Implementation:Scikit learn Scikit learn NumpyDocScrape

Knowledge Sources	Scikit_learn Scikit-learn Docs
Domains	Documentation, Utilities
Last Updated	2026-02-08 15:00 GMT

Overview

Concrete tool for extracting and parsing NumPy-style docstrings into structured data provided by scikit-learn.

Description

This module provides classes for parsing NumPy-style docstrings into an abstract structured representation. The main class NumpyDocString parses docstring sections (Parameters, Returns, Examples, etc.) into a mapping from section titles to structured data. It also includes a Reader class for line-based string reading and a FunctionDoc class that extracts documentation directly from function objects.

Usage

Use this module when you need to programmatically parse, inspect, or validate NumPy-style docstrings in Python code, particularly within the scikit-learn documentation toolchain.

Code Reference

Source Location

• Repository: scikit-learn
• File: sklearn/externals/_numpydoc/docscrape.py

Signature

class Reader:
    """A line-based string reader."""
    def __init__(self, data):
        ...

class NumpyDocString(Mapping):
    """Parses a numpydoc string to an abstract representation."""
    def __init__(self, docstring, config=None):
        ...

class FunctionDoc(NumpyDocString):
    def __init__(self, func, role="func", doc=None, config=None):
        ...

Import

from sklearn.externals._numpydoc.docscrape import NumpyDocString, FunctionDoc

I/O Contract

Inputs

Name	Type	Required	Description
docstring	str	Yes	NumPy-style docstring text to parse
config	dict	No	Configuration options for parsing behavior
func	callable	Yes (FunctionDoc)	Function object to extract documentation from
role	str	No	Role descriptor like 'func' or 'meth' (default: 'func')

Outputs

Name	Type	Description
NumpyDocString instance	Mapping	Mapping from section names to structured content
Parameters section	list	List of (name, type, description) tuples
Returns section	list	List of (name, type, description) tuples
Examples section	str	Raw example code text

Usage Examples

Basic Usage

from sklearn.externals._numpydoc.docscrape import NumpyDocString

doc = NumpyDocString("""
Summary line.

Parameters
----------
x : int
    The input value.
y : float, optional
    Another parameter.

Returns
-------
result : bool
    The output.
""")

print(doc['Parameters'])
print(doc['Returns'])

Related Pages

• Principle:Scikit_learn_Scikit_learn_Dataset_Loading