New in version 0.7.

You can filter token streams coming from lexers to improve or annotate the output. For example, you can highlight special words in comments, convert keywords to upper or lowercase to enforce a style guide etc.

To apply a filter, you can use the add_filter() method of a lexer:

>>> from pygments.lexers import PythonLexer
>>> l = PythonLexer()
>>> # add a filter given by a string and options
>>> l.add_filter('codetagify', case='lower')
>>> l.filters
[<pygments.filters.CodeTagFilter object at 0xb785decc>]
>>> from pygments.filters import KeywordCaseFilter
>>> # or give an instance
>>> l.add_filter(KeywordCaseFilter(case='lower'))

The add_filter() method takes keyword arguments which are forwarded to the constructor of the filter.

To get a list of all registered filters by name, you can use the get_all_filters() function from the pygments.filters module that returns an iterable for all known filters.

If you want to write your own filter, have a look at Write your own filter.

Builtin Filters

class RaiseOnErrorTokenFilter

Raise an exception when the lexer generates an error token.

Options accepted:

excclass : Exception class
The exception class to raise. The default is pygments.filters.ErrorToken.

New in version 0.8.

class VisibleWhitespaceFilter

Convert tabs, newlines and/or spaces to visible characters.

Options accepted:

spaces : string or bool
If this is a one-character string, spaces will be replaces by this string. If it is another true value, spaces will be replaced by · (unicode MIDDLE DOT). If it is a false value, spaces will not be replaced. The default is False.
tabs : string or bool
The same as for spaces, but the default replacement character is » (unicode RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK). The default value is False. Note: this will not work if the tabsize option for the lexer is nonzero, as tabs will already have been expanded then.
tabsize : int
If tabs are to be replaced by this filter (see the tabs option), this is the total number of characters that a tab should be expanded to. The default is 8.
newlines : string or bool
The same as for spaces, but the default replacement character is (unicode PILCROW SIGN). The default value is False.
wstokentype : bool
If true, give whitespace the special Whitespace token type. This allows styling the visible whitespace differently (e.g. greyed out), but it can disrupt background colors. The default is True.

New in version 0.8.

class TokenMergeFilter

Merges consecutive tokens with the same token type in the output stream of a lexer.

New in version 1.2.

class NameHighlightFilter

Highlight a normal Name (and Name.*) token with a different token type.


filter = NameHighlightFilter(
    names=['foo', 'bar', 'baz'],

This would highlight the names “foo”, “bar” and “baz” as functions. Name.Function is the default token type.

Options accepted:

names : list of strings
A list of names that should be given the different token type. There is no default.
tokentype : TokenType or string
A token type or a string containing a token type name that is used for highlighting the strings in names. The default is Name.Function.
class GobbleFilter

Gobbles source code lines (eats initial characters).

This filter drops the first n characters off every line of code. This may be useful when the source code fed to the lexer is indented by a fixed amount of space that isn’t desired in the output.

Options accepted:

n : int
The number of characters to gobble.

New in version 1.2.

class CodeTagFilter

Highlight special code tags in comments and docstrings.

Options accepted:

codetags : list of strings
A list of strings that are flagged as code tags. The default is to highlight XXX, TODO, BUG and NOTE.
class KeywordCaseFilter

Convert keywords to lowercase or uppercase or capitalize them, which means first letter uppercase, rest lowercase.

This can be useful e.g. if you highlight Pascal code and want to adapt the code to your styleguide.

Options accepted:

case : string
The casing to convert keywords to. Must be one of 'lower', 'upper' or 'capitalize'. The default is 'lower'.