f2format 0.4.1

Back-port compiler for Python 3.6 f-string literals.

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 literal 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-string, f2format provides an intelligent, yet imperfect, solution of a backport compiler by converting f-strings to str.format literals, which guarantees you to always write f-string in Python 3.6 flavour then compile for compatibility later.

  f2format is inspired and assisted by my 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

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
                        (default is '${CWD}')

optional arguments:
  -h, --help            show this help message and exit
  -V, --version         show program's version number and exit

archive options:
  duplicate original files in case there's any issue

  -n, --no-archive      do not archive original files
  -p PATH, --archive-path PATH
                        path to archive original files (default is '${CWD}/archive')

convert options:
  compatibility configuration for none-unicode files

  -c CODING, --encoding CODING
                        encoding to open source files (default is '${ENCODING}')
  -v VERSION, --python-version VERSION
                        Python version of lexical scanner to analyse source
                        files (default is Python ${PYTHONVERSION})

  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"))

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

APIs

f2format.f2format(filename)

Wrapper works for conversion.

Args:

• filename -- str, file to be converted

f2format.convert(string, lineno)

The main conversion process.

Args:

• string -- str, context to be converted
• lineno -- dict<int: int>, line number to actual offset mapping

Returns:

• str -- converted string

Codec

  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.

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.