YAPF is a Python formatter based on clang-format (developed by Daniel Jasper). In essence, the algorithm takes the code and calculates the best formatting that conforms to the configured style. It takes away a lot of the drudgery of maintaining your code.

The ultimate goal is that the code YAPF produces is as good as the code that a programmer would write if they were following the style guide.

Note YAPF is not an official Google product (experimental or otherwise), it is just code that happens to be owned by Google.

To install YAPF from PyPI:

$ pip install yapf

YAPF is still considered in "beta" stage, and the released version may change often; therefore, the best way to keep up-to-date with the latest development is to clone this repository or install directly from github:

$ pip install git+https://github.com/google/yapf.git

Note that if you intend to use YAPF as a command-line tool rather than as a library, installation is not necessary. YAPF supports being run as a directory by the Python interpreter. If you cloned/unzipped YAPF into DIR, it's possible to run:

$ PYTHONPATH=DIR python DIR/yapf [options] ...

YAPF is supported by multiple editors via community extensions or plugins. See Editor Support for more info.

YAPF supports Python 3.7+.

usage: yapf [-h] [-v] [-d | -i | -q] [-r | -l START-END] [-e PATTERN]
            [--style STYLE] [--style-help] [--no-local-style] [-p] [-m] [-vv]
            [files ...]

Formatter for Python code.

positional arguments:
  files                 reads from stdin when no files are specified.

optional arguments:
  -h, --help            show this help message and exit
  -v, --version         show program's version number and exit
  -d, --diff            print the diff for the fixed source
  -i, --in-place        make changes to files in place
  -q, --quiet           output nothing and set return value
  -r, --recursive       run recursively over directories
  -l START-END, --lines START-END
                        range of lines to reformat, one-based
  -e PATTERN, --exclude PATTERN
                        patterns for files to exclude from formatting
  --style STYLE         specify formatting style: either a style name (for
                        example "pep8" or "google"), or the name of a file
                        with style settings. The default is pep8 unless a
                        .style.yapf or setup.cfg or pyproject.toml file
                        located in the same directory as the source or one of
                        its parent directories (for stdin, the current
                        directory is used).
  --style-help          show style settings and exit; this output can be saved
                        to .style.yapf to make your settings permanent
  --no-local-style      don't search for local style definition
  -p, --parallel        run YAPF in parallel when formatting multiple files.
  -m, --print-modified  print out file names of modified files
  -vv, --verbose        print out file names while processing

Normally YAPF returns zero on successful program termination and non-zero otherwise.

If --diff is supplied, YAPF returns zero when no changes were necessary, non-zero otherwise (including program error). You can use this in a CI workflow to test that code has been YAPF-formatted.

In addition to exclude patterns provided on commandline, YAPF looks for additional patterns specified in a file named .yapfignore or pyproject.toml located in the working directory from which YAPF is invoked.

.yapfignore's syntax is similar to UNIX's filename pattern matching:

*       matches everything
?       matches any single character
[seq]   matches any character in seq
[!seq]  matches any character not in seq

Note that no entry should begin with ./.

If you use pyproject.toml, exclude patterns are specified by ignore_patterns key in [tool.yapfignore] section. For example:

[tool.yapfignore]
ignore_patterns = [
  "temp/**/*.py",
  "temp2/*.py"
]

The formatting style used by YAPF is configurable and there are many "knobs" that can be used to tune how YAPF does formatting. See the style.py module for the full list.

To control the style, run YAPF with the --style argument. It accepts one of the predefined styles (e.g., pep8 or google), a path to a configuration file that specifies the desired style, or a dictionary of key/value pairs.

The config file is a simple listing of (case-insensitive) key = value pairs with a [style] heading. For example:

[style]
based_on_style = pep8
spaces_before_comment = 4
split_before_logical_operator = true

The based_on_style setting determines which of the predefined styles this custom style is based on (think of it like subclassing). Four styles are predefined:

• pep8 (default)
• google (based off of the Google Python Style Guide)
• yapf (for use with Google open source projects)
• facebook

See _STYLE_NAME_TO_FACTORY in style.py for details.

It's also possible to do the same on the command line with a dictionary. For example:

--style='{based_on_style: pep8, indent_width: 2}'

This will take the pep8 base style and modify it to have two space indentations.

YAPF will search for the formatting style in the following manner:

1. Specified on the command line
2. In the [style] section of a .style.yapf file in either the current directory or one of its parent directories.
3. In the [yapf] section of a setup.cfg file in either the current directory or one of its parent directories.
4. In the [tool.yapf] section of a pyproject.toml file in either the current directory or one of its parent directories.
5. In the [style] section of a ~/.config/yapf/style file in your home directory.

If none of those files are found, the default style PEP8 is used.

An example of the type of formatting that YAPF can do, it will take this ugly code:

x = {  'a':37,'b':42,

'c':927}

y = 'hello ''world'
z = 'hello '+'world'
a = 'hello {}'.format('world')
class foo  (     object  ):
  def f    (self   ):
    return       37*-+2
  def g(self, x,y=42):
      return y
def f  (   a ) :
  return      37+-+a[42-x :  y**3]

and reformat it into:

x = {'a': 37, 'b': 42, 'c': 927}

y = 'hello ' 'world'
z = 'hello ' + 'world'
a = 'hello {}'.format('world')


class foo(object):
    def f(self):
        return 37 * -+2

    def g(self, x, y=42):
        return y


def f(a):
    return 37 + -+a[42 - x:y**3]

The two main APIs for calling YAPF are FormatCode and FormatFile, these share several arguments which are described below:

>>> from yapf.yapflib.yapf_api import FormatCode  # reformat a string of code

>>> formatted_code, changed = FormatCode("f ( a = 1, b = 2 )")
>>> formatted_code
'f(a=1, b=2)\n'
>>> changed
True

A style_config argument: Either a style name or a path to a file that contains formatting style settings. If None is specified, use the default style as set in style.DEFAULT_STYLE_FACTORY.

>>> FormatCode("def g():\n  return True", style_config='pep8')[0]
'def g():\n    return True\n'

A lines argument: A list of tuples of lines (ints), [start, end], that we want to format. The lines are 1-based indexed. It can be used by third-party code (e.g., IDEs) when reformatting a snippet of code rather than a whole file.

>>> FormatCode("def g( ):\n    a=1\n    b = 2\n    return a==b", lines=[(1, 1), (2, 3)])[0]
'def g():\n    a = 1\n    b = 2\n    return a==b\n'

A print_diff (bool): Instead of returning the reformatted source, return a diff that turns the formatted source into reformatted source.

>>> print(FormatCode("a==b", filename="foo.py", print_diff=True)[0])
--- foo.py (original)
+++ foo.py (reformatted)
@@ -1 +1 @@
-a==b
+a == b

Note: the filename argument for FormatCode is what is inserted into the diff, the default is <unknown>.

FormatFile returns reformatted code from the passed file along with its encoding:

>>> from yapf.yapflib.yapf_api import FormatFile  # reformat a file

>>> print(open("foo.py").read())  # contents of file
a==b

>>> reformatted_code, encoding, changed = FormatFile("foo.py")
>>> formatted_code
'a == b\n'
>>> encoding
'utf-8'
>>> changed
True

The in_place argument saves the reformatted code back to the file:

>>> FormatFile("foo.py", in_place=True)[:2]
(None, 'utf-8')

>>> print(open("foo.py").read())  # contents of file (now fixed)
a == b

Options:

usage: yapf-diff [-h] [-i] [-p NUM] [--regex PATTERN] [--iregex PATTERN][-v]
                 [--style STYLE] [--binary BINARY]

This script reads input from a unified diff and reformats all the changed
lines. This is useful to reformat all the lines touched by a specific patch.
Example usage for git/svn users:

  git diff -U0 --no-color --relative HEAD^ | yapf-diff -i
  svn diff --diff-cmd=diff -x-U0 | yapf-diff -p0 -i

It should be noted that the filename contained in the diff is used
unmodified to determine the source file to update. Users calling this script
directly should be careful to ensure that the path in the diff is correct
relative to the current working directory.

optional arguments:
  -h, --help            show this help message and exit
  -i, --in-place        apply edits to files instead of displaying a diff
  -p NUM, --prefix NUM  strip the smallest prefix containing P slashes
  --regex PATTERN       custom pattern selecting file paths to reformat
                        (case sensitive, overrides -iregex)
  --iregex PATTERN      custom pattern selecting file paths to reformat
                        (case insensitive, overridden by -regex)
  -v, --verbose         be more verbose, ineffective without -i
  --style STYLE         specify formatting style: either a style name (for
                        example "pep8" or "google"), or the name of a file
                        with style settings. The default is pep8 unless a
                        .style.yapf or setup.cfg or pyproject.toml file
                        located in the same directory as the source or one of
                        its parent directories (for stdin, the current
                        directory is used).
  --binary BINARY       location of binary to use for YAPF

• Python 3.12 – PEP 695 – Type Parameter Syntax – YAPF #1170
• Python 3.12 – PEP 701 – Syntactic formalization of f-strings – YAPF #1136

Align closing bracket with visual indentation.

Allow lambdas to be formatted on more than one line.

Allow dictionary keys to exist on multiple lines. For example:

    x = {
        ('this is the first element of a tuple',
         'this is the second element of a tuple'):
             value,
    }

Allow splitting before a default / named assignment in an argument list.

Allow splits before the dictionary value.

Let spacing indicate operator precedence. For example:

    a = 1 * 2 + 3 / 4
    b = 1 / 2 - 3 * 4
    c = (1 + 2) * (3 - 4)
    d = (1 - 2) / (3 + 4)
    e = 1 * 2 - 3
    f = 1 + 2 + 3 + 4

will be formatted as follows to indicate precedence:

    a = 1*2 + 3/4
    b = 1/2 - 3*4
    c = (1+2) * (3-4)
    d = (1-2) / (3+4)
    e = 1*2 - 3
    f = 1 + 2 + 3 + 4

Sets the number of desired blank lines surrounding top-level function and class definitions. For example:

    class Foo:
        pass
                       # <------ having two blank lines here
                       # <------ is the default setting
    class Bar:
        pass

Insert a blank line before a class-level docstring.

Insert a blank line before a module docstring.

Insert a blank line before a def or class immediately nested within another def or class. For example:

    class Foo:
                       # <------ this blank line
        def method():
            pass

Sets the number of desired blank lines between top-level imports and variable definitions. Useful for compatibility with tools like isort.

Do not split consecutive brackets. Only relevant when DEDENT_CLOSING_BRACKETS or INDENT_CLOSING_BRACKETS is set. For example:

    call_func_that_takes_a_dict(
        {
            'key1': 'value1',
            'key2': 'value2',
        }
    )

would reformat to:

    call_func_that_takes_a_dict({
        'key1': 'value1',
        'key2': 'value2',
    })

The column limit (or max line-length)

The style for continuation alignment. Possible values are:

• SPACE: Use spaces for continuation alignment. This is default behavior.
• FIXED: Use fixed number (CONTINUATION_INDENT_WIDTH) of columns (i.e. CONTINUATION_INDENT_WIDTH/INDENT_WIDTH tabs or CONTINUATION_INDENT_WIDTH spaces) for continuation alignment.
• VALIGN-RIGHT: Vertically align continuation lines to multiple of INDENT_WIDTH columns. Slightly right (one tab or a few spaces) if cannot vertically align continuation lines with indent characters.

Indent width used for line continuations.

Put closing brackets on a separate line, dedented, if the bracketed expression can't fit in a single line. Applies to all kinds of brackets, including function definitions and calls. For example:

    config = {
        'key1': 'value1',
        'key2': 'value2',
    }  # <--- this bracket is dedented and on a separate line

    time_series = self.remote_client.query_entity_counters(
        entity='dev3246.region1',
        key='dns.query_latency_tcp',
        transform=Transformation.AVERAGE(window=timedelta(seconds=60)),
        start_ts=now()-timedelta(days=3),
        end_ts=now(),
    )  # <--- this bracket is dedented and on a separate line

Disable the heuristic which places each list element on a separate line if the list is comma-terminated.

Note: The behavior of this flag changed in v0.40.3. Before, if this flag was true, we would split lists that contained a trailing comma or a comment. Now, we have a separate flag, DISABLE_SPLIT_LIST_WITH_COMMENT, that controls splitting when a list contains a comment. To get the old behavior, set both flags to true. More information in CHANGELOG.md.

Don't put every element on a new line within a list that contains interstitial comments.

Without this flag (default):

[
  a,
  b,  #
  c
]

With this flag:

[
  a, b,  #
  c
]

This mirrors the behavior of clang-format and is useful for forming "logical groups" of elements in a list. It also works in function declarations.

Place each dictionary entry onto its own line.

Respect EACH_DICT_ENTRY_ON_SEPARATE_LINE even if the line is shorter than COLUMN_LIMIT.

The regex for an internationalization comment. The presence of this comment stops reformatting of that line, because the comments are required to be next to the string they translate.

The internationalization function call names. The presence of this function stops reformatting on that line, because the string it has cannot be moved away from the i18n comment.

Set to True to prefer indented blank lines rather than empty

Put closing brackets on a separate line, indented, if the bracketed expression can't fit in a single line. Applies to all kinds of brackets, including function definitions and calls. For example:

    config = {
        'key1': 'value1',
        'key2': 'value2',
        }  # <--- this bracket is indented and on a separate line

    time_series = self.remote_client.query_entity_counters(
        entity='dev3246.region1',
        key='dns.query_latency_tcp',
        transform=Transformation.AVERAGE(window=timedelta(seconds=60)),
        start_ts=now()-timedelta(days=3),
        end_ts=now(),
        )  # <--- this bracket is indented and on a separate line

Indent the dictionary value if it cannot fit on the same line as the dictionary key. For example:

    config = {
        'key1':
            'value1',
        'key2': value1 +
                value2,
    }

The number of columns to use for indentation.

Join short lines into one line. E.g., single line if statements.

Do not include spaces around selected binary operators. For example:

    1 + 2 * 3 - 4 / 5

will be formatted as follows when configured with *, /:

    1 + 2*3 - 4/5

Insert a space between the ending comma and closing bracket of a list, etc.

Use spaces inside brackets, braces, and parentheses.  For example:

        method_call( 1 )
        my_dict[ 3 ][ 1 ][ get_index( *args, **kwargs ) ]
        my_set = { 1, 2, 3 }

Set to True to prefer spaces around the assignment operator for default or keyword arguments.

Adds a space after the opening '{' and before the ending '}' dict delimiters.

        {1: 2}

will be formatted as:

        { 1: 2 }

Adds a space after the opening '[' and before the ending ']' list delimiters.

    [1, 2]

will be formatted as:

    [ 1, 2 ]

Set to True to prefer using spaces around **.

Use spaces around the subscript / slice operator. For example:

    my_list[1 : 10 : 2]

Adds a space after the opening '(' and before the ending ')' tuple delimiters.

    (1, 2, 3)

will be formatted as:

    ( 1, 2, 3 )

The number of spaces required before a trailing comment. This can be a single value (representing the number of spaces before each trailing comment) or list of values (representing alignment column values; trailing comments within a block will be aligned to the first column value that is greater than the maximum line length within the block).

Note: Lists of values may need to be quoted in some contexts (eg. shells or editor config files).

For example, with spaces_before_comment=5:

    1 + 1 # Adding values

will be formatted as:

    1 + 1     # Adding values <-- 5 spaces between the end of the statement and comment

with spaces_before_comment="15, 20":

    1 + 1 # Adding values
    two + two # More adding

    longer_statement # This is a longer statement
    short # This is a shorter statement

    a_very_long_statement_that_extends_beyond_the_final_column # Comment
    short # This is a shorter statement

will be formatted as:

    1 + 1          # Adding values <-- end of line comments in block aligned to col 15
    two + two      # More adding

    longer_statement    # This is a longer statement <-- end of line comments in block aligned to col 20
    short               # This is a shorter statement

    a_very_long_statement_that_extends_beyond_the_final_column  # Comment <-- the end of line comments are aligned based on the line length
    short                                                       # This is a shorter statement

If a comma separated list (dict, list, tuple, or function def) is on a line that is too long, split such that each element is on a separate line.

Variation on SPLIT_ALL_COMMA_SEPARATED_VALUES in which, if a subexpression with a comma fits in its starting line, then the subexpression is not split. This avoids splits like the one for b in this code:

    abcdef(
        aReallyLongThing: int,
        b: [Int,
            Int])

with the new knob this is split as:

    abcdef(
        aReallyLongThing: int,
        b: [Int, Int])

Split before arguments if the argument list is terminated by a comma.

Set to True to prefer splitting before +, -, *, /, //, or @ rather than after.

Set to True to prefer splitting before &, | or ^ rather than after.

Split before the closing bracket if a list or dict literal doesn't fit on a single line.

Split before a dictionary or set generator (comp_for). For example, note the split before the for:

    foo = {
        variable: 'Hello world, have a nice day!'
        for variable in bar if variable != 42
    }

Split before the . if we need to split a longer expression:

    foo = ('This is a really long string: {}, {}, {}, {}'.format(a, b, c, d))

would reformat to something like:

    foo = ('This is a really long string: {}, {}, {}, {}'
           .format(a, b, c, d))

Split after the opening paren which surrounds an expression if it doesn't fit on a single line.

If an argument / parameter list is going to be split, then split before the first argument.

Set to True to prefer splitting before and or or rather than after.

Split named assignments onto individual lines.

For list comprehensions and generator expressions with multiple clauses (e.g multiple for calls, if filter expressions) and which need to be reflowed, split each clause onto its own line. For example:

    result = [
        a_var + b_var for a_var in xrange(1000) for b_var in xrange(1000)
        if a_var % b_var]

would reformat to something like:

    result = [
        a_var + b_var
        for a_var in xrange(1000)
        for b_var in xrange(1000)
        if a_var % b_var]

The penalty for splitting right after the opening bracket.

The penalty for splitting the line after a unary operator.

The penalty of splitting the line around the +, -, *, /, //, %, and @ operators.

The penalty for splitting right before an if expression.

The penalty of splitting the line around the &, |, and ^ operators.

The penalty for splitting a list comprehension or generator expression.

The penalty for characters over the column limit.

The penalty incurred by adding a line split to the logical line. The more line splits added the higher the penalty.

The penalty of splitting a list of import as names. For example:

    from a_very_long_or_indented_module_name_yada_yad import (long_argument_1,
                                                              long_argument_2,
                                                              long_argument_3)

would reformat to something like:

    from a_very_long_or_indented_module_name_yada_yad import (
        long_argument_1, long_argument_2, long_argument_3)

The penalty of splitting the line around the and and or operators.

Use the Tab character for indentation.

YAPF tries very hard to get the formatting correct. But for some code, it won't be as good as hand-formatting. In particular, large data literals may become horribly disfigured under YAPF.

The reasons for this are manyfold. In short, YAPF is simply a tool to help with development. It will format things to coincide with the style guide, but that may not equate with readability.

What can be done to alleviate this situation is to indicate regions YAPF should ignore when reformatting something:

# yapf: disable
FOO = {
    # ... some very large, complex data literal.
}

BAR = [
    # ... another large data literal.
]
# yapf: enable

You can also disable formatting for a single literal like this:

BAZ = {
    (1, 2, 3, 4),
    (5, 6, 7, 8),
    (9, 10, 11, 12),
}  # yapf: disable

To preserve the nice dedented closing brackets, use the dedent_closing_brackets in your style. Note that in this case all brackets, including function definitions and calls, are going to use that style. This provides consistency across the formatted codebase.

We wanted to use clang-format's reformatting algorithm. It's very powerful and designed to come up with the best formatting possible. Existing tools were created with different goals in mind, and would require extensive modifications to convert to using clang-format's algorithm.

Please do! YAPF was designed to be used as a library as well as a command line tool. This means that a tool or IDE plugin is free to use YAPF.

YAPF tries very hard to be fully PEP 8 compliant. However, it is paramount to not risk altering the semantics of your code. Thus, YAPF tries to be as safe as possible and does not change the token stream (e.g., by adding parentheses). All these cases however, can be easily fixed manually. For instance,

from my_package import my_function_1, my_function_2, my_function_3, my_function_4, my_function_5

FOO = my_variable_1 + my_variable_2 + my_variable_3 + my_variable_4 + my_variable_5 + my_variable_6 + my_variable_7 + my_variable_8

won't be split, but you can easily get it right by just adding parentheses:

from my_package import (my_function_1, my_function_2, my_function_3,
                        my_function_4, my_function_5)

FOO = (my_variable_1 + my_variable_2 + my_variable_3 + my_variable_4 +
       my_variable_5 + my_variable_6 + my_variable_7 + my_variable_8)

The main data structure in YAPF is the LogicalLine object. It holds a list of FormatToken\s, that we would want to place on a single line if there were no column limit. An exception being a comment in the middle of an expression statement will force the line to be formatted on more than one line. The formatter works on one LogicalLine object at a time.

An LogicalLine typically won't affect the formatting of lines before or after it. There is a part of the algorithm that may join two or more LogicalLine\s into one line. For instance, an if-then statement with a short body can be placed on a single line:

if a == 42: continue

YAPF's formatting algorithm creates a weighted tree that acts as the solution space for the algorithm. Each node in the tree represents the result of a formatting decision --- i.e., whether to split or not to split before a token. Each formatting decision has a cost associated with it. Therefore, the cost is realized on the edge between two nodes. (In reality, the weighted tree doesn't have separate edge objects, so the cost resides on the nodes themselves.)

For example, take the following Python code snippet. For the sake of this example, assume that line (1) violates the column limit restriction and needs to be reformatted.

def xxxxxxxxxxx(aaaaaaaaaaaa, bbbbbbbbb, cccccccc, dddddddd, eeeeee):  # 1
    pass                                                               # 2

For line (1), the algorithm will build a tree where each node (a FormattingDecisionState object) is the state of the line at that token given the decision to split before the token or not. Note: the FormatDecisionState objects are copied by value so each node in the graph is unique and a change in one doesn't affect other nodes.

Heuristics are used to determine the costs of splitting or not splitting. Because a node holds the state of the tree up to a token's insertion, it can easily determine if a splitting decision will violate one of the style requirements. For instance, the heuristic is able to apply an extra penalty to the edge when not splitting between the previous token and the one being added.

There are some instances where we will never want to split the line, because doing so will always be detrimental (i.e., it will require a backslash-newline, which is very rarely desirable). For line (1), we will never want to split the first three tokens: def, xxxxxxxxxxx, and (. Nor will we want to split between the ) and the : at the end. These regions are said to be "unbreakable." This is reflected in the tree by there not being a "split" decision (left hand branch) within the unbreakable region.

Now that we have the tree, we determine what the "best" formatting is by finding the path through the tree with the lowest cost.

And that's it!