Skip to main content

Parse/serialize any markup format

Project description

Build Status Code Health Coverage

Parse or serialize any markup. Currently supports ini, json, json5, toml, xml and yaml. Report bugs and new functionality requests at


>>> import anymarkup
>>> anymarkup.parse('foo: bar')
{'foo': 'bar'}
>>> anymarkup.parse_file('foo/bar.ini')
{'section': {'subsection': {'opt2': 'bar'}, 'opt1': 'foo'}}

$ cat foo/bar.ini


>>> import anymarkup
>>> anymarkup.serialize({'foo': 'bar'}, 'json')
b'{\n  "foo": "bar"\n}'
>>> anymarkup.serialize_file({'foo': 'bar'}, 'foo/bar.json')

$ cat foo/bar.json
  "foo": "bar"

anymarkup is licensed under BSD license. You can download official releases from or install them via pip install anymarkup.

anymarkup works with Python 2.7 and >= 3.3.

Automatic Markup Language Recognition

When using anymarkup.parse(input), anymarkup will try to guess markup language of input. This usually works fine except:

  • ini vs toml: These two look almost the same and in fact have common subset (which, however, yields different parsing result). Because of this, anything with an ini-like look will be parsed with ini parser. If you want an input string to be parsed as toml, you have to explicitly specify that using format=toml (see below for examples).

  • json vs json5: json5 is superset of json, but not very widely used. Because of practicality of json usage, everything that looks like json is parsed as json. If you want input string to be parsed as json5, you have to explicitly specify that using format=json5.

When using anymarkup.parse_file(path), anymarkup will try to guess format based on file extension and then fallback to guessing as explained above. This means that if the file has .toml pr .json5 extension, you don’t have to provide format=<format> explicitly.

Notes on Parsing Basic Types

When parsing, anymarkup recognizes basic types - NoneType, int, float and bool (and long on Python 2) and converts all values to these types. If you want to get everything as strings, just use force_types=False with parse or parse_file. Finally, you can also use force_types=None to get whatever the parsing backend returned:

>>> anymarkup.parse('a: 1')
{'a': 1}
>>> anymarkup.parse('a: 1', force_types=False)
{'a': '1'}
>>> anymarkup.parse('a: 1', force_types=None)
{'a': 1}


To install the CLI, run the following command:

pip install anymarkup

Example of conversion from JSON to XML:

anymarkup convert –from-format json –to-format xml <somefile.json

For full help on the CLI run the following commands:

anymarkup –help anymarkup convert –help


anymarkup uses:

Notes on OrderedDict

Parsing certain types of markup can yield Python’s OrderedDict type - namely XML documents and YAML !!omap (see anymarkup handles this without a problem, but note that if you serialize these as JSON or INI and then parse again, you’ll lose the ordering information (meaning you’ll get just dict back).

This is because JSON and INI parsers (to my knowledge) don’t consider ordering key-value structures important and there’s no direct means in these markup languages to express ordering key-value structures.

Notes on Dependencies

Read this section if you want anymarkup functionality only for subset of supported markup languages without the need to install all parsers.

Since version 0.5.0, anymarkup is just a wrapper library around anymarkup-core ( and doesn’t actually contain any code, except of imports from anymarkup-core.

anymarkup-core goal is to not explicitly depend on any of the parsers, so people can install it with only a specified subset of dependencies. For example, you can install anymarkup-core only with PyYAML, if you know you’ll only be parsing YAML.

If you install anymarkup, you will always get a full set of dependencies and you will be able to parse any markup language that’s supported.

The CLI requires click as indicated in the requirements.txt file.


Parsing examples:

ini = """
foo = bar"""

json = """
{"a": {
    "foo": "bar"

xml = """<?xml version="1.0" encoding="UTF-8"?>

yaml = """
  foo: bar

# these will all yield the same value (except that xml parsing will yield OrderedDict)

# explicitly specify a type of format to expect and/or encoding (utf-8 is default)
anymarkup.parse('foo: bar', format='yaml', encoding='ascii')

# by default, anymarkup recognizes basic types (None, booleans, ints and floats)
#   if you want to get everything as strings, just use force_types=False

# will yield {'a': 1, 'b': True, 'c': None}
anymarkup.parse('a: 1\nb: True\nc: None')
# will yield {'a': '1', 'b': 'True', 'c': 'None'}
anymarkup.parse('a: 1\nb: True\nc: None', force_types=False)

# or parse a file

# if a file doesn't have a format extension, pass it explicitly
anymarkup.parse_file('foo', format='json')

# you can also pass encoding explicitly (utf-8 is default)
anymarkup.parse_file('bar', format='xml', encoding='ascii')

Serializing examples:

struct = {'a': ['b', 'c']}

for fmt in ['ini', 'json', 'xml', 'yaml']:
    # any of the above formats can be used for serializing
    anymarkup.serialize(struct, fmt)

# explicitly specify encoding (utf-8 is default)
anymarkup.serialize(struct, 'json', encoding='utf-8')

# or serialize directly to a file
anymarkup.serialize_file(struct, 'foo/bar.ini')

# if a file doesn't have a format extension, pass it explicitly
anymarkup.serialize_file(struct, 'foo/bar', format='json')

# you can also pass encoding explicitly (utf-8 is default)
anymarkup.serialize_file(struct, 'foo/bar', format='json', encoding='ascii')

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

anymarkup-0.8.1.tar.gz (7.7 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page