tomledit 0.7.0

Format-preserving TOML parser

tomledit

A format-preserving TOML editor for Python, powered by Rust's toml_edit.

Parse a TOML document, modify it, and write it back - comments, whitespace, and ordering are preserved.

Quick start

Given a TOML file like this:

[project]
name = "my-app"
version = "1.0.0"
# Search terms for the package index
keywords = ["python", "toml"]

[project.optional-dependencies]
dev = ["pytest"]

You can parse, modify, and write it back - comments, whitespace, and ordering are all preserved:

from pathlib import Path

from tomledit import Document

text = Path("pyproject.toml").read_text(encoding="utf-8")
doc = Document.parse(text)

doc["project"]["version"].comment = "# Bumped for release"
doc["project"]["version"] = "2.0.0"
doc["project"]["keywords"].append("important-keyword")
doc["project"]["keywords"].inline_comment = "# updated"

Path("pyproject.toml").write_text(str(doc), encoding="utf-8")

The result:

[project]
name = "my-app"
# Bumped for release
version = "2.0.0"
# Search terms for the package index
keywords = ["python", "toml", "important-keyword"] # updated

[project.optional-dependencies]
dev = ["pytest"]

Truth in advertising

It is intended that using a Document feels just like using a native Python dictionary, and that the Items you get from it can also be treated as ordinary dictionaries and lists and suchlike.

Under the hood, though, every Item is really a path back into the shared document.

This is mostly invisible, but sometimes the implementation leaks out. In particular: because items are paths, they can go stale when the document changes underneath them.

doc = Document.parse('arr = ["a", "b"]')
first = doc["arr"][0]       # a path to arr[0]
doc["arr"][0] = "changed"   # mutates the document
print(first)                # RuntimeError: this Item is stale

The item that performs the mutation stays valid:

arr = doc["arr"]
arr[0] = "changed"
print(arr)                  # still works - arr itself did the mutating

If you just need the plain Python value, grab it with .value before the document changes:

first = doc["arr"][0].value   # "a" - a plain str, not a path
doc["arr"][0] = "changed"
print(first)                  # still "a"