Skip to main content

Derives type annotations from Sphinx comments in Python source

Project description

What?

Extracts type annotations from Python source documented with Sphinx-compatible docstrings.

Why?

Existing tools like PyAnnotate require you to write a driver to exercise your application, while a special hook collects type information and converts it to annotations. This works really well, but requires significant effort in writing a comprehensive driver that covers your entire application.

If your application already has type information in its docstrings, it may be simpler to convert that information to annotations directly. Kutuzov will help you with that.

How?

Kutuzov does this:

  • Scans packages and submodules for docstrings

  • Parses the docstrings, extracts type information about each parameter

  • Writes a PyAnnotate-compatible JSON file

What remains for you is:

  • Run PyAnnotate and consume the generated JSON file

  • Tweak the results by hand

  • PROFIT

The amount of tweaking you have to do may vary. It significantly depends on the quality and accuracy of your docstrings.

Example

We will be annotating the example.py file. You will need pyannotate (pip install pyannotate) to proceed.

$ python -m kutuzov example.py > type_info.json
$ cat type_info.json
[
    {
        "func_name": "Dog.__init__",
        "line": 2,
        "path": "example.py",
        "samples": 1,
        "type_comments": [
            "(str) -> None"
        ]
    },
    {
        "func_name": "Dog.bark",
        "line": 16,
        "path": "example.py",
        "samples": 1,
        "type_comments": [
            "(str) -> str"
        ]
    }
]
$ pyannotate example.py --py3
Refactored example.py
--- example.py  (original)
+++ example.py  (refactored)
@@ -1,5 +1,5 @@
 class Dog:
-    def __init__(self, name):
+    def __init__(self, name: str) -> None:
         """
         :param str name: The name of this dog.
         """
@@ -13,7 +13,7 @@
         """
         return self._name

-    def bark(self, language='en'):
+    def bark(self, language: str = 'en') -> str:
         """
         Make some noise!

Files that need to be modified:
example.py
NOTE: this was a dry run; use -w to write files

You may have noticed that [example.py](example.py) already contains a PyAnnotate driver in the mainline.

$ cat type_info.json
[
    {
        "path": "example.py",
        "line": 2,
        "func_name": "Dog.__init__",
        "type_comments": [
            "(str) -> None"
        ],
        "samples": 1
    },
    {
        "path": "example.py",
        "line": 8,
        "func_name": "name",
        "type_comments": [
            "() -> str"
        ],
        "samples": 1
    },
    {
        "path": "example.py",
        "line": 16,
        "func_name": "Dog.bark",
        "type_comments": [
            "(str) -> str"
        ],
        "samples": 1
    }
]

If you compare that to the previously generated JSON, you will see that it is mostly similar. The greatest difference is that Kutuzov missed the name property - it doesn’t know how to handle those yet.

What’s in a name?

Mikhail Kutuzov was a contemporary of Napoleon Bonaparte. A particular painting depicts Napoleon before the Sphinx. And Sphinx, of course, is the project that we all love for our Python documentation needs.

This project was partially inspired by sphinx.ext.napoleon, which does something similar, but for a different docstring format.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

kutuzov-0.1.0.tar.gz (5.6 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page