pytomlpp
This is an unofficial python wrapper for tomlplusplus (https://marzer.github.io/tomlplusplus/).
Some points you may want to know before use:
- Using
tomlplusplus
means that this module is fully compatible with TOML v1.0.0. - We convert toml structure to native python data structures (dict/list etc.) when parsing, this is more inline with what
json
module does. - The binding is using pybind11.
- The project is tested using toml-test and pytest.
Example
In [1]: import pytomlpp
In [2]: toml_string = 'hello = "世界"'
In [3]: pytomlpp.loads(toml_string)
Out[3]: {'hello': '世界'}
In [4]: type(_)
Out[4]: dict
In [6]: pytomlpp.dumps({"你好": "world"})
Out[6]: '"你好" = "world"'
Why bother?
There are some exisitng python TOML parser on the market but from my experience they are all purely implemented in python which is a bit slow.
In [1]: import pytomlpp
In [2]: import toml
In [3]: def run_parser(parser_func):
...: for i in range(1000):
...: parser_func('Cargo.toml')
...:
In [4]: %timeit run_parser(pytomlpp.load)
310 ms ± 56.7 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
In [5]: %timeit run_parser(toml.load)
3.5 s ± 162 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
In [6]: pytomlpp.lib_version
Out[6]: '1.3.2'
Installing
We recommand you to use conda
to install this package:
conda install -c dorafmon pytomlpp
If you are not using conda
then please install from source:
git clone [email protected]:bobfang1992/pytomlpp.git
cd pytomlpp
pip install .
Why not pypi?
Pypi has some rules on how to distribute pre-compiled binary for different platforms. I do not have enough experties in this area. I would love to see contribution to make this package avaliable on pypi.