blob: 63afc7556850a7c531a3269ed1496505fb5bd400 [file] [log] [blame]
# $Id: misc.py 6185 2009-10-25 20:43:42Z milde $
# Authors: David Goodger <goodger@python.org>; Dethe Elza
# Copyright: This module has been placed in the public domain.
"""Miscellaneous directives."""
__docformat__ = 'reStructuredText'
import sys
import os.path
import re
import time
from docutils import io, nodes, statemachine, utils
from docutils.parsers.rst import Directive, convert_directive_function
from docutils.parsers.rst import directives, roles, states
from docutils.transforms import misc
class Include(Directive):
"""
Include content read from a separate source file.
Content may be parsed by the parser, or included as a literal
block. The encoding of the included file can be specified. Only
a part of the given file argument may be included by specifying
start and end line or text to match before and/or after the text
to be used.
"""
required_arguments = 1
optional_arguments = 0
final_argument_whitespace = True
option_spec = {'literal': directives.flag,
'encoding': directives.encoding,
'tab-width': int,
'start-line': int,
'end-line': int,
'start-after': directives.unchanged_required,
'end-before': directives.unchanged_required}
standard_include_path = os.path.join(os.path.dirname(states.__file__),
'include')
def run(self):
"""Include a reST file as part of the content of this reST file."""
if not self.state.document.settings.file_insertion_enabled:
raise self.warning('"%s" directive disabled.' % self.name)
source = self.state_machine.input_lines.source(
self.lineno - self.state_machine.input_offset - 1)
source_dir = os.path.dirname(os.path.abspath(source))
path = directives.path(self.arguments[0])
if path.startswith('<') and path.endswith('>'):
path = os.path.join(self.standard_include_path, path[1:-1])
path = os.path.normpath(os.path.join(source_dir, path))
path = utils.relative_path(None, path)
path = nodes.reprunicode(path)
encoding = self.options.get(
'encoding', self.state.document.settings.input_encoding)
tab_width = self.options.get(
'tab-width', self.state.document.settings.tab_width)
try:
self.state.document.settings.record_dependencies.add(path)
include_file = io.FileInput(
source_path=path, encoding=encoding,
error_handler=(self.state.document.settings.\
input_encoding_error_handler),
handle_io_errors=None)
except IOError, error:
raise self.severe('Problems with "%s" directive path:\n%s: %s.' %
(self.name, error.__class__.__name__, str(error)))
# Hack: Since Python 2.6, the string interpolation returns a
# unicode object if one of the supplied %s replacements is a
# unicode object. IOError has no `__unicode__` method and the
# fallback `__repr__` does not report the file name. Explicitely
# converting to str fixes this for now::
# print '%s\n%s\n%s\n' %(error, str(error), repr(error))
startline = self.options.get('start-line', None)
endline = self.options.get('end-line', None)
try:
if startline or (endline is not None):
lines = include_file.readlines()
rawtext = ''.join(lines[startline:endline])
else:
rawtext = include_file.read()
except UnicodeError, error:
raise self.severe(
'Problem with "%s" directive:\n%s: %s'
% (self.name, error.__class__.__name__, error))
# start-after/end-before: no restrictions on newlines in match-text,
# and no restrictions on matching inside lines vs. line boundaries
after_text = self.options.get('start-after', None)
if after_text:
# skip content in rawtext before *and incl.* a matching text
after_index = rawtext.find(after_text)
if after_index < 0:
raise self.severe('Problem with "start-after" option of "%s" '
'directive:\nText not found.' % self.name)
rawtext = rawtext[after_index + len(after_text):]
before_text = self.options.get('end-before', None)
if before_text:
# skip content in rawtext after *and incl.* a matching text
before_index = rawtext.find(before_text)
if before_index < 0:
raise self.severe('Problem with "end-before" option of "%s" '
'directive:\nText not found.' % self.name)
rawtext = rawtext[:before_index]
if 'literal' in self.options:
# Convert tabs to spaces, if `tab_width` is positive.
if tab_width >= 0:
text = rawtext.expandtabs(tab_width)
else:
text = rawtext
literal_block = nodes.literal_block(rawtext, text, source=path)
literal_block.line = 1
return [literal_block]
else:
include_lines = statemachine.string2lines(
rawtext, tab_width, convert_whitespace=1)
self.state_machine.insert_input(include_lines, path)
return []
class Raw(Directive):
"""
Pass through content unchanged
Content is included in output based on type argument
Content may be included inline (content section of directive) or
imported from a file or url.
"""
required_arguments = 1
optional_arguments = 0
final_argument_whitespace = True
option_spec = {'file': directives.path,
'url': directives.uri,
'encoding': directives.encoding}
has_content = True
def run(self):
if (not self.state.document.settings.raw_enabled
or (not self.state.document.settings.file_insertion_enabled
and ('file' in self.options
or 'url' in self.options))):
raise self.warning('"%s" directive disabled.' % self.name)
attributes = {'format': ' '.join(self.arguments[0].lower().split())}
encoding = self.options.get(
'encoding', self.state.document.settings.input_encoding)
if self.content:
if 'file' in self.options or 'url' in self.options:
raise self.error(
'"%s" directive may not both specify an external file '
'and have content.' % self.name)
text = '\n'.join(self.content)
elif 'file' in self.options:
if 'url' in self.options:
raise self.error(
'The "file" and "url" options may not be simultaneously '
'specified for the "%s" directive.' % self.name)
source_dir = os.path.dirname(
os.path.abspath(self.state.document.current_source))
path = os.path.normpath(os.path.join(source_dir,
self.options['file']))
path = utils.relative_path(None, path)
try:
self.state.document.settings.record_dependencies.add(path)
raw_file = io.FileInput(
source_path=path, encoding=encoding,
error_handler=(self.state.document.settings.\
input_encoding_error_handler),
handle_io_errors=None)
except IOError, error:
raise self.severe('Problems with "%s" directive path:\n%s.'
% (self.name, error))
try:
text = raw_file.read()
except UnicodeError, error:
raise self.severe(
'Problem with "%s" directive:\n%s: %s'
% (self.name, error.__class__.__name__, error))
attributes['source'] = path
elif 'url' in self.options:
source = self.options['url']
# Do not import urllib2 at the top of the module because
# it may fail due to broken SSL dependencies, and it takes
# about 0.15 seconds to load.
import urllib2
try:
raw_text = urllib2.urlopen(source).read()
except (urllib2.URLError, IOError, OSError), error:
raise self.severe(
'Problems with "%s" directive URL "%s":\n%s.'
% (self.name, self.options['url'], error))
raw_file = io.StringInput(
source=raw_text, source_path=source, encoding=encoding,
error_handler=(self.state.document.settings.\
input_encoding_error_handler))
try:
text = raw_file.read()
except UnicodeError, error:
raise self.severe(
'Problem with "%s" directive:\n%s: %s'
% (self.name, error.__class__.__name__, error))
attributes['source'] = source
else:
# This will always fail because there is no content.
self.assert_has_content()
raw_node = nodes.raw('', text, **attributes)
return [raw_node]
class Replace(Directive):
has_content = True
def run(self):
if not isinstance(self.state, states.SubstitutionDef):
raise self.error(
'Invalid context: the "%s" directive can only be used within '
'a substitution definition.' % self.name)
self.assert_has_content()
text = '\n'.join(self.content)
element = nodes.Element(text)
self.state.nested_parse(self.content, self.content_offset,
element)
if ( len(element) != 1
or not isinstance(element[0], nodes.paragraph)):
messages = []
for node in element:
if isinstance(node, nodes.system_message):
node['backrefs'] = []
messages.append(node)
error = self.state_machine.reporter.error(
'Error in "%s" directive: may contain a single paragraph '
'only.' % (self.name), line=self.lineno)
messages.append(error)
return messages
else:
return element[0].children
class Unicode(Directive):
r"""
Convert Unicode character codes (numbers) to characters. Codes may be
decimal numbers, hexadecimal numbers (prefixed by ``0x``, ``x``, ``\x``,
``U+``, ``u``, or ``\u``; e.g. ``U+262E``), or XML-style numeric character
entities (e.g. ``&#x262E;``). Text following ".." is a comment and is
ignored. Spaces are ignored, and any other text remains as-is.
"""
required_arguments = 1
optional_arguments = 0
final_argument_whitespace = True
option_spec = {'trim': directives.flag,
'ltrim': directives.flag,
'rtrim': directives.flag}
comment_pattern = re.compile(r'( |\n|^)\.\. ')
def run(self):
if not isinstance(self.state, states.SubstitutionDef):
raise self.error(
'Invalid context: the "%s" directive can only be used within '
'a substitution definition.' % self.name)
substitution_definition = self.state_machine.node
if 'trim' in self.options:
substitution_definition.attributes['ltrim'] = 1
substitution_definition.attributes['rtrim'] = 1
if 'ltrim' in self.options:
substitution_definition.attributes['ltrim'] = 1
if 'rtrim' in self.options:
substitution_definition.attributes['rtrim'] = 1
codes = self.comment_pattern.split(self.arguments[0])[0].split()
element = nodes.Element()
for code in codes:
try:
decoded = directives.unicode_code(code)
except ValueError, err:
raise self.error(
'Invalid character code: %s\n%s: %s'
% (code, err.__class__.__name__, err))
element += nodes.Text(decoded)
return element.children
class Class(Directive):
"""
Set a "class" attribute on the directive content or the next element.
When applied to the next element, a "pending" element is inserted, and a
transform does the work later.
"""
required_arguments = 1
optional_arguments = 0
final_argument_whitespace = True
has_content = True
def run(self):
try:
class_value = directives.class_option(self.arguments[0])
except ValueError:
raise self.error(
'Invalid class attribute value for "%s" directive: "%s".'
% (self.name, self.arguments[0]))
node_list = []
if self.content:
container = nodes.Element()
self.state.nested_parse(self.content, self.content_offset,
container)
for node in container:
node['classes'].extend(class_value)
node_list.extend(container.children)
else:
pending = nodes.pending(
misc.ClassAttribute,
{'class': class_value, 'directive': self.name},
self.block_text)
self.state_machine.document.note_pending(pending)
node_list.append(pending)
return node_list
class Role(Directive):
has_content = True
argument_pattern = re.compile(r'(%s)\s*(\(\s*(%s)\s*\)\s*)?$'
% ((states.Inliner.simplename,) * 2))
def run(self):
"""Dynamically create and register a custom interpreted text role."""
if self.content_offset > self.lineno or not self.content:
raise self.error('"%s" directive requires arguments on the first '
'line.' % self.name)
args = self.content[0]
match = self.argument_pattern.match(args)
if not match:
raise self.error('"%s" directive arguments not valid role names: '
'"%s".' % (self.name, args))
new_role_name = match.group(1)
base_role_name = match.group(3)
messages = []
if base_role_name:
base_role, messages = roles.role(
base_role_name, self.state_machine.language, self.lineno,
self.state.reporter)
if base_role is None:
error = self.state.reporter.error(
'Unknown interpreted text role "%s".' % base_role_name,
nodes.literal_block(self.block_text, self.block_text),
line=self.lineno)
return messages + [error]
else:
base_role = roles.generic_custom_role
assert not hasattr(base_role, 'arguments'), (
'Supplemental directive arguments for "%s" directive not '
'supported (specified by "%r" role).' % (self.name, base_role))
try:
converted_role = convert_directive_function(base_role)
(arguments, options, content, content_offset) = (
self.state.parse_directive_block(
self.content[1:], self.content_offset, converted_role,
option_presets={}))
except states.MarkupError, detail:
error = self.state_machine.reporter.error(
'Error in "%s" directive:\n%s.' % (self.name, detail),
nodes.literal_block(self.block_text, self.block_text),
line=self.lineno)
return messages + [error]
if 'class' not in options:
try:
options['class'] = directives.class_option(new_role_name)
except ValueError, detail:
error = self.state_machine.reporter.error(
'Invalid argument for "%s" directive:\n%s.'
% (self.name, detail), nodes.literal_block(
self.block_text, self.block_text), line=self.lineno)
return messages + [error]
role = roles.CustomRole(new_role_name, base_role, options, content)
roles.register_local_role(new_role_name, role)
return messages
class DefaultRole(Directive):
"""Set the default interpreted text role."""
required_arguments = 0
optional_arguments = 1
final_argument_whitespace = False
def run(self):
if not self.arguments:
if '' in roles._roles:
# restore the "default" default role
del roles._roles['']
return []
role_name = self.arguments[0]
role, messages = roles.role(role_name, self.state_machine.language,
self.lineno, self.state.reporter)
if role is None:
error = self.state.reporter.error(
'Unknown interpreted text role "%s".' % role_name,
nodes.literal_block(self.block_text, self.block_text),
line=self.lineno)
return messages + [error]
roles._roles[''] = role
# @@@ should this be local to the document, not the parser?
return messages
class Title(Directive):
required_arguments = 1
optional_arguments = 0
final_argument_whitespace = True
def run(self):
self.state_machine.document['title'] = self.arguments[0]
return []
class Date(Directive):
has_content = True
def run(self):
if not isinstance(self.state, states.SubstitutionDef):
raise self.error(
'Invalid context: the "%s" directive can only be used within '
'a substitution definition.' % self.name)
format = '\n'.join(self.content) or '%Y-%m-%d'
text = time.strftime(format)
return [nodes.Text(text)]
class TestDirective(Directive):
"""This directive is useful only for testing purposes."""
required_arguments = 0
optional_arguments = 1
final_argument_whitespace = True
option_spec = {'option': directives.unchanged_required}
has_content = True
def run(self):
if self.content:
text = '\n'.join(self.content)
info = self.state_machine.reporter.info(
'Directive processed. Type="%s", arguments=%r, options=%r, '
'content:' % (self.name, self.arguments, self.options),
nodes.literal_block(text, text), line=self.lineno)
else:
info = self.state_machine.reporter.info(
'Directive processed. Type="%s", arguments=%r, options=%r, '
'content: None' % (self.name, self.arguments, self.options),
line=self.lineno)
return [info]
# Old-style, functional definition:
#
# def directive_test_function(name, arguments, options, content, lineno,
# content_offset, block_text, state, state_machine):
# """This directive is useful only for testing purposes."""
# if content:
# text = '\n'.join(content)
# info = state_machine.reporter.info(
# 'Directive processed. Type="%s", arguments=%r, options=%r, '
# 'content:' % (name, arguments, options),
# nodes.literal_block(text, text), line=lineno)
# else:
# info = state_machine.reporter.info(
# 'Directive processed. Type="%s", arguments=%r, options=%r, '
# 'content: None' % (name, arguments, options), line=lineno)
# return [info]
#
# directive_test_function.arguments = (0, 1, 1)
# directive_test_function.options = {'option': directives.unchanged_required}
# directive_test_function.content = 1