# -*- coding: utf-8 -*-
# ----------------------------------------------------------------------------
#
# TITLE : numpy Parse tools
# PROJECT : utilipy
#
# ----------------------------------------------------------------------------
# Docstring
"""Numpy parse tools.
routines to parse docstrings according to the Napoleon format specifications.
code sourced from [1]_, but modified to merge, not replace, docstring sections.
Routine Listings
----------------
merge_numpy_docs
Merge two numpy-style docstrings into a single docstring
References
----------
The code source [1]_.
.. [1] https://github.com/rsokl/custom_inherit
TODO
----
https://stackoverflow.com/a/41083968
"""
__author__ = "Ryan Soklaski"
__license__ = "MIT"
__version__ = "2.2.2"
__email__ = "https://github.com/rsokl/custom_inherit"
__all__ = ["merge_numpy_docs"]
###############################################################################
# IMPORTS
# GENERAL
from collections import OrderedDict
from inspect import cleandoc
import typing
from typing import Union
from typing_extensions import Literal
###############################################################################
# CODE
###############################################################################
def parse_numpy_doc(doc):
"""Extract the text from the sections of a numpy-formatted docstring.
Parameters
----------
doc: Union[str, None]
Returns
-------
OrderedDict[str, Union[None,str]]
The extracted numpy-styled docstring sections.
"""
doc_sections = OrderedDict(
[
("Short Summary", None),
("Deprecation Warning", None),
("Attributes", None),
("Extended Summary", None),
("Parameters", None),
("Returns", None),
("Yields", None),
("Other Parameters", None),
("Raises", None),
("See Also", None),
("Notes", None),
("References", None),
("Examples", None),
]
)
if not doc:
return doc_sections
doc = cleandoc(doc)
lines = iter(doc.splitlines())
key = "Short Summary"
body = []
while True:
try:
line = next(lines).rstrip()
if line in doc_sections:
doc_sections[key] = "\n".join(body).rstrip() if body else None
body = []
key = line
next(lines) # skip section delimiter
else:
body.append(line)
except StopIteration:
doc_sections[key] = "\n".join(body)
break
return doc_sections
# /def
def merge_section(
key: str,
prnt_sec: Union[str, None],
child_sec: Union[str, None],
method: Union[Literal["merge"], Literal["replace"]] = "replace",
) -> Union[str, None]:
"""Synthesize a output numpy docstring section.
Parameters
----------
key: str
The numpy-section being merged.
prnt_sec: Optional[str]
The docstring section from the parent's attribute.
child_sec: Optional[str]
The docstring section from the child's attribute.
method: {merge, replace}, optional
'merge' merges the sections.
'replace' (default) parent with child sections, if exists.
Returns
-------
Optional[str]
The output docstring section.
"""
if prnt_sec is None and child_sec is None:
return None
if key == "Short Summary":
header = ""
else:
header = "\n".join((key, "".join("-" for i in range(len(key))), ""))
if method == "replace":
body = prnt_sec if child_sec is None else child_sec
elif method == "merge":
# TODO more intelligent merge strategy,
# respecting arg vs kwarg order, overriding repeated args, etc.
if prnt_sec is None:
body = child_sec or ""
elif child_sec is None:
body = prnt_sec or ""
else:
body = prnt_sec + "\n" + child_sec
return header + body
# /def
def merge_all_sections(
prnt_sctns: typing.MutableMapping[str, Union[None, str]],
child_sctns: typing.MutableMapping[str, Union[None, str]],
method: Union[Literal["merge"], Literal["replace"]] = "replace",
) -> Union[str, None]:
"""Merge the doc-sections of the parent's and child's attribute.
Parameters
----------
prnt_sctns: OrderedDict[str, Union[None,str]]
child_sctns: OrderedDict[str, Union[None,str]]
method: {merge, replace}, optional
'merge' merges the sections.
'replace' (default) parent with child sections, if exists.
Returns
-------
str
Output docstring of the merged docstrings.
"""
doc = []
prnt_only_raises = prnt_sctns["Raises"] and not (
prnt_sctns["Returns"] or prnt_sctns["Yields"]
)
if prnt_only_raises and (child_sctns["Returns"] or child_sctns["Yields"]):
prnt_sctns["Raises"] = None
for key in prnt_sctns:
sect = merge_section(
key, prnt_sctns[key], child_sctns[key], method=method
)
if sect is not None:
doc.append(sect)
return "\n\n".join(doc) if doc else None
# /def
[docs]def merge_numpy_docs(
prnt_doc: Union[str, None] = None,
child_doc: Union[str, None] = None,
method: Union[Literal["merge"], Literal["replace"]] = "replace",
) -> Union[str, None]:
"""Merge two numpy-style docstrings into a single docstring.
Given the numpy-style docstrings from a parent and child's attributes,
merge the docstring sections such that the child's section is used,
wherever present, otherwise the parent's section is used.
Any whitespace that can be uniformly removed from a docstring's second line
and onwards is removed. Sections will be separated by a single blank line.
Parameters
----------
prnt_doc: Optional[str]
The docstring from the parent.
child_doc: Optional[str]
The docstring from the child.
method: {merge, replace}, optional
'merge' merges the sections.
'replace' (default) parent with child sections, if exists.
Returns
-------
Union[str, None]
The merged docstring.
"""
return merge_all_sections(
parse_numpy_doc(prnt_doc), parse_numpy_doc(child_doc), method=method
)
# /def
###############################################################################
# END