docstr-coverage
is a simple tool that lets you measure your Python source code's
docstring coverage.
It shows which of your functions, classes, methods, and modules don't have docstrings.
It also provide statistics about overall docstring coverage for individual files, and for your entire project.
>>> HunterMcGushion$ docstr-coverage /docstr_coverage/
File: "docstr_coverage/setup.py"
- No module docstring
- No docstring for `readme`
Needed: 2; Found: 0; Missing: 2; Coverage: 0.0%
File: "docstr_coverage/docstr_coverage/__init__.py"
- No module docstring
Needed: 1; Found: 0; Missing: 1; Coverage: 0.0%
File: "docstr_coverage/docstr_coverage/coverage.py"
- No docstring for `DocStringCoverageVisitor.__init__`
Needed: 11; Found: 10; Missing: 1; Coverage: 90.9%
Overall statistics for 3 files:
Docstrings needed: 14; Docstrings found: 10; Docstrings missing: 4
Total docstring coverage: 71.4%; Grade: Very good
General usage is: docstr-coverage <path to dir or module> [options]
To test a single module, named some_module.py
, run:
docstr-coverage some_module.py
To test a directory (recursively), just supply the directory some_project/src
instead:
docstr-coverage some_project/src
- --destination=<type>, -dst <type> - Set the results output destination (default stdout)
- stdout - Output to standard STDOUT.
- file - Save output to file.
- --format=<type>, -frm <type> - Set output style (default text)
- text - Output in simple style.
- markdown - Output in Markdown notation.
- --skip-magic, -m - Ignore all magic methods (except
__init__
) - --skip-init, -i - Ignore all
__init__
methods - --skip-file-doc, -f - Ignore module docstrings (at the top of files)
- --skip-private, -P - Ignore private functions (starting with a single underscore)
- --skip-class-def, -c - Ignore docstrings of class definitions
- --skip-property, -sp - Ignore functions with
@property
decorator - --include-setter, -is - Include functions with
@setter
decorator (skipped by default) - --include-deleter, -idel - Include functions with
@deleter
decorator (skipped by default) - --accept-empty, -a - Exit with code 0 if no Python files are found (default: exit code 1)
- --exclude=<regex>, -e <regex> - Filepath pattern to exclude from analysis
- To exclude the contents of a virtual environment
env
and yourtests
directory, run:docstr-coverage some_project/ -e ".*/(env|tests)"
- To exclude the contents of a virtual environment
- --verbose=<level>, -v <level> - Set verbosity level (0-3, default: 3)
- 0 - Silence
- 1 - Print overall statistics
- 2 - Also print individual statistics for each file
- 3 - Also print missing docstrings (function names, class names, etc.)
- 4 - Also print information about present docstrings
- --fail-under=<int|float>, -F <int|float> - Fail if under a certain percentage of coverage (default: 100.0)
- --badge=<filepath>, -b <filepath> - Generate a docstring coverage percent badge as an SVG saved to a given filepath
- Include the badge in a repo's README using
[![docstr_coverage](<filepath/of/your/saved/badge.svg>)](https://github.com/HunterMcGushion/docstr_coverage)
, where<filepath/of/your/saved/badge.svg>
is the path provided to the--badge
option
- Include the badge in a repo's README using
- --follow-links, -l - Follow symlinks
- --percentage-only, -p - Output only the overall coverage percentage as a float, silencing all other logging
- --help, -h - Display CLI options
All options can be saved in a config file. A file named .docstr.yaml
in the folder in which docstr-coverage
is executed is picked up automatically.
Other locations can be passed using docstr-coverage -C path/to/config.yml
or the long version --config
.
Example:
paths: # list or string
- docstr_coverage
badge: docs # Path
exclude: .*/test # regex
verbose: 3 # int (0-4)
skip_magic: True # Boolean
skip_file_doc: True # Boolean
skip_init: True # Boolean
skip_class_def: True # Boolean
skip_private: True # Boolean
follow_links: True # Boolean
accept_empty: True # Boolean
ignore_names_file: .*/test # regex
fail_under: 90 # int
percentage_only: True # Boolean
ignore_patterns: # Dict with key/value pairs of file-pattern/node-pattern
.*: method_to_ignore_in_all_files
FileWhereWeWantToIgnoreAllSpecialMethods: "__.+__"
SomeFile:
- method_to_ignore1
- method_to_ignore2
- method_to_ignore3
a_very_important_view_file:
- "^get$"
- "^set$"
- "^post$"
detect_.*:
- "get_val.*"
equivalent to
docstr-coverage docstr_coverage -e ".*/test" --skip-magic --skip-init --badge="docs" --skip-class-def etc...
Note that options passed as command line arguments have precedence over options configured in a config file.
In your config files, using ignore_patterns
, you can specify regex patterns for files names and nodes (methods, ...)
which should be ignored. See config file example above.
Note that docstr-coverage
can not parse
dynamically added documentation (e.g. through class extension).
Thus, some of your code which deliberately has no docstring might be counted as uncovered.
You can override this by adding either # docstr-coverage:inherited
(intended for use if a docstring is provided in the corresponding superclass method)
or a generic excuse with a reason, like # docstr-coverage:excused `My probably bad excuse`
.
These have to be stated right above any class or function definition
(or above the functions annotations, if applicable).
Such class or function would then be counted as if they had a docstring.
# docstr-coverage:excused `no one is reading this anyways`
class FooBarChild(FooBar):
# docstr-coverage:inherited
def function(self):
pass
You can use docstr-coverage
as a pre-commit hook by adding the following to your .pre-commit-config.yaml
file
and configuring the paths
section of the .docstr.yaml
config.
This is preferrable over pre-commit args,
as it facilitates the use of the same config in CI, pre-commit and manual runs.
repos:
- repo: https://github.com/HunterMcGushion/docstr_coverage
rev: v2.3.2 # most recent docstr-coverage release or commit sha
hooks:
- id: docstr-coverage
args: ["--verbose", "2"] # override the .docstr.yaml to see less output
You can also use docstr-coverage
as a part of your project by importing it thusly.
It will supply you with overall and per-file coverages:
from docstr_coverage import get_docstring_coverage
my_coverage = get_docstring_coverage(['some_dir/file_0.py', 'some_dir/file_1.py'])
If you want more fine grained information, try the experimental docstr_coverage.analyze()
from docstr_coverage import analyze
coverage_report = analyze(['some_dir/file_0.py', 'some_dir/file_1.py'])
coverage = coverage_report.count_aggregate().coverage()
- Thorough documentation is important to help others (and even yourself) understand your code
- As a developer, improve your code's maintainability for when you need to make updates and fix bugs
- As a user, instantly know how easy it's going to be to understand a new library * If its documentation coverage is low, you may need to figure a lot out for yourself
pip install docstr-coverage
If you like being on the cutting-edge, and you want all the latest developments, run:
pip install git+https://github.com/HunterMcGushion/docstr_coverage.git
Thank you to Alexey "DataGreed" Strelkov, and James Harlow for doing all the hard work.
docstr-coverage
simply revives and brings their efforts to Python 3. See 'THANKS.txt' for more information.