Quickstart¶

Parse, edit, dump¶

import tomlrt

with open("pyproject.toml", "rb") as f:
    doc = tomlrt.load(f)

doc["project"]["version"] = "0.2.0"
doc["project"]["dependencies"].append("requests>=2")

with open("pyproject.toml", "wb") as f:
    tomlrt.dump(doc, f)

For in-memory round-trips, use loads / dumps:

doc = tomlrt.loads(text)
text_again = tomlrt.dumps(doc)
assert text == text_again

tomlrt.loads and tomlrt.parse are equivalent. tomlrt.dumps(doc) and doc.render() are equivalent.

Reading values¶

A Document behaves like a dict; nested tables are Table (also a dict subclass), inline arrays are Array (a list subclass), and arrays-of-tables are AoT (a list of Table). Plain reads with doc["key"] work as you'd expect — see Typed access when you want typechecker-friendly traversal.

Writing values¶

For the difference between snapshot and live containers, see Editing documents.