def read_doc(app: "Sphinx", env: BuildEnvironment,
filename: str) -> nodes.document:
"""Parse a document and convert to doctree."""
# set up error_handler for the target document
error_handler = UnicodeDecodeErrorHandler(env.docname)
codecs.register_error('sphinx', error_handler) # type: ignore
reader = SphinxStandaloneReader()
reader.setup(app)
filetype = get_filetype(app.config.source_suffix, filename)
parser = app.registry.create_source_parser(app, filetype)
if parser.__class__.__name__ == 'CommonMarkParser' and parser.settings_spec == (
):
# a workaround for recommonmark
# If recommonmark.AutoStrictify is enabled, the parser invokes reST parser
# internally. But recommonmark-0.4.0 does not provide settings_spec for reST
# parser. As a workaround, this copies settings_spec for RSTParser to the
# CommonMarkParser.
parser.settings_spec = RSTParser.settings_spec
input_class = app.registry.get_source_input(filetype)
if input_class:
# Sphinx-1.8 style
source = input_class(
app,
env,
source=None,
source_path=filename, # type: ignore
encoding=env.config.source_encoding)
pub = Publisher(
reader=reader,
parser=parser,
writer=SphinxDummyWriter(),
source_class=SphinxDummySourceClass, # type: ignore
destination=NullOutput())
pub.process_programmatic_settings(None, env.settings, None)
pub.set_source(source, filename)
else:
# Sphinx-2.0 style
pub = Publisher(reader=reader,
parser=parser,
writer=SphinxDummyWriter(),
source_class=SphinxFileInput,
destination=NullOutput())
pub.process_programmatic_settings(None, env.settings, None)
pub.set_source(source_path=filename)
pub.publish()
return pub.document
def publish_cmdline(reader=None,
reader_name='standalone',
parser=None,
parser_name='restructuredtext',
writer=None,
writer_name='pseudoxml',
settings=None,
settings_spec=None,
settings_overrides=None,
config_section=None,
enable_exit_status=1,
argv=None,
usage=default_usage,
description=default_description):
"""
See docutils.core.publish_cmdline.
We just modified this function to return the parsed destination file.
"""
pub = Publisher(reader, parser, writer, settings=settings)
pub.set_components(reader_name, parser_name, writer_name)
output = pub.publish(argv,
usage,
description,
settings_spec,
settings_overrides,
config_section=config_section,
enable_exit_status=enable_exit_status)
return output, pub.settings._source, pub.settings._destination
def __init__(self,
settings_spec=None,
settings_overrides=dict(report_level=5, halt_level=5),
config_section='general'):
self.pub = Publisher(reader=None,
parser=None,
writer=None,
settings=None,
source_class=io.StringInput,
destination_class=io.StringOutput)
self.pub.set_components(reader_name='standalone',
parser_name='restructuredtext',
writer_name='html')
# hack: JEP-0071 does not allow HTML char entities, so we hack our way
# out of it.
# — == u"\u2014"
# a setting to only emit charater entities in the writer would be nice
# FIXME: several are emitted, and they are explicitly forbidden
# in the JEP
# == u"\u00a0"
self.pub.writer.translator_class.attribution_formats['dash'] = (
'\u2014', '')
self.pub.process_programmatic_settings(settings_spec,
settings_overrides,
config_section)
def publish_cmdline_to_binary(reader=None, reader_name='standalone',
parser=None, parser_name='restructuredtext',
writer=None, writer_name='pseudoxml',
settings=None, settings_spec=None,
settings_overrides=None, config_section=None,
enable_exit_status=1, argv=None,
usage=default_usage, description=default_description,
destination=None, destination_class=BinaryFileOutput
):
"""
Set up & run a `Publisher` for command-line-based file I/O (input and
output file paths taken automatically from the command line). Return the
encoded string output also.
This is just like publish_cmdline, except that it uses
io.BinaryFileOutput instead of io.FileOutput.
Parameters: see `publish_programmatically` for the remainder.
- `argv`: Command-line argument list to use instead of ``sys.argv[1:]``.
- `usage`: Usage string, output if there's a problem parsing the command
line.
- `description`: Program description, output for the "--help" option
(along with command-line option descriptions).
"""
pub = Publisher(reader, parser, writer, settings=settings,
destination_class=destination_class)
pub.set_components(reader_name, parser_name, writer_name)
output = pub.publish(
argv, usage, description, settings_spec, settings_overrides,
config_section=config_section, enable_exit_status=enable_exit_status)
return output
def publish(writer_name):
reader=None
reader_name='standalone'
parser=None
parser_name='restructuredtext'
writer=None
settings=None
settings_spec=None
settings_overrides=options[writer_name]
config_section=None
enable_exit=1
argv=[]
usage=default_usage
pub = Publisher(reader, parser, writer, settings=settings)
pub.set_components(reader_name, parser_name, writer_name)
settings = pub.get_settings(settings_spec=settings_spec,
config_section=config_section)
if settings_overrides:
settings._update(settings_overrides, 'loose')
source = file(source_path)
pub.set_source(source, source_path)
destination_path = 'pg1.' + extensions[writer_name]
destination = file(destination_path, 'w')
pub.set_destination(destination, destination_path)
pub.publish(argv, usage, description, settings_spec, settings_overrides,
config_section=config_section, enable_exit=enable_exit)
def _docutils_rest_to(rest, writer):
""" Uses docutils to convert a ReST string to HTML. Returns a tuple
containg the HTML string and the list of warning nodes that were
removed from the HTML.
"""
# Make sure any Sphinx polution of docutils has been removed.
if Sphinx is not None:
for key, value in docutils_roles.items():
if value.__module__.startswith('sphinx'):
docutils_roles.pop(key)
pub = Publisher(source_class=docutils.io.StringInput,
destination_class=docutils.io.StringOutput)
pub.set_reader('standalone', None, 'restructuredtext')
pub.set_writer(writer)
pub.writer.default_stylesheet_path=''
pub.get_settings() # Get the default settings
pub.settings.halt_level = 6 # Don't halt on errors
pub.settings.warning_stream = StringIO()
pub.set_source(rest)
pub.set_destination()
pub.document = pub.reader.read(pub.source, pub.parser, pub.settings)
pub.apply_transforms()
# Walk the node structure of a docutils document and remove 'problematic'
# and 'system_message' nodes. Save the system_message nodes.
warning_nodes = []
for node in pub.document.traverse(docutils.nodes.problematic):
node.parent.replace(node, node.children[0])
for node in pub.document.traverse(docutils.nodes.system_message):
warning_nodes.append(node)
node.parent.remove(node)
return pub.writer.write(pub.document, pub.destination), warning_nodes
def convertWWW(src, language, options=None):
if language == 'el':
encoding = 'iso-8859-7'
elif language == 'pl':
encoding = 'iso-8859-2'
elif language == 'ru':
encoding = 'windows-1251'
elif language == 'cs':
encoding = 'iso-8859-2'
else:
encoding = 'iso-8859-15'
arguments = [
'--no-generator', '--language=' + language, '--no-source-link',
'--no-datestamp', '--input-encoding=' + encoding,
'--output-encoding=' + encoding, '--target-suffix=' + 'php', src, ''
]
if options:
for option in options:
arguments.insert(0, option)
publisher = Publisher(destination_class=NullOutput)
publisher.set_reader('standalone', None, 'restructuredtext')
publisher.set_writer('html')
publisher.publish(argv=arguments)
return ''.join(publisher.writer.body_pre_docinfo +
publisher.writer.body).encode(encoding)
def processHTML(src, depth):
src = os.path.normpath(src)
prefix = os.path.splitext(src)[0]
suffix = os.path.splitext(src)[1][1:]
if suffix != DEFAULTLANG:
return
dst = prefix + '.html' #.' + suffix
dst_abs = os.path.normpath(os.path.join(TRGROOT, dst))
src_abs = os.path.normpath(os.path.join(SRCROOT, src))
dst_dir = os.path.dirname(dst_abs)
makedir(dst_dir)
if newer([src_abs], dst_abs):
reportBuilding(src)
arguments = [
'--no-generator', '--language=' + suffix, '--no-source-link',
'--no-datestamp', '--output-encoding=iso-8859-15',
'--target-suffix=html',
'--stylesheet=' + '../' * depth + 'aros.css', '--link-stylesheet',
src_abs, dst_abs
]
publisher = Publisher()
publisher.set_reader('standalone', None, 'restructuredtext')
publisher.set_writer('html')
publisher.publish(argv=arguments)
else:
reportSkipping(dst)
def read_doc(app, env, filename):
# type: (Sphinx, BuildEnvironment, unicode) -> nodes.document
"""Parse a document and convert to doctree."""
filetype = get_filetype(app.config.source_suffix, filename)
input_class = app.registry.get_source_input(filetype)
reader = SphinxStandaloneReader(app)
source = input_class(app, env, source=None, source_path=filename,
encoding=env.config.source_encoding)
parser = app.registry.create_source_parser(app, filetype)
if parser.__class__.__name__ == 'CommonMarkParser' and parser.settings_spec == ():
# a workaround for recommonmark
# If recommonmark.AutoStrictify is enabled, the parser invokes reST parser
# internally. But recommonmark-0.4.0 does not provide settings_spec for reST
# parser. As a workaround, this copies settings_spec for RSTParser to the
# CommonMarkParser.
parser.settings_spec = RSTParser.settings_spec
pub = Publisher(reader=reader,
parser=parser,
writer=SphinxDummyWriter(),
source_class=SphinxDummySourceClass,
destination=NullOutput())
pub.set_components(None, 'restructuredtext', None)
pub.process_programmatic_settings(None, env.settings, None)
pub.set_source(source, filename)
pub.publish()
return pub.document
def parse_docstring(doc):
p = Publisher(source=doc, source_class=io.StringInput)
p.set_reader('standalone', p.parser, 'restructuredtext')
p.writer = Writer()
p.process_programmatic_settings(None, None, None)
p.set_source(doc, None)
return p.publish()
def main(prog_args):
argv = None
reader = standalone.Reader()
reader_name = 'standalone'
writer = Writer()
writer_name = 'pseudoxml'
parser = None
parser_name = 'restructuredtext'
settings = None
settings_spec = None
settings_overrides = None
config_section = None
enable_exit_status = 1
usage = default_usage
publisher = Publisher(reader,
parser,
writer,
settings,
destination_class=BinaryFileOutput)
publisher.set_components(reader_name, parser_name, writer_name)
description = ('Generates OpenDocument/OpenOffice/ODF slides from '
'standalone reStructuredText sources. ' +
default_description)
output = publisher.publish(argv,
usage,
description,
settings_spec,
settings_overrides,
config_section=config_section,
enable_exit_status=enable_exit_status)
def main(args):
argv = None
reader = standalone.Reader()
reader_name = 'standalone'
writer = EpubWriter()
writer_name = 'epub2'
parser = Parser()
parser_name = 'restructuredtext'
settings = None
settings_spec = None
settings_overrides = None
config_section = None
enable_exit_status = 1
usage = default_usage
publisher = Publisher(reader,
parser,
writer,
settings,
destination_class=EpubFileOutput)
publisher.set_components(reader_name, parser_name, writer_name)
description = ('Generates epub books from reStructuredText sources. ' +
default_description)
output = publisher.publish(argv,
usage,
description,
settings_spec,
settings_overrides,
config_section=config_section,
enable_exit_status=enable_exit_status)
def get_sphinx():
sphinx = getattr(local_data, 'sphinx', None)
if sphinx is None:
sphinx = Sphinx(tempdir,
tempdir,
tempdir,
tempdir,
'json',
status=None,
warning=None)
sphinx.builder.translator_class = CustomHTMLTranslator
sphinx.env.patch_lookup_functions()
sphinx.env.temp_data['docname'] = 'text'
sphinx.env.temp_data['default_domain'] = 'py'
pub = Publisher(reader=None,
parser=None,
writer=HTMLWriter(sphinx.builder),
source_class=io.StringInput,
destination_class=io.NullOutput)
pub.set_components('standalone', 'restructuredtext', None)
pub.process_programmatic_settings(None, sphinx.env.settings, None)
pub.set_destination(None, None)
sphinx.publisher = pub
local_data.sphinx = sphinx
return sphinx, sphinx.publisher
def render_partial(self, node):
"""Utility: Render a lone doctree node."""
if node is None:
return {'fragment': ''}
doc = new_document(b'<partial node>')
doc.append(node)
if self._publisher is None:
self._publisher = Publisher(
source_class=DocTreeInput,
destination_class=StringOutput)
self._publisher.set_components('standalone',
'restructuredtext', 'pseudoxml')
pub = self._publisher
pub.reader = DoctreeReader()
pub.writer = MarkdownWriter(self)
pub.process_programmatic_settings(
None, {}, None)
#pub.process_programmatic_settings(
# None, {'output_encoding': 'unicode'}, None)
pub.set_source(doc, None)
pub.set_destination(None, None)
pub.publish()
return pub.writer.parts
def render(self):
key = 'cia.apps.blog.%d' % self.id
parts = cache.get(key)
if not parts:
# Convert the reST markup to a document tree
document = publish_doctree(source=self.content)
visitor = ImageTranslator(document)
document.walkabout(visitor)
#
# Publish that document tree as HTML. We can't use any of
# the simpler methods in docutils.core, since we need
# access to writer.parts
#
reader = doctree.Reader(parser_name='null')
pub = Publisher(reader,
None,
None,
source=DocTreeInput(document),
destination_class=StringOutput)
pub.set_writer('html4css1')
pub.process_programmatic_settings(None, {
'cloak_email_addresses': True,
'initial_header_level': 2,
}, None)
pub.publish()
parts = pub.writer.parts
cache.set(key, parts)
return parts
def create_publisher(app: "Sphinx", filetype: str) -> Publisher:
reader = SphinxStandaloneReader()
reader.setup(app)
parser = app.registry.create_source_parser(app, filetype)
if parser.__class__.__name__ == 'CommonMarkParser' and parser.settings_spec == (
):
# a workaround for recommonmark
# If recommonmark.AutoStrictify is enabled, the parser invokes reST parser
# internally. But recommonmark-0.4.0 does not provide settings_spec for reST
# parser. As a workaround, this copies settings_spec for RSTParser to the
# CommonMarkParser.
from docutils.parsers.rst import Parser as RSTParser
parser.settings_spec = RSTParser.settings_spec
pub = Publisher(reader=reader,
parser=parser,
writer=SphinxDummyWriter(),
source_class=SphinxFileInput,
destination=NullOutput())
# Propagate exceptions by default when used programmatically:
defaults = {"traceback": True, **app.env.settings}
# Set default settings
if docutils.__version_info__[:2] >= (0, 19):
pub.get_settings(**defaults) # type: ignore[arg-type]
else:
pub.settings = pub.setup_option_parser(
**defaults).get_default_values() # type: ignore
return pub
def read_doc(app: "Sphinx", env: BuildEnvironment,
filename: str) -> nodes.document:
"""Parse a document and convert to doctree."""
warnings.warn('sphinx.io.read_doc() is deprecated.',
RemovedInSphinx70Warning,
stacklevel=2)
# set up error_handler for the target document
error_handler = UnicodeDecodeErrorHandler(env.docname)
codecs.register_error('sphinx', error_handler) # type: ignore
reader = SphinxStandaloneReader()
reader.setup(app)
filetype = get_filetype(app.config.source_suffix, filename)
parser = app.registry.create_source_parser(app, filetype)
if parser.__class__.__name__ == 'CommonMarkParser' and parser.settings_spec == (
):
# a workaround for recommonmark
# If recommonmark.AutoStrictify is enabled, the parser invokes reST parser
# internally. But recommonmark-0.4.0 does not provide settings_spec for reST
# parser. As a workaround, this copies settings_spec for RSTParser to the
# CommonMarkParser.
parser.settings_spec = RSTParser.settings_spec
pub = Publisher(reader=reader,
parser=parser,
writer=SphinxDummyWriter(),
source_class=SphinxFileInput,
destination=NullOutput())
pub.process_programmatic_settings(None, env.settings, None)
pub.set_source(source_path=filename)
pub.publish()
return pub.document
def render_rst(directory, name, meta_parser, template):
# check if that file actually exists
path = safe_join(directory, name + '.rst')
if not os.path.exists(path):
abort(404)
# read file
with codecs.open(path, encoding='utf-8') as fd:
content = fd.read()
if not template:
# Strip out RST
content = content.replace('.. meta::\n', '')
content = content.replace('.. contents::\n\n', '')
content = content.replace('.. raw:: html\n\n', '')
content = content.replace('\n.. [', '\n[')
content = content.replace(']_.', '].')
content = content.replace(']_,', '],')
content = content.replace(']_', '] ')
# Change highlight formatter
content = content.replace('{% highlight', "{% highlight formatter='textspec'")
# Metatags
for (metatag, label) in METATAG_LABELS.items():
content = content.replace(' :%s' % metatag, label)
# render the post with Jinja2 to handle URLs etc.
rendered_content = render_template_string(content)
rendered_content = rendered_content.replace('</pre></div>', ' </pre></div>')
if not template:
# Send response
r = make_response(rendered_content)
r.mimetype = 'text/plain'
return r
# Render the ToC
doctree = publish_doctree(source=rendered_content)
bullet_list = doctree[1][1]
doctree.clear()
doctree.append(bullet_list)
reader = Reader(parser_name='null')
pub = Publisher(reader, None, None,
source=io.DocTreeInput(doctree),
destination_class=io.StringOutput)
pub.set_writer('html')
pub.publish()
toc = pub.writer.parts['fragment']
# Remove the ToC from the main document
rendered_content = rendered_content.replace('.. contents::\n', '')
# publish the spec with docutils
parts = publish_parts(source=rendered_content, source_path=directory, writer_name="html")
meta = meta_parser(parts['meta'])
if (directory == PROPOSAL_DIR):
meta['num'] = int(name[:3])
return render_template(template, title=parts['title'], toc=toc, body=parts['fragment'], name=name, meta=meta)
def __init__(self):
self._publisher = Publisher(source_class=StringInput,
destination_class=StringOutput)
self._publisher.set_components('standalone', 'restructuredtext',
'html')
self._publisher.writer.translator_class = MyHTMLTranslator
self._publisher.process_programmatic_settings(None, extra_params, None)
def process_description(source, output_encoding='unicode'):
"""Given an source string, returns an HTML fragment as a string.
The return value is the contents of the <body> tag.
Parameters:
- `source`: A multi-line text string; required.
- `output_encoding`: The desired encoding of the output. If a Unicode
string is desired, use the default value of "unicode" .
"""
# Dedent all lines of `source`.
source = trim_docstring(source)
settings_overrides = {
'raw_enabled': 0, # no raw HTML code
'file_insertion_enabled': 0, # no file/URL access
'halt_level': 2, # at warnings or errors, raise an exception
'report_level': 5, # never report problems with the reST code
}
parts = None
# Convert reStructuredText to HTML using Docutils.
document = publish_doctree(source=source,
settings_overrides=settings_overrides)
for node in document.traverse():
if node.tagname == '#text':
continue
if node.hasattr('refuri'):
uri = node['refuri']
elif node.hasattr('uri'):
uri = node['uri']
else:
continue
o = urlparse(uri)
if o.scheme not in ALLOWED_SCHEMES:
raise TransformError('link scheme not allowed: {0}'.format(uri))
# now turn the transformed document into HTML
reader = readers.doctree.Reader(parser_name='null')
pub = Publisher(reader,
source=io.DocTreeInput(document),
destination_class=io.StringOutput)
pub.set_writer('html')
pub.process_programmatic_settings(None, settings_overrides, None)
pub.set_destination(None, None)
pub.publish()
parts = pub.writer.parts
output = parts['body']
if output_encoding != 'unicode':
output = output.encode(output_encoding)
return output
def run(self):
log.debug("[%s.%s] Compiling manual from RST sources." %
(__name__, self.__class__.__name__))
pub = Publisher(writer=manpage.Writer())
pub.set_components(reader_name='standalone',
parser_name='restructuredtext',
writer_name='pseudoxml')
pub.publish(argv=[
u'%s' % get_path([DOCDIR, 'man', 'tribus.rst']),
u'%s' % get_path([DOCDIR, 'man', 'tribus.1'])
])
def get_html(self, body_only=True, content_only=False, noclasses=False):
import sys
import pygments_rest
from docutils.core import Publisher
from docutils.io import StringInput, StringOutput
from cStringIO import StringIO
settings = {
'doctitle_xform': 1,
'pep_references': 1,
'rfc_references': 1,
'footnote_references': 'superscript',
'output_encoding': 'unicode',
'report_level':
2, # 2=show warnings, 3=show only errors, 5=off (docutils.utils
}
if content_only:
post_rst = self.get_rst(noclasses=noclasses)
else:
post_rst = render_to('post_single.rst',
post=self,
noclasses=noclasses)
pub = Publisher(reader=None,
parser=None,
writer=None,
settings=None,
source_class=StringInput,
destination_class=StringOutput)
pub.set_components(reader_name='standalone',
parser_name='restructuredtext',
writer_name='html')
pub.process_programmatic_settings(settings_spec=None,
settings_overrides=settings,
config_section=None)
pub.set_source(post_rst, source_path=self.module_path)
pub.set_destination(None, None)
errors_io = StringIO()
real_stderr = sys.stderr
sys.stderr = errors_io
try:
html_full = pub.publish(enable_exit_status=False)
html_body = ''.join(pub.writer.html_body)
finally:
sys.stderr = real_stderr
errors = errors_io.getvalue()
self._process_rest_errors(errors)
errors_io.close()
return html_body if body_only else html_full
def inspect(
filename,
source_path=None,
):
"returns the document object before any transforms)"
from docutils.core import Publisher
pub = Publisher(source_class=io.FileInput, )
pub.set_reader('standalone', None, "restructuredtext")
pub.process_programmatic_settings(None, None, None)
pub.set_source(source_path=source_path)
pub.set_io()
return pub.reader.read(pub.source, pub.parser, pub.settings)
def convert(infilename, outfilename):
print "converting %s to %s" % (infilename, outfilename)
pub = Publisher()
pub.set_components(
'standalone', # reader
'restructuredtext', # parser
'latex') # writer (arg, will be discarded)
pub.reader = OptikReader()
pub.writer = PyLaTeXWriter()
pub.process_programmatic_settings(None, None, None)
pub.set_source(source_path=infilename)
pub.set_destination(destination_path=outfilename)
pub.publish()
def process_labels(site, logger, source, post):
site.processing_labels = True
pub = Publisher(reader=Reader(), parser=None, writer=None)
pub.set_components(None, 'restructuredtext', 'html')
# Reading the file will generate output/errors that we don't care about
# at this stage. The report_level = 5 means no output
pub.process_programmatic_settings(
settings_spec=None,
settings_overrides={'report_level': 5},
config_section=None,
)
pub.set_source(None, source)
pub.publish()
document = pub.document
site.processing_labels = False
# Code based on Sphinx std domain
for name, is_explicit in document.nametypes.items():
if not is_explicit:
continue
labelid = document.nameids[name]
if labelid is None:
continue
node = document.ids[labelid]
if node.tagname == 'target' and 'refid' in node:
node = document.ids.get(node['refid'])
labelid = node['names'][0]
if node.tagname == 'footnote' or 'refuri' in node or node.tagname.startswith(
'desc_'):
continue
if name in site.ref_labels:
logger.warn(
'Duplicate label {dup}, other instance in {other}'.format(
dup=name, other=site.ref_labels[name][0]))
site.anon_ref_labels[name] = post.permalink(), labelid
def clean_astext(node):
"""Like node.astext(), but ignore images.
Taken from sphinx.util.nodes"""
node = node.deepcopy()
for img in node.traverse(nodes.image):
img['alt'] = ''
for raw in node.traverse(nodes.raw):
raw.parent.remove(raw)
return node.astext()
if node.tagname in ('section', 'rubric'):
sectname = clean_astext(node[0])
else:
continue
site.ref_labels[name] = post.permalink(), labelid, sectname
def render_readme_like_pypi(source, output_encoding='unicode'):
"""
Render a ReST document just like PyPI does.
"""
# Dedent all lines of `source`.
source = trim_docstring(source)
settings_overrides = {
'raw_enabled': 0, # no raw HTML code
'file_insertion_enabled': 0, # no file/URL access
'halt_level': 2, # at warnings or errors, raise an exception
'report_level': 5, # never report problems with the reST code
}
parts = None
# Convert reStructuredText to HTML using Docutils.
document = publish_doctree(source=source,
settings_overrides=settings_overrides)
for node in document.traverse():
if node.tagname == '#text':
continue
if node.hasattr('refuri'):
uri = node['refuri']
elif node.hasattr('uri'):
uri = node['uri']
else:
continue
o = urlparse(uri)
if o.scheme not in ALLOWED_SCHEMES:
raise TransformError('link scheme not allowed: {0}'.format(uri))
# now turn the transformed document into HTML
reader = readers.doctree.Reader(parser_name='null')
pub = Publisher(reader,
source=io.DocTreeInput(document),
destination_class=io.StringOutput)
pub.set_writer('html')
pub.process_programmatic_settings(None, settings_overrides, None)
pub.set_destination(None, None)
pub.publish()
parts = pub.writer.parts
output = parts['body']
if output_encoding != 'unicode':
output = output.encode(output_encoding)
return output
def get_doctree(path):
"""
Obtain a Sphinx doctree from the RST file at ``path``.
Performs no Releases-specific processing; this code would, ideally, be in
Sphinx itself, but things there are pretty tightly coupled. So we wrote
this.
:param str path: A relative or absolute file path string.
:returns:
A two-tuple of the generated ``sphinx.application.Sphinx`` app and the
doctree (a ``docutils.document`` object).
"""
root, filename = os.path.split(path)
docname, _ = os.path.splitext(filename)
# TODO: this only works for top level changelog files (i.e. ones where
# their dirname is the project/doc root)
app = make_app(srcdir=root)
# Create & init a BuildEnvironment. Mm, tasty side effects.
app._init_env(freshenv=True)
env = app.env
env.update(config=app.config, srcdir=root, doctreedir=app.doctreedir, app=app)
# Code taken from sphinx.environment.read_doc; easier to manually call
# it with a working Environment object, instead of doing more random crap
# to trick the higher up build system into thinking our single changelog
# document was "updated".
env.temp_data['docname'] = docname
env.app = app
# NOTE: SphinxStandaloneReader API changed in 1.4 :(
reader_kwargs = {'app': env.app, 'parsers': env.config.source_parsers}
if sphinx.version_info[:2] < (1, 4):
del reader_kwargs['app']
reader = SphinxStandaloneReader(**reader_kwargs)
pub = Publisher(reader=reader,
writer=SphinxDummyWriter(),
destination_class=NullOutput)
pub.set_components(None, 'restructuredtext', None)
pub.process_programmatic_settings(None, env.settings, None)
# NOTE: docname derived higher up, from our given path
src_path = env.doc2path(docname)
source = SphinxFileInput(app, env, source=None, source_path=src_path,
encoding=env.config.source_encoding)
pub.source = source
pub.settings._source = src_path
pub.set_destination(None, None)
pub.publish()
return app, pub.document
def _get_publisher(self, source_path):
extra_params = {
'initial_header_level': '2',
'syntax_highlight': 'short',
'input_encoding': 'utf-8'
}
user_params = self.settings.get('DOCUTILS_SETTINGS')
if user_params:
extra_params.update(user_params)
pub = Publisher(destination_class=StringOutput)
pub.set_components('standalone', 'restructuredtext', 'html')
pub.writer.translator_class = HTMLTranslator
pub.process_programmatic_settings(None, extra_params, None)
pub.set_source(source_path=source_path)
pub.publish()
return pub
def pypi_rest2html(source, output_encoding="unicode"):
"""
>>> pypi_rest2html("test!")
u'<p>test!</p>\n'
"""
settings_overrides = {
"raw_enabled": 0, # no raw HTML code
"file_insertion_enabled": 0, # no file/URL access
"halt_level": 2, # at warnings or errors, raise an exception
"report_level": 5, # never report problems with the reST code
}
# Convert reStructuredText to HTML using Docutils.
document = publish_doctree(source=source,
settings_overrides=settings_overrides)
for node in document.traverse():
if node.tagname == "#text":
continue
if node.hasattr("refuri"):
uri = node["refuri"]
elif node.hasattr("uri"):
uri = node["uri"]
else:
continue
o = urlparse(uri)
if o.scheme not in ALLOWED_SCHEMES:
raise TransformError("link scheme not allowed")
# now turn the transformed document into HTML
reader = readers.doctree.Reader(parser_name="null")
pub = Publisher(reader,
source=io.DocTreeInput(document),
destination_class=io.StringOutput)
pub.set_writer("html")
pub.process_programmatic_settings(None, settings_overrides, None)
pub.set_destination(None, None)
pub.publish()
parts = pub.writer.parts
output = parts["body"]
if output_encoding != "unicode":
output = output.encode(output_encoding)
return output
def publish_doctree(source, logger):
# type: (str, Any) -> Node
pub = Publisher(reader=None,
parser=None,
writer=None,
settings=None,
source_class=io.StringInput,
destination_class=io.NullOutput)
pub.reader = Reader(None, 'restructuredtext')
pub.reader.doc_logger = logger
pub.set_writer('null')
pub.parser = pub.reader.parser
pub.process_programmatic_settings(None, None, None)
pub.set_source(source, None)
pub.set_destination(None, None)
output = pub.publish(enable_exit_status=False)
return pub.document
