Front matter offers a useful way to add arbitrary, structured metadata to text documents, regardless of type.
This is a small package to load and parse files (or just text) with YAML (or JSON, TOML or other) front matter.
# with pip
pip install python-frontmatter
# or uv
uv add python-frontmatter
# local development, with uv
uv sync>>> import frontmatterLoad a post from a filename:
>>> post = frontmatter.load('tests/yaml/hello-world.txt')Or a file (or file-like object):
>>> with open('tests/yaml/hello-world.txt') as f:
... post = frontmatter.load(f)Or load from text:
>>> with open('tests/yaml/hello-world.txt') as f:
... post = frontmatter.loads(f.read())If the file has a Byte-Order Mark (BOM), strip it off first. An easy way to do this is by using the utf-8-sig encoding:
>>> with open('tests/yaml/hello-world.txt', encoding="utf-8-sig") as f:
... post = frontmatter.load(f)Access content:
>>> print(post.content)
Well, hello there, world.
# this works, too
>>> print(post)
Well, hello there, world.Use metadata (metadata gets proxied as post keys):
>>> print(post['title'])
Hello, world!Metadata is a dictionary, with some handy proxies:
>>> sorted(post.keys())
['layout', 'title']
>>> from pprint import pprint
>>> post['excerpt'] = 'tl;dr'
>>> pprint(post.metadata)
{'excerpt': 'tl;dr', 'layout': 'post', 'title': 'Hello, world!'}If you don't need the whole post object, just parse:
>>> with open('tests/yaml/hello-world.txt') as f:
... metadata, content = frontmatter.parse(f.read())
>>> print(metadata['title'])
Hello, world!Write back to plain text, too:
>>> print(frontmatter.dumps(post)) # doctest: +NORMALIZE_WHITESPACE
---
excerpt: tl;dr
layout: post
title: Hello, world!
---
Well, hello there, world.Or write to a file (or file-like object):
>>> from io import StringIO
>>> f = StringIO()
>>> frontmatter.dump(post, f)
>>> print(f.getvalue()) # doctest: +NORMALIZE_WHITESPACE
---
excerpt: tl;dr
layout: post
title: Hello, world!
---
Well, hello there, world.For more examples, see files in the tests/ directory. Each sample file has a corresponding .result.json file showing the expected parsed output. See also the examples/ directory, which covers more ways to customize input and output.