Last updated:
0 purchases
pythonf2format 0.8.4
f2format
Write f-string in Python 3.6 flavour, and let f2format worry about back-port issues :beer:
Since PEP 498, Python introduced
f-string
literals in version 3.6. Though released ever since
December 23, 2016, Python
3.6 is still not widely used as expected. For those who are now used to f-strings, f2format
provides an intelligent, yet imperfect, solution of a backport compiler by converting
f-strings to str.format expressions, which guarantees you to always write f-strings in Python
3.6 flavour then compile for compatibility later.
f2format is inspired and assisted by my good mate @gousaiyang.
It functions by tokenising and parsing Python code into multiple abstract syntax trees (AST),
through which it shall synthesise and extract expressions from f-string literals, and then
reassemble the original string using str.format method. Besides
conversion and
format specification, f2format
also considered and resolved
string concatenation.
Also, it always tries to maintain the original layout of source code, and accuracy of syntax.
Installation
Note that f2format only supports Python versions since 3.3 🐍
For macOS users, f2format is now available through Homebrew:
brew tap jarryshaw/tap
brew install f2format
# or simply, a one-liner
brew install jarryshaw/tap/f2format
Simply run the following to install the current version from PyPI:
pip install f2format
Or install the latest version from the git repository:
git clone https://github.com/JarryShaw/f2format.git
cd f2format
pip install -e .
# and to update at any time
git pull
Basic Usage
CLI
It is fairly straightforward to use f2format:
context in ${...} changes dynamically according to runtime environment
usage: f2format [options] <python source files and folders...>
Convert f-string to str.format for Python 3 compatibility.
positional arguments:
SOURCE python source files and folders to be converted (${CWD})
optional arguments:
-h, --help show this help message and exit
-V, --version show program's version number and exit
-q, --quiet run in quiet mode
archive options:
duplicate original files in case there's any issue
-na, --no-archive do not archive original files
-p PATH, --archive-path PATH
path to archive original files (${CWD}/archive)
convert options:
compatibility configuration for none-unicode files
-c CODING, --encoding CODING
encoding to open source files (${LOCALE_ENCODING})
-v VERSION, --python VERSION
convert against Python version (${LATEST_VERSION})
f2format will read then convert all f-string literals in every Python file under this
path. In case there might be some problems with the conversion, f2format will duplicate all
original files it is to modify into archive directory ahead of the process, if -n not set.
For instance, the code will be converted as follows.
# the original code
var = f'foo{(1+2)*3:>5}bar{"a", "b"!r}boo'
# after `f2format`
var = 'foo{:>5}bar{!r}boo'.format((1+2)*3, ("a", "b"))
Docker
Well... it's not published to the Docker Hub yet ;)
Considering f2format may be used in scenerios where Python is not reachable.
We provide also a Docker image for those poor little guys.
See
Dockerfile for more
information.
Bundled Executable
Coming soooooooooooon...
For the worst case, we also provide bundled executables of f2format. In such case,
you may simply download it then, voilà, it's ready for you.
Special thanks to PyInstaller ❤️
Developer Reference
Automator
make-demo.sh provides a
demo script, which may help integrate f2format in your development and distribution circle.
NB: make-demo.sh is not an integrated automation script. It should be revised by design.
It assumes
all source files in /src directory
using GitHub for repository management
having release branch under /release directory
already installed f2format and twine
permission to these files and folders granted
And it will
copy setup.py and src to release directory
run f2format for Python files under release
distribute to PyPI and TestPyPI using twine
upload to release branch on GitHub
upload original files to GitHub
Environments
f2format currently supports three environment arguments:
F2FORMAT_QUIET -- run in quiet mode (same as --quiet option in CLI)
F2FORMAT_VERSION -- convert against Python version (same as --python option in CLI)
F2FORMAT_ENCODING -- encoding to open source files (same as --encoding option in CLI)
APIs
f2format -- wrapper works for conversion
f2format(filename)
Args:
filename -- str, file to be converted
Envs:
F2FORMAT_QUIET -- run in quiet mode (same as --quiet option in CLI)
F2FORMAT_ENCODING -- encoding to open source files (same as --encoding option in CLI)
F2FORMAT_VERSION-- convert against Python version (same as --python option in CLI)
Raises:
ConvertError -- when parso.ParserSyntaxError raised
convert -- the main conversion process
convert(string, source='<unknown>')
Args:
string -- str, context to be converted
source -- str, source of the context
Envs:
F2FORMAT_VERSION-- convert against Python version (same as --python option in CLI)
Returns:
str -- converted string
Raises:
ConvertError -- when parso.ParserSyntaxError raised
ConvertError -- f2format internal exception
class ConvertError(SyntaxError):
pass
Codec
NB: this project is now stalled, because I just cannot figure out how to play w/ codecs :)
f2format-codec registers a codec in Python
interpreter, which grants you the compatibility to write directly in Python 3.6 f-string syntax
even through running with a previous version of Python.
Test
The current test samples are under /test
folder. test_driver.py is the main entry point for tests.
For unittests, see test.py.
Known bugs
Since f2format is currently based on parso project,
it had encountered several compatibility and parsing issues.
Parsing f-strings with nested format specifiers produces incorrect SyntaxError (#74)
This issue has been resolved since parso version 0.5.0.
Parsing f-strings with invalid quotes in expression part does not raise SyntaxError (#86)
Parsing f-strings with seeming assignment expressions produces incorrect SyntaxError (#87)
Contribution
Contributions are very welcome, especially fixing bugs and providing test cases, which
@gousaiyang is to help with, so to speak. Note that code must
remain valid and reasonable.
See Also
babel
poseur
walrus
vermin
For personal and professional use. You cannot resell or redistribute these repositories in their original state.
There are no reviews.