out_line
    # The tokenize module doesn't recognize @ sign before a decorator
    return out

# NOTE: This isn't used anymore...  Just here for reference in case someone
# searches the internet looking for a way to remove similarly-styled end-of-line
# comments from non-python code.  It also acts as an edge case of sorts with
# that raw triple quoted string inside the "quoted_string" assignment.
def remove_comment(single_line):
    """
    Removes the comment at the end of the line (if any) and returns the result.
    """
    quoted_string = re.compile(
        r'''((?<!\\)".*?(?<!\\)")|((?<!\\)'.*?(?<!\\)')'''
    )
    # This divides the line up into sections:
    #   Those inside single quotes and those that are not
    split_line = quoted_string.split(single_line)
    # Remove empty items:
    split_line = [a for a in split_line if a]
    out_line = ""
    for section in split_line:
        if section.startswith("'") or section.startswith('"'):
            # This is a quoted string; leave it alone
            out_line += section
        elif '#' in section: # A '#' not in quotes?  There's a comment here!
            # Get rid of everything after the # including the # itself:
            out_line += section.split('#')[0]
            break # No reason to bother the rest--it's all comments
        else:
            # This isn't a quoted string OR a comment; leave it as-is
            out_line += section
    return out_line.rstrip() # Strip trailing whitespace before returning

def join_multiline_pairs(text, pair="()"):
    """
    Finds and removes newlines in multiline matching pairs of characters in
    'text'.  For example, "(.*\n.*), {.*\n.*}, or [.*\n.*]".

    By default it joins parens () but it will join any two characters given via
    the 'pair' variable.

    **Note:** Doesn't remove extraneous whitespace that ends up between the pair.
    Use reduce_operators() for that.

    Example:

    .. code-block:: python

        test = (
            "This is inside a multi-line pair of parentheses"
        )

    Will become:

    .. code-block:: python

        test = (            "This is inside a multi-line pair of parentheses"        )
    """
    # Readability variables
    opener = pair[0]
    closer = pair[1]

    # Tracking variables
    inside_pair = False
    inside_quotes = False
    inside_double_quotes = False
    inside_single_quotes = False
    quoted_string = False
    openers = 0
    closers = 0
    linecount = 0

    # Regular expressions
    opener_regex = re.compile('\%s' % opener)
    closer_regex = re.compile('\%s' % closer)

    output = ""

    for line in text.split('\n'):
        escaped = False
        # First we rule out multi-line strings
        multline_match = multiline_quoted_string.search(line)
        not_quoted_string_match = not_quoted_string.search(line)
        if multline_match and not not_quoted_string_match and not quoted_string:
            if len(line.split('"""')) > 1 or len(line.split("'''")):
                # This is a single line that uses the triple quotes twice
                # Treat it as if it were just a regular line:
                output += line + '\n'
                quoted_string = False
            else:
                output += line + '\n'
                quoted_string = True
        elif quoted_string and multiline_quoted_string.search(line):
            output += line + '\n'
            quoted_string = False
        # Now let's focus on the lines containing our opener and/or closer:
        elif not quoted_string:
            if opener_regex.search(line) or closer_regex.search(line) or inside_pair:
                for character in line:
                    if character == opener:
                        if not escaped and not inside_quotes:
                            openers += 1
                            inside_pair = True
                            output += character
                        else:
                            escaped = False
                            output += character
                    elif character == closer:
                        if not escaped and not inside_quotes:
                            if openers and openers == (closers + 1):
                                closers = 0
                                openers = 0
                                inside_pair = False
                                output += character
                            else:
                                closers += 1
                                output += character
                        else:
                            escaped = False
                            output += character
                    elif character == '\\':
                        if escaped:
                            escaped = False
                            output += character
                        else:
                            escaped = True
                            output += character
                    elif character == '"' and escaped:
                        output += character
                        escaped = False
                    elif character == "'" and escaped:
                        output += character
                        escaped = False
                    elif character == '"' and inside_quotes:
                        if inside_single_quotes:
                            output += character
                        else:
                            inside_quotes = False
                            inside_double_quotes = False
                            output += character
                    elif character == "'" and inside_quotes:
                        if inside_double_quotes:
                            output += character
                        else:
                            inside_quotes = False
                            inside_single_quotes = False
                            output += character
                    elif character == '"' and not inside_quotes:
                        inside_quotes = True
                        inside_double_quotes = True
                        output += character
                    elif character == "'" and not inside_quotes:
                        inside_quotes = True
                        inside_single_quotes = True
                        output += character
                    elif character == ' ' and inside_pair and not inside_quotes:
                        if not output[-1] in [' ', opener]:
                            output += ' '
                    else:
                        if escaped:
                            escaped = False
                        output += character
                if inside_pair == False:
                    output += '\n'
            else:
                output += line + '\n'
        else:
            output += line + '\n'

    # Clean up
    output = trailing_newlines.sub('\n', output)

    return output

def dedent(source):
    """
    Minimizes indentation to save precious bytes

    Example:

    .. code-block:: python

        def foo(bar):
            test = "This is a test"

    Will become:

    .. code-block:: python

        def foo(bar):
         test = "This is a test"
    """
    io_obj = cStringIO.StringIO(source)
    out = ""
    last_lineno = -1
    last_col = 0
    prev_start_line = 0
    indentation = ""
    indentation_level = 0
    for i,tok in enumerate(tokenize.generate_tokens(io_obj.readline)):
        token_type = tok[0]
        token_string = tok[1]
        start_line, start_col = tok[2]
        end_line, end_col = tok[3]
        if start_line > last_lineno:
            last_col = 0
        if token_type == tokenize.INDENT:
            indentation_level += 1
            continue
        if token_type == tokenize.DEDENT:
            indentation_level -= 1
            continue
        indentation = " " * indentation_level
        if start_line > prev_start_line:
            out += indentation + token_string
        elif start_col > last_col:
            out += " " + token_string
        else:
            out += token_string
        prev_start_line = start_line
        last_col = end_col
        last_lineno = end_line
    return out

def fix_empty_methods(source):
    """
    Appends 'pass' to empty methods/functions (i.e. where there was nothing but
    a docstring before we removed it =).

    Example:

    .. code-block:: python

        # Note: This triple-single-quote inside a triple-double-quote is also a
        # pyminifier self-test
        def myfunc():
            '''This is just a placeholder function.'''

    Will become:

    .. code-block:: python

        def myfunc(): pass
    """
    def_indentation_level = 0
    output = ""
    just_matched = False
    previous_line = None
    method = re.compile(r'^\s*def\s*.*\(.*\):.*$')
    for line in source.split('\n'):
        if len(line.strip()) > 0: # Don't look at blank lines
            if just_matched == True:
                this_indentation_level = len(line.rstrip()) - len(line.strip())
                if def_indentation_level == this_indentation_level:
                    # This method is empty, insert a 'pass' statement
                    output += "%s pass\n%s\n" % (previous_line, line)
                else:
                    output += "%s\n%s\n" % (previous_line, line)
                just_matched = False
            elif method.match(line):
                def_indentation_level = len(line) - len(line.strip()) # A commment
                just_matched = True
                previous_line = line
            else:
                output += "%s\n" % line # Another self-test
        else:
            output += "\n"
    return output

def remove_blank_lines(source):
    """
    Removes blank lines from 'source' and returns the result.

    Example:

    .. code-block:: python

        test = "foo"

        test2 = "bar"

    Will become:

    .. code-block:: python

        test = "foo"
        test2 = "bar"
    """
    io_obj = cStringIO.StringIO(source)
    source = [a for a in io_obj.readlines() if a.strip()]
    return "".join(source)

def minify(source):
    """
    Remove all docstrings, comments, blank lines, and minimize code
    indentation from 'source' then prints the result.
    """
    preserved_shebang = None
    preserved_encoding = None

    # This is for things like shebangs that must be precisely preserved
    for line in source.split('\n')[0:2]:
        # Save the first comment line if it starts with a shebang
        # (e.g. '#!/usr/bin/env python') <--also a self test!
        if shebang.match(line): # Must be first line
            preserved_shebang = line
            continue
        # Save the encoding string (must be first or second line in file)
        if encoding.match(line):
            preserved_encoding = line

    # Remove multilines (e.g. lines that end with '\' followed by a newline)
    source = multiline_indicator.sub('', source)

    # Remove docstrings (Note: Must run before fix_empty_methods())
    source = remove_comments_and_docstrings(source)

    # Remove empty (i.e. single line) methods/functions
    source = fix_empty_methods(source)

    # Join multiline pairs of parens, brackets, and braces
    source = join_multiline_pairs(source)
    source = join_multiline_pairs(source, '[]')
    source = join_multiline_pairs(source, '{}')

    # Remove whitespace between operators:
    source = reduce_operators(source)

    # Minimize indentation
    source = dedent(source)

    # Re-add preseved items
    if preserved_encoding:
        source = preserved_encoding + "\n" + source
    if preserved_shebang:
        source = preserved_shebang + "\n" + source

    # Remove blank lines
    source = remove_blank_lines(source).rstrip('\n') # Stubborn last newline

    return source

def bz2_pack(source):
    "Returns 'source' as a bzip2-compressed, self-extracting python script."
    import bz2, base64
    out = ""
    compressed_source = bz2.compress(source)
    out += 'import bz2, base64\n'
    out += "exec bz2.decompress(base64.b64decode('"
    out += base64.b64encode((compressed_source))
    out += "'))\n"
    return out

def gz_pack(source):
    "Returns 'source' as a gzip-compressed, self-extracting python script."
    import zlib, base64
    out = ""
    compressed_source = zlib.compress(source)
    out += 'import zlib, base64\n'
    out += "exec zlib.decompress(base64.b64decode('"
    out += base64.b64encode((compressed_source))
    out += "'))\n"
    return out

# The test.+() functions below are for testing pyminifer...
def test_decorator(f):
    """Decorator that does nothing"""
    return f

def test_reduce_operators():
    """Test the case where an operator such as an open paren starts a line"""
    (a, b) = 1, 2 # The indentation level should be preserved
    pass

def test_empty_functions():
    """
    This is a test method.
    This should be replaced with 'def empty_method: pass'
    """

class test_class(object):
    "Testing indented decorators"

    @test_decorator
    def foo(self):
        pass

def test_function():
    """
    This function encapsulates the edge cases to prevent them from invading the
    global namespace.
    """
    foo = ("The # character in this string should " # This comment
           "not result in a syntax error") # ...and this one should go away
    test_multi_line_list = [
        'item1',
        'item2',
        'item3'
    ]
    test_multi_line_dict = {
        'item1': 1,
        'item2': 2,
        'item3': 3
    }
    # It may seem strange but the code below tests our docstring removal code.
    test_string_inside_operators = imaginary_function(
        "This string was indented but the tokenizer won't see it that way."
    ) # To understand how this could mess up docstring removal code see the
      # remove_comments_and_docstrings() function starting at this line:
      #     "elif token_type == tokenize.STRING:"
    # This tests remove_extraneous_spaces():
    this_line_has_leading_indentation    = '''<--That extraneous space should be
                                              removed''' # But not these spaces

def main():
    usage = '%prog [options] "<input file>"'
    parser = OptionParser(usage=usage, version=__version__)
    parser.disable_interspersed_args()
    parser.add_option(
        "-o", "--outfile",
        dest="outfile",
        default=None,
        help="Save output to the given file.",
        metavar="<file path>"
    )
    parser.add_option(
        "--bzip2",
        action="store_true",
        dest="bzip2",
        default=False,
        help="bzip2-compress the result into a self-executing python script."
    )
    parser.add_option(
        "--gzip",
        action="store_true",
        dest="gzip",
        default=False,
        help="gzip-compress the result into a self-executing python script."
    )
    options, args = parser.parse_args()
    try:
        source = open(args[0]).read()
    except Exception, e:
        print e
        parser.print_help()
        sys.exit(2)
    # Minify our input script
    result = minify(source)
    # Compress it if we were asked to do so
    if options.bzip2:
        result = bz2_pack(result)
    elif options.gzip:
        result = gz_pack(result)
    # Either save the result to the output file or print it to stdout
    if options.outfile:
        f = open(options.outfile, 'w')
        f.write(result)
        f.close()
    else:
        print result

if __name__ == "__main__":
    main()