Import Upstream version 0.5.4

This commit is contained in:
su-fang 2022-10-10 17:08:27 +08:00
commit 3f079e203d
11 changed files with 722 additions and 0 deletions

10
.gitignore vendored Normal file
View File

@ -0,0 +1,10 @@
*.pyc
build/
dist/
docs/build/output/
*.orig
alembic.ini
tox.ini
.venv
.egg-info
.coverage

28
.pre-commit-config.yaml Normal file
View File

@ -0,0 +1,28 @@
# See https://pre-commit.com for more information
# See https://pre-commit.com/hooks.html for more hooks
repos:
- repo: https://github.com/python/black
rev: 20.8b1
hooks:
- id: black
- repo: https://github.com/sqlalchemyorg/zimports
rev: master
hooks:
- id: zimports
- repo: https://github.com/pycqa/flake8
rev: master
hooks:
- id: flake8
additional_dependencies:
- flake8-import-order
- flake8-builtins
- flake8-docstrings
- flake8-rst-docstrings
- pydocstyle
- pygments

19
LICENSE Normal file
View File

@ -0,0 +1,19 @@
This is the MIT license: http://www.opensource.org/licenses/mit-license.php
Copyright (C) by Michael Bayer.
Permission is hereby granted, free of charge, to any person obtaining a copy of this
software and associated documentation files (the "Software"), to deal in the Software
without restriction, including without limitation the rights to use, copy, modify, merge,
publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons
to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or
substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE
FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
DEALINGS IN THE SOFTWARE.

7
MANIFEST.in Normal file
View File

@ -0,0 +1,7 @@
recursive-include sphinx_paramlinks *.py *.css
include README* LICENSE
exclude pyproject.toml

141
README.rst Normal file
View File

@ -0,0 +1,141 @@
==================
Sphinx Paramlinks
==================
A `Sphinx <http://sphinx.pocoo.org/>`_ extension which allows ``:param:``
directives within Python documentation to be linkable.
This is an experimental extension that's used by the
`SQLAlchemy <http://www.sqlalchemy.org>`_ project and related projects.
Configuration
=============
Just turn it on in ``conf.py``::
extensions = [
'sphinx_paramlinks',
# your other sphinx extensions
# ...
]
Since version 0.5.3, you can modify how clickable hyperlinks are placed around the names of
the parameter using the ``paramlinks_hyperlink_param`` setting in ``conf.py``::
paramlinks_hyperlink_param='name'
This parameter accepts the following values:
* ``'none'``: No link will be be inserted. The parameter still has a target
attached to it so that you can e.g. jump to it from the search.
* ``'name'``: The parameter name is a clickable hyperlink.
* ``'link_symbol'``: A clickable link symbol is inserted after the parameter
name (but before an eventual type specification). By default, this symbol
only shows when hovering the parameter description (see below)
* ``'name_and_symbol'``: link both the name and also generate a link symbol.
The default is ``paramlinks_hyperlink_param = 'link_symbol'``.
Features
========
* ``:param:`` directives within Sphinx function/method descriptions
will be given a paragraph link so that they can be linked
to externally.
* A new text role ``:paramref:`` is added, which works like ``:meth:``,
``:func:``, etc. Just append the parameter name as an additional token::
:paramref:`.EnvironmentContext.configure.transactional_ddl`
The directive makes use of the existing Python role to do the method/function
lookup, searching first the ``:meth:``, then the ``:class:``, and then the
``:func:`` role; then the parameter name is applied separately to produce the
final reference link. (new in 0.3.4, search for ``:meth:`` / ``:func:`` /
``:class:`` individually rather than using ``:obj:`` which catches lots of
things that don't have parameters)
* The paramlinks are also added to the master index as well as the list
of domain objects, which allows them to be searchable through the
searchindex.js system. (new in 0.3.0)
Stylesheet
==========
The paragraph link involves a short stylesheet, to allow the links to
be visible when hovered. This sheet is called
``sphinx_paramlinks.css`` and the plugin will copy it to the ``_static``
directory of the output automatically. The stylesheet is added to the
``css_files`` list present in the template namespace for Sphinx via the
``Sphinx.add_stylesheet()`` hook.
Customization
-------------
To customize the link styling, you can override the configuration of
``sphinx_paramlinks.css`` by adding a custom style sheet via::
app.add_css_file("path/to/custom.css")
If the parameter name is a hyperlink, the HTML code will look something like
this::
<a class="paramname reference internal" href="#package.method.params.parameter_name">
<strong>parameter_name</strong>
</a>
The class ``paramname`` is defined by ``sphinx-paramlinks`` and can be used to
customize the styling.
If a link symbol is inserted after the hyperlink, the HTML code will look
something like this::
<a class="paramlink headerlink reference internal" href="#package.method.params.parameter_name">¶</a>
The class ``paramlink`` is defined by ``sphinx-paramlinks`` and can be used to
customize the styling.
Compatibility
=============
Python Compatibility
--------------------
sphinx-paramlinks is fully Python 3 compatible.
Sphinx Compatibility
--------------------
I've tried *very* hard to make as few assumptions as possible about Sphinx
and to use only very simple public APIs, so that architectural changes in future
Sphinx versions won't break this plugin. To come up with this plugin I
spent many hours with Sphinx source and tried many different approaches to
various elements of functionality; hopefully what's here is as simple and
stable as possible based on the current extension capabilities of Sphinx.
One element that involves using a bit of internals is the usage of the
``sphinx.domains.python.PyXRefRole`` class, which is currently the
Sphinx class that defines roles for things like ``:meth:``,
``:func:``, etc. The object is used as-is in order to define the
``:paramref:`` role; the product of this role is later transformed
using standard hooks.
Another assumption is that in order to locate the RST nodes Sphinx
creates for the ``:param:`` tags, we look at ``nodes.strong``,
assuming that this is the type of node currently used to render
``:param:`` within RST. If this changes, or needs to be expanded to
support other domains, this traversal can be opened up as needed.
This part was difficult as Sphinx really doesn't provide any hooks
into how the "Info Field List" aspect of domains is handled.
Overall, the approach here is to apply extra information to constructs
going into the Sphinx system, then do some transformations as the data
comes back out. This relies on as little of how Sphinx does its
thing as possible, rather than going with custom domains and heavy use
of injected APIs which may change in future releases.

2
pyproject.toml Normal file
View File

@ -0,0 +1,2 @@
[tool.black]
line-length = 79

15
setup.cfg Normal file
View File

@ -0,0 +1,15 @@
[flake8]
show-source = false
enable-extensions = G
# E203 is due to https://github.com/PyCQA/pycodestyle/issues/373
ignore =
A003,
D,
E203,E305,E711,E712,E721,E722,E741,
N801,N802,N806,
RST304,RST303,RST299,RST399,
W503,W504
exclude = .venv,.git,.tox,dist,doc,*egg,build
import-order-style = google
application-import-names = sqlalchemy,test

45
setup.py Normal file
View File

@ -0,0 +1,45 @@
#!/usr/bin/env python
import os
import re
from setuptools import setup
v = open(
os.path.join(os.path.dirname(__file__), "sphinx_paramlinks/__init__.py")
)
VERSION = (
re.compile(r".*__version__ = [\"'](.*?)[\"']", re.S)
.match(v.read())
.group(1)
)
v.close()
readme = os.path.join(os.path.dirname(__file__), "README.rst")
setup(
name="sphinx-paramlinks",
version=VERSION,
description="Allows param links in Sphinx function/method "
"descriptions to be linkable",
long_description=open(readme).read(),
classifiers=[
"Development Status :: 3 - Alpha",
"Environment :: Console",
"Intended Audience :: Developers",
"Programming Language :: Python",
"Programming Language :: Python :: 3",
"Programming Language :: Python :: Implementation :: CPython",
"Programming Language :: Python :: Implementation :: PyPy",
"Topic :: Documentation",
],
keywords="sphinx",
author="Mike Bayer",
author_email="mike@zzzcomputing.com",
url="http://github.com/sqlalchemyorg/sphinx-paramlinks",
license="MIT",
packages=["sphinx_paramlinks"],
include_package_data=True,
zip_safe=False,
)

View File

@ -0,0 +1,3 @@
__version__ = "0.5.3"
from .sphinx_paramlinks import setup # noqa

View File

@ -0,0 +1,13 @@
li > a.paramlink {
visibility: hidden;
}
p:hover > a.headerlink, p:hover > a.paramlink, li:hover > a.paramlink, li:hover > a.headerlink {
visibility: visible;
}
a.paramname {
color:inherit;
}

View File

@ -0,0 +1,439 @@
#!coding: utf-8
from distutils.version import LooseVersion
from enum import Enum
import os
import re
from docutils import nodes
from docutils.transforms import Transform
from sphinx import __version__
from sphinx import addnodes
from sphinx.domains import ObjType
from sphinx.domains.python import ObjectEntry
from sphinx.domains.python import PythonDomain
from sphinx.domains.python import PyXRefRole
from sphinx.util import logging
from sphinx.util.console import bold
from sphinx.util.osutil import copyfile
# the searchindex.js system relies upon the object types
# in the PythonDomain to create search entries
PythonDomain.object_types["parameter"] = ObjType("parameter", "param")
LOG = logging.getLogger(__name__)
def _is_html(app):
return app.builder.name in ("html", "readthedocs")
# https://www.sphinx-doc.org/en/master/extdev/deprecated.html
# Constants for link styles
class HyperlinkStyle(Enum):
NONE = "none"
NAME = "name"
LINK_SYMBOL = "link_symbol"
NAME_AND_SYMBOL = "name_and_symbol"
def _indexentries(env):
return env.domains["index"].entries
def _tempdata(app):
if "_sphinx_paramlinks_index" in _indexentries(app.env):
idx = _indexentries(app.env)["_sphinx_paramlinks_index"]
else:
_indexentries(app.env)["_sphinx_paramlinks_index"] = idx = {}
return idx
def autodoc_process_docstring(app, what, name, obj, options, lines):
# locate :param: lines within docstrings. augment the parameter
# name with that of the parent object name plus a token we can
# spot later. Also put an index entry in a temporary collection.
idx = _tempdata(app)
docname = app.env.temp_data.get("docname")
if not docname:
return
if docname in idx:
doc_idx = idx[docname]
else:
idx[docname] = doc_idx = []
def _cvt_param(name, line):
if name.endswith(".__init__"):
# kill off __init__ if present, the links are always
# off the class
name = name[0:-9]
def cvt(m):
role, modifier, objname, paramname = m.group(1), m.group(2) or "", name, m.group(3)
refname = _refname_from_paramname(paramname, strip_markup=True)
item = (
"single",
"%s (%s parameter)" % (refname, objname),
"%s.params.%s" % (objname, refname),
"",
)
if LooseVersion(__version__) >= LooseVersion("1.4.0"):
item += (None,)
doc_idx.append(item)
return ":%s %s_sphinx_paramlinks_%s.%s:" % (
role,
modifier,
objname,
paramname,
)
def secondary_cvt(m):
role, modifier, objname, paramname = m.group(1), m.group(2) or "", name, m.group(3)
return ":%s %s_sphinx_paramlinks_%s.%s:" % (
role,
modifier,
objname,
paramname,
)
line = re.sub(r"^:(keyword|param) ([^:]+? )?([^:]+?):", cvt, line)
line = re.sub(r"^:(kwtype|type) ([^:]+? )?([^:]+?):", secondary_cvt, line)
return line
if what in ("function", "method", "class"):
lines[:] = [_cvt_param(name, line) for line in lines]
def _refname_from_paramname(paramname, strip_markup=False):
literal_match = re.match(r"^``(.+?)``$", paramname)
if literal_match:
paramname = literal_match.group(1)
refname = paramname
eq_match = re.match(r"(.+?)=.+$", refname)
if eq_match:
refname = eq_match.group(1)
if strip_markup:
refname = re.sub(r"\\", "", refname)
return refname
class ApplyParamPrefix(Transform):
"""Obfuscate the token name inside of a ":paramref:" reference
to prevent Sphinx from resolving the node
and generating a reference node that doesn't look the way we want it to.
Ensure that our own missing-reference resolver is used.
"""
default_priority = 210
def apply(self):
for ref in self.document.traverse(addnodes.pending_xref):
# look only at paramref
if ref["reftype"] != "paramref":
continue
# for params that explicitly have ".params." in the reference
# source, let those just resolve normally. this is not
# really expected but it seems to work already.
if "params." in ref["reftarget"]:
continue
target_tokens = ref["reftarget"].split(".")
# apply a token to the link that will completely prevent
# Sphinx from ever resolving this node, because WE want to
# resolve and render the reference node, ALWAYS, THANK YOU!
target_tokens[-1] = "_sphinx_paramlinks_" + target_tokens[-1]
ref["reftarget"] = ".".join(target_tokens)
class LinkParams(Transform):
# apply references targets and optional references
# to nodes that contain our target text.
default_priority = 210
def apply(self):
config_value = (
self.document.settings.env.app.config.paramlinks_hyperlink_param
)
try:
link_style = HyperlinkStyle[config_value.upper()]
except KeyError as exc:
raise ValueError(
f"Unknown value {repr(config_value)} for "
f"'paramlinks_hyperlink_param'. "
f"Must be one of "
f"""{
', '.join(repr(member.value) for member in HyperlinkStyle)
}."""
) from exc
if link_style is HyperlinkStyle.NONE:
return
is_html = _is_html(self.document.settings.env.app)
# search <strong> nodes, which will include the titles for
# those :param: directives, looking for our special token.
# then fix up the text within the node.
for ref in self.document.traverse(nodes.strong):
text = ref.astext()
if text.startswith("_sphinx_paramlinks_"):
components = re.match(r"_sphinx_paramlinks_(.+)\.(.+)$", text)
location, paramname = components.group(1, 2)
refname = _refname_from_paramname(paramname)
refid = "%s.params.%s" % (location, refname)
ref.parent.insert(0, nodes.target("", "", ids=[refid]))
del ref[0]
ref.insert(0, nodes.Text(paramname, paramname))
if is_html:
# add the "p" thing only if we're the HTML builder.
# using a real ¶, surprising, right?
# http://docutils.sourceforge.net/FAQ.html
# #how-can-i-represent-esoteric-characters-
# e-g-character-entities-in-a-document
# "For example, say you want an em-dash (XML
# character entity &mdash;, Unicode character
# U+2014) in your document: use a real em-dash.
# Insert concrete characters (e.g. type a real em-
# dash) into your input file, using whatever
# encoding suits your application, and tell
# Docutils the input encoding. Docutils uses
# Unicode internally, so the em-dash character is
# a real em-dash internally." OK !
for pos, node in enumerate(ref.parent.children):
# try to figure out where the node with the
# paramname is. thought this was simple, but
# readthedocs proving..it's not.
# TODO: need to take into account a type name
# with the parens.
if (
isinstance(node, nodes.TextElement)
and node.astext() == paramname
):
break
else:
return
refparent = ref.parent
if link_style in (
HyperlinkStyle.NAME,
HyperlinkStyle.NAME_AND_SYMBOL,
):
# If the parameter name should be a href, we wrap it
# into an <a></a> tag
element = refparent.pop(pos)
# note this is expected to be the same....
# assert element is ref
newnode = nodes.reference(
"",
"",
# needed to avoid recursion overflow
element.deepcopy(),
refid=refid,
classes=["paramname"],
)
refparent.insert(pos, newnode)
if link_style in (
HyperlinkStyle.LINK_SYMBOL,
HyperlinkStyle.NAME_AND_SYMBOL,
):
# If there should be a link symbol after the parameter
# name, insert it here
refparent.insert(
pos + 1,
nodes.reference(
"",
"",
nodes.Text("", ""),
refid=refid,
# paramlink is our own CSS class, headerlink
# is theirs. Trying to get everything we can
# for existing symbols...
classes=["paramlink", "headerlink"],
),
)
def lookup_params(app, env, node, contnode):
# here, we catch the "pending xref" nodes that we created with
# the "paramref" role we added. The resolve_xref() routine
# knows nothing about this node type so it never finds anything;
# the Sphinx BuildEnvironment then gives us one more chance to do a lookup
# here.
if node["reftype"] != "paramref":
return None
target = node["reftarget"]
tokens = target.split(".")
# if we just have :paramref:`arg` and not :paramref:`namespace.arg`,
# we must assume that the current namespace is meant.
if tokens == [target]:
#
# node.source is expected to look like:
# /path/to/file.py:docstring of module.clsname.methname
#
docstring_match = re.match(r".*?:docstring of (.*)", node.source)
if docstring_match:
full_attr_path = docstring_match.group(1)
fn_name = full_attr_path.split(".")[-1]
tokens.insert(0, fn_name)
resolve_target = ".".join(tokens[0:-1])
# we are now cleared of Sphinx's resolver.
# remove _sphinx_paramlinks_ token from refid so we can search
# for the node normally.
paramname = tokens[-1].replace("_sphinx_paramlinks_", "")
# Remove _sphinx_paramlinks_ obfuscation string from all text nodes
# for rendering.
for replnode in (node, contnode):
for text_node in replnode.traverse(nodes.Text):
text_node.parent.replace(
text_node,
nodes.Text(text_node.replace("_sphinx_paramlinks_", "")),
)
# emulate the approach within
# sphinx.environment.BuildEnvironment.resolve_references
try:
domain = env.domains[node["refdomain"]] # hint: this will be 'py'
except KeyError:
return None
# BuildEnvironment doesn't pass us "fromdocname" here as the
# fallback, oh well
refdoc = node.get("refdoc", None)
# we call the same "resolve_xref" that BuildEnvironment just tried
# to call for us, but we load the call with information we know
# it can find, e.g. the "object" role (or we could use :meth:/:func:)
# along with the classname/methodname/funcname minus the parameter
# part.
for search in ["meth", "class", "func"]:
newnode = domain.resolve_xref(
env, refdoc, app.builder, search, resolve_target, node, contnode
)
if newnode is not None:
break
if newnode is not None:
# assuming we found it, tack the paramname back onto to the final
# URI
if "refuri" in newnode:
newnode["refuri"] += ".params." + paramname
elif "refid" in newnode:
newnode["refid"] += ".params." + paramname
return newnode
def add_stylesheet(app):
# changed in 1.8 from add_stylesheet()
# https://www.sphinx-doc.org/en/master/extdev/appapi.html
# #sphinx.application.Sphinx.add_css_file
app.add_css_file("sphinx_paramlinks.css")
def copy_stylesheet(app, exception):
LOG.info(
bold("The name of the builder is: %s" % app.builder.name), nonl=True
)
if not _is_html(app) or exception:
return
LOG.info(bold("Copying sphinx_paramlinks stylesheet... "), nonl=True)
source = os.path.abspath(os.path.dirname(__file__))
# the '_static' directory name is hardcoded in
# sphinx.builders.html.StandaloneHTMLBuilder.copy_static_files.
# would be nice if Sphinx could improve the API here so that we just
# give it the path to a .css file and it does the right thing.
dest = os.path.join(app.builder.outdir, "_static", "sphinx_paramlinks.css")
copyfile(os.path.join(source, "sphinx_paramlinks.css"), dest)
LOG.info("done")
def build_index(app, doctree):
entries = _tempdata(app)
for docname in entries:
doc_entries = entries[docname]
_indexentries(app.env)[docname].extend(doc_entries)
if LooseVersion(__version__) >= LooseVersion("4.0.0"):
for entry in doc_entries:
sing, desc, ref, extra = entry[:4]
app.env.domains["py"].data["objects"][ref] = ObjectEntry(
docname, ref, "parameter", False
)
elif LooseVersion(__version__) >= LooseVersion("3.0.0"):
for entry in doc_entries:
sing, desc, ref, extra = entry[:4]
app.env.domains["py"].data["objects"][ref] = ObjectEntry(
docname,
ref,
"parameter",
)
else:
for entry in doc_entries:
sing, desc, ref, extra = entry[:4]
app.env.domains["py"].data["objects"][ref] = (
docname,
"parameter",
)
_indexentries(app.env).pop("_sphinx_paramlinks_index")
def setup(app):
app.add_transform(LinkParams)
app.add_transform(ApplyParamPrefix)
# Make sure that default is are the same as in LinkParams
# When config changes, the whole env needs to be rebuild since
# LinkParams is applied while building the doctrees
app.add_config_value(
"paramlinks_hyperlink_param",
HyperlinkStyle.LINK_SYMBOL.name,
"env",
[str],
)
# PyXRefRole is what the sphinx Python domain uses to set up
# role nodes like "meth", "func", etc. It produces a "pending xref"
# sphinx node along with contextual information.
app.add_role_to_domain("py", "paramref", PyXRefRole())
app.connect("autodoc-process-docstring", autodoc_process_docstring)
app.connect("builder-inited", add_stylesheet)
app.connect("build-finished", copy_stylesheet)
app.connect("doctree-read", build_index)
app.connect("missing-reference", lookup_params)
return {
"parallel_read_safe": True,
"parallel_write_safe": True,
}