npdoc2md

A simple python utility for auto-converting numpy-style python docstrings to markdown for use with mkdocs, an entire package at a time. It also includes a second utility, code2npdoc, which will attempt to generate docstring templates given code for an entire project.

Installation

The preferred method for installing npdoc2md is to use pip. (Python 3.5+ required)

pip install npdoc2md

If pip defaults to Python2, use pip3. You can also install from source:

git clone https://github.com/jwlodek/npdoc2md
cd npdoc2md
pip install .

Usage

Below is the result of running npdoc2md with the -h flag:

usage: npdoc2md [-h] [-v] -i INPUT -o OUTPUT [-s SKIP [SKIP ...]] [-d]

optional arguments:
  -h, --help            show this help message and exit
  -v, --version         Use this flag to print out npdoc2md version info.
  -i INPUT, --input INPUT
                        The path to the target python project or file to
                        convert.
  -o OUTPUT, --output OUTPUT
                        The output directory where the markdown files should
                        be placed.
  -s SKIP [SKIP ...], --skip SKIP [SKIP ...]
                        List of filenames/directories to skip.
  -d, --debug           Add this flag to print detailed log messages during
                        conversion.

Basic usage will require at least target and output locations that are valid.

npdoc2md -i C:\Users\jwlodek\demo -o C:\Users\jwlodek\demo_output

You can also specify to enable debug printing with -d, and files to skip with -s followed by a list of files. For example, to autogenerate py_cui docs, the following command is run:

npdoc2md -i ..\..\..\py_cui -o ..\..\DocstringGenerated -s statusbar.py errors.py

which will ignore the statusbar.py and errors.py files.

The npdoc2md script will recursively search the target (if it is a folder) for files ending with the .py extension, and will generate a markdown file for each one not specified in the ignore section. Any encountered __init__.py files will use their containing folder's name for the name of the output markdown file.

Doc Rules

You must follow strict docstring style rules to use npdoc2md:

• Each class of function docstring must start and end with """, and the initial description must be right after the initial """. Ex: """Hello this is a function
• Use numpy style guidelines for Attributes, Parameters, Returns, Raises lists
• The Returns list should give a return value name and type with the doc message. If it doesn't, a generic name will be assigned to the return variable

Examples

As stated previously, py_cui uses npdoc2md to auto-generate documentation to use with mkdocs. You may also see the Npdoc2md.md file in this repository which was generated by running this script on itself:

npdoc2md -i .\npdoc2md.py -o .\example\.

Generating template docs

Writing out all of the docstrings for a project is a lengthy process, so a second helper script was written to help with generating template np docs. It has the following usage:

usage: code2npdoc [-h] [-v] -i INPUT [-c] [-s SKIP [SKIP ...]] [-d]

A utility for auto-creating base numpy style docstrings for an entire python
project.

optional arguments:
  -h, --help            show this help message and exit
  -v, --version         Add this flag for displaying version information.
  -i INPUT, --input INPUT
                        Path to target python project or file
  -c, --createtemp      If this flag is set, code2npdoc will create a
                        temporary conversion folder without overriding your
                        sources.
  -s SKIP [SKIP ...], --skip SKIP [SKIP ...]
                        List of filenames/directories to skip.
  -d, --debug           Add this flag to print detailed log messages during
                        conversion.

With similar flags as the base npdoc2md script. Add the -c flag if you don't want your original file to be auto-overwritten. An example of running this script on the npdoc2md file is as follows:

code2npdoc -i npdoc2md.py -c

License

MIT License
Copyright (c) 2020, Jakub Wlodek