rst3.py :  » Development » Leo » Leo-4.7.1-final » leo » plugins » Python Open Source

Home
Python Open Source
1.3.1.2 Python
2.Ajax
3.Aspect Oriented
4.Blog
5.Build
6.Business Application
7.Chart Report
8.Content Management Systems
9.Cryptographic
10.Database
11.Development
12.Editor
13.Email
14.ERP
15.Game 2D 3D
16.GIS
17.GUI
18.IDE
19.Installer
20.IRC
21.Issue Tracker
22.Language Interface
23.Log
24.Math
25.Media Sound Audio
26.Mobile
27.Network
28.Parser
29.PDF
30.Project Management
31.RSS
32.Search
33.Security
34.Template Engines
35.Test
36.UML
37.USB Serial
38.Web Frameworks
39.Web Server
40.Web Services
41.Web Unit
42.Wiki
43.Windows
44.XML
Python Open Source » Development » Leo 
Leo » Leo 4.7.1 final » leo » plugins » rst3.py
#@+leo-ver=4-thin
#@+node:ekr.20050805162550:@thin rst3.py
#@<< docstring >>
#@+node:ekr.20050805162550.1:<< docstring >>
'''The rst3 plugin creates output files from LeooutlinescontainingrST import 
(reStructuredText) markup.

This is a powerful plugin with many options.  The full documentation is at:
http://webpages.charter.net/edreamleo/rstplugin3.html

The rst3 plugin adds the Write Restructured Text command to Leo's Edit menu. This
command searches the selected outline looking for **rst root nodes** whose
headline have the form @rst <filename>. The plugin then creates the named file
in various ways depending which rst3 options are in effect.

By default, the rst3 plugin creates rST headings automatically from outlinesso import 
the higher-level nodes in the outline correspond to higher-level sections in the
output. Creating rST headings automatically eliminates one of the most tedious
chores associated with rST markup.

To use this plugin effectively, Python's docutils_ module must be installed. The
rst3 plugin will use the SilverCity_ syntax coloring package if it installed.

This plugin sends .htm, .html or .tex files to the docutils module for further
processing. Docutils generates HTML files or LaTeX files depending on the file's
extension. HTML files generated by docutils refer to three .css (cascading style
sheet) files that should exist in the same directory as the generated HTML file.
You can control the formatting of the HTML file by altering these .css files.

rst3 options control most aspects of this plugin's operations. You can set
options in @settings trees, in headlines and in body text. There are too many
details to discuss here. For full details see:
http://webpages.charter.net/edreamleo/rstplugin3.html
'''
#@nonl
#@-node:ekr.20050805162550.1:<< docstring >>
#@nl

# Original rst code by Josef Dalcolmo:
# contributed under the same licensed as Leo.py itself.

# rst3.py based on rst2.py v2.4.

from __future__ import generators# To make this plugin work with Python 2.2.

bwm_file = None
if 0:
    bwm_file = open("bwm_file", "w")

__version__ = '1.23'

#@<< imports >>
#@+node:ekr.20050805162550.2:<< imports >>
import leo.core.leoGlobals as g
import leo.core.leoPlugins as leoPlugins
import leo.core.leoTest as leoTest

if g.isPython3:
    import html.parser as HTMLParser
else:
    import HTMLParser

import os
import pprint

if g.isPython3:
    import io
    StringIO = io.StringIO
else:
    import StringIO
import sys

# Make sure the present directory in in sys.path.
dir,junk = os.path.split(__file__)
if dir not in sys.path: sys.path.append(dir)

try:
    import mod_http
except ImportError:
    mod_http = None

try:
    import docutils
    import docutils.parsers.rst
    import docutils.core
    import docutils.io
except ImportError:
    # This message given in init.
    # g.pr('rst3 plugin: can not import docutils')
    docutils = None

try:
    import SilverCity
except ImportError:
    if '--silent' not in sys.argv and not g.unitTesting and not g.app.batchMode:
        g.pr('rst3 plugin: SilverCity not loaded')
    SilverCity = None
#@-node:ekr.20050805162550.2:<< imports >>
#@nl
#@<< change log >>
#@+node:ekr.20050805162550.3:<< change log >>
#@@nocolor
#@@color
#@+at
# 
# 1.14 EKR: processTopTree ignores @rst- nodes so rst3 button works properly 
# when such nodes are selected.
# 1.15 EKR: added support for @rst-doc-only and doc_only_mode.  An important 
# new feature.
# 1.16 EKR: fixed bug that ignored whatever followed @space or @doc in 
# doc-only mode.
# 1.17 EKR:
# - Support show_headlines in doc-only mode.
# - .txt files are now written to default_path directory, just like special 
# files.
# 1.18 BWM: Added support for mod_scripting plugin.
# 1.19 EKR: Fixed crash that happens when invoked from menu.
# 1.20 EKR: Registers the write-restructured-text command.
# 1.21 EKR: Added rst3-publish-argv-for-missing-stylesheets setting.
# 1.22 EKR: Fixed bug that caused the plugin not to find default.css.
# 1.23 EKR: Use g.makeAllNonExistentDirectories instead of os.mkdir.
#@-at
#@nonl
#@-node:ekr.20050805162550.3:<< change log >>
#@nl
#@<< to do >>
#@+node:ekr.20050806162146:<< to do >>
#@@nocolor
#@+at
# 
# - Specify option for the spelling of special doc parts.
#     - The present code assumes to much about these commands.
# 
# - Warn if option gets set twice in same vnode.
# 
# Later or never:
#     - show_context option.
#     - encoding option: can override @encoding directives
#     - Support docutils config files.
# 
#@-at
#@@c

#@+others
#@+node:ekr.20050804081215:More options
#@@nocolor
#@+at
# http://sourceforge.net/forum/message.php?msg_id=3276408
# 
# Leo has great potential for smoothing some of the bumps
# outside of learning the details of the markup.
# Here is my wish list;
# 
# (done) - Section level handling.
#   - one of the problems with using rst comes when a new
#     section level is inserted into a document. Leo can
#     take care of this effortlessly, providing a correct
#    underline character for each section heading.
# 
#     Other than the section management, I want Leo to
#     assume that I'm writing rst, and should be rendered
#     as-is. If I want to include some pretty Python, I put
#     it in a ..  python-code:: block (or some such)
# 
#     On top of this base could be built special case
#     handling.
# 
# - Assist in viewing rendered output
#   - easy access to the command to render a tree to  html.
#   - I would like an option which automatically inserted
#     a @url node which pointed to the rendered file, or
#     some other trick to make switching between source
#     and output painless.
# 
# - Assist in working with the CSS file
#   - a CSS file is now required by docutils. A @setting
#     could specify the file. I would like some kind of
#     sugar which made it easy to edit the CSS file, maybe
#     an option to create a @file node automatically,
#     like the @url node above. Again, I want to be able
#     to effortlessly edit the CSS and see the results.
# 
# - The third file of interest when writing rst is the config
#    file which controls many behaviours of the renderer.
#    Again, some intelligence which located it, made it
#    convenient to edit it, maybe even linked to
#    documentation for it, would be very nice.
# 
# (done) - I would like a setting which would save the file with the
#   rst markup and make it easy to examine and edit.
# 
# - On the someday/maybe list would be a @publish
#   feature, which configured a directory to ftp the rendered
#   file to.
#@-at
#@nonl
#@-node:ekr.20050804081215:More options
#@-others
#@nonl
#@-node:ekr.20050806162146:<< to do >>
#@nl

controllers = {} # For use by @button rst3 code.

#@+others
#@+node:ekr.20050805162550.4:Module level
#@+node:ekr.20050805162550.5: init
def init ():

    ok = docutils is not None # Ok for unit testing.

    if ok:
        leoPlugins.registerHandler('after-create-leo-frame', onCreate)
        g.plugin_signon(__name__)
    else:
        s = 'rst3 plugin not loaded: can not load docutils'
        g.es_print(s,color='red')

    return ok
#@-node:ekr.20050805162550.5: init
#@+node:ekr.20050805162550.6:onCreate
def onCreate(tag, keywords):

    c = keywords.get('new_c') or keywords.get('c')

    # g.trace('(rst3 plugin)',c)

    if c:
        global controllers
        controllers[c] = rstClass(c)

        # Warning: Do not return anything but None here!
        # Doing so suppresses the loadeing of other 'new' or 'open2' hooks!
#@nonl
#@-node:ekr.20050805162550.6:onCreate
#@+node:ekr.20050806101253:code_block
def code_block (name,arguments,options,content,lineno,content_offset,block_text,state,state_machine):

    '''Implement the code-block directive for docutils.'''

    try:
        language = arguments [0]
        # g.trace(language)
        # See http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/252170
        module = SilverCity and getattr(SilverCity,language)
        generator = module and getattr(module,language+"HTMLGenerator")
        if generator:
            io = StringIO()
            generator().generate_html(io,'\n'.join(content))
            html = '<div class="code-block">\n%s\n</div>\n' % io.getvalue()
        else:
            html = '<div class="code-block">\n%s\n</div>\n' % '<br>\n'.join(content)
        raw = docutils.nodes.raw('',html,format='html')
        return [raw]
    except Exception: # Return html as shown.  Lines are separated by <br> elements.
        g.es_trace('exception in rst3:code_block()')
        g.es_exception()
        return [None]

# See http://docutils.sourceforge.net/spec/howto/rst-directives.html
code_block.arguments = (
    1, # Number of required arguments.
    0, # Number of optional arguments.
    0) # True if final argument may contain whitespace.

# A mapping from option name to conversion function.
if docutils:
    code_block.options = {
        'language':
        docutils.parsers.rst.directives.unchanged # Return the text argument, unchanged.
    }
    code_block.content = 1 # True if content is allowed.

    # Register the directive with docutils.
    docutils.parsers.rst.directives.register_directive('code-block',code_block)
else:
    code_block.options = {}

#@-node:ekr.20050806101253:code_block
#@+node:ekr.20090429055156.63:runUnitTests
def runUnitTests(c):

    controller = rstClass(c)
    p = g.findNodeAnywhere('UnitTests')
    if p:
        leoTest.doTests(c,p=p
        )

#@-node:ekr.20090429055156.63:runUnitTests
#@-node:ekr.20050805162550.4:Module level
#@+node:ekr.20050805162550.39:html parser classes
#@+doc
# The parser classes are used to construct the html code for nodes. The 
# algorithm has two phases:
#     1. In the first phase, the html code for each node is identified.
#     2. The second phase identifies all links and checks if these links need 
# to be modified.
# The first phase of scanning is done by the anchor_hmlParserClass. The second 
# phase of this algorithm is
# done with the link_htmlParserClass.
#@-doc
#@@code
#@<< class linkAnchorParserClass >>
#@+node:ekr.20050805162550.40: << class linkAnchorParserClass >> (subclass of HTMLParser.HTMLParser)
class linkAnchorParserClass (HTMLParser.HTMLParser):

    '''
    A class to recognize anchors and links in HTML documents.
    A special marker is the "node_marker" which demarkates the border between 
    node and the next.
    '''

    #@    @+others
    #@+node:ekr.20050805162550.41:__init__
    def __init__(self,rst):

        HTMLParser.HTMLParser.__init__(self) # Init the base class.

        self.rst = rst

        # Set ivars from options.  This works only if we don't change nodes!
        self.node_begin_marker      = rst.getOption('node_begin_marker')
        self.clear_http_attributes  = rst.getOption('clear_http_attributes')

        self.current_file = rst.outputFileName
    #@nonl
    #@-node:ekr.20050805162550.41:__init__
    #@+node:ekr.20050805162550.42:is_anchor
    def is_anchor(self, tag, attrs):
        """
        Check if the current tag is an anchor.
        Returns *all* anchors.
        Works with docutils 0.4
        """

        if tag == 'a':
            return True

        if self.is_node_marker(attrs):
            return True
        return tag == "span"
    #@-node:ekr.20050805162550.42:is_anchor
    #@+node:ekr.20050805162550.43:is_link
    def is_link(self, tag, attrs):
        '''
        Return True if tag, attrs is represents a link.
        '''

        if tag != 'a':
            return False

        result = 'href' in dict(attrs)
        return result
    #@-node:ekr.20050805162550.43:is_link
    #@+node:ekr.20050815164715:is_node_marker
    def is_node_marker (self,attrs):
        '''
        Return the name of the anchor, if this is an anchor for the beginning of a node,
        False otherwise.
        '''

        d = dict(attrs)
        result = 'id' in d and d['id'].startswith(self.node_begin_marker)
        if result:
            return d['id']
        return result
    #@-node:ekr.20050815164715:is_node_marker
    #@-others
#@nonl
#@-node:ekr.20050805162550.40: << class linkAnchorParserClass >> (subclass of HTMLParser.HTMLParser)
#@nl

#@+others
#@+node:ekr.20050805162550.44:class htmlParserClass (linkAnchorParserClass)
class htmlParserClass (linkAnchorParserClass):

    '''
    The responsibility of the html parser is:
        1. Find out which html code belongs to which node.
        2. Keep a stack of open tags which apply to the current node.
        3. Keep a list of tags which should be included in the nodes, even
           though they might be closed.
           The <style> tag is one example of that.

    Later, we have to relocate inter-file links: if a reference to another location
    is in a file, we must change the link.

    '''

    #@    @+others
    #@+node:ekr.20050805162550.45:__init__
    def __init__ (self,rst):

        linkAnchorParserClass.__init__(self,rst) # Init the base class.

        self.stack = None
        # The stack contains lists of the form:
            # [text1, text2, previous].
            # text1 is the opening tag
            # text2 is the closing tag
            # previous points to the previous stack element

        self.node_marker_stack = []
        # self.node_marker_stack.pop() returns True for a closing
        # tag if the opening tag identified an anchor belonging to a vnode.

        self.node_code = []
            # Accumulated html code.
            # Once the hmtl code is assigned a vnode, it is deleted here.

        self.deleted_lines = 0 # Number of lines deleted in self.node_code

        self.endpos_pending = False
        # Do not include self.node_code[0:self.endpos_pending] in the html code.

        self.last_position = None
        # Last position; we must attach html code to this node.

        self.last_marker = None
    #@-node:ekr.20050805162550.45:__init__
    #@+node:ekr.20050805162550.46:handle_starttag
    def handle_starttag (self,tag,attrs):
        '''
        1. Find out if the current tag is an achor.
        2. If it is an anchor, we check if this anchor marks the beginning of a new 
           node
        3. If a new node begins, then we might have to store html code for the previous
           node.
        4. In any case, put the new tag on the stack.
        '''
        is_node_marker = False
        if self.is_anchor(tag,attrs) and self.is_node_marker(attrs):
            is_node_marker = self.is_node_marker(attrs)
            # g.trace(tag,attrs)
            line, column = self.getpos()
            if self.last_position:
                lines = self.node_code [:]
                lines [0] = lines [0] [self.startpos:]
                del lines [line-self.deleted_lines-1:]
                # g.trace('Storing in %s...\n%s' % self.last_position, lines)
                mod_http.get_http_attribute(self.last_position).extend(lines)
                #@            << trace the unknownAttribute >>
                #@+node:ekr.20050815164715.1:<< trace the unknownAttribute >>
                if 0:
                    g.pr("rst3: unknownAttributes[self.http_attributename]")
                    g.pr("For:", self.last_position)
                    pprint.pprint(mod_http.get_http_attribute(self.last_position))
                #@nonl
                #@-node:ekr.20050815164715.1:<< trace the unknownAttribute >>
                #@nl
            if self.deleted_lines < line-1:
                del self.node_code [: line-1-self.deleted_lines]
                self.deleted_lines = line-1
                self.endpos_pending = True
        # g.trace("rst2: handle_starttag:", tag, attrs, is_node_marker)
        starttag = self.get_starttag_text()
        self.stack = [starttag, None, self.stack]
        self.node_marker_stack.append(is_node_marker)
    #@nonl
    #@-node:ekr.20050805162550.46:handle_starttag
    #@+node:ekr.20050805162550.47:handle_endtag
    def handle_endtag(self, tag):
        '''
        1. Set the second element of the current top of stack.
        2. If this is the end tag for an anchor for a node,
           store the current stack for that node.
        '''
        self.stack[1] = "</" + tag + ">"

        # g.trace(tag,g.listToString(self.stack))
        if self.endpos_pending:
            line, column = self.getpos()
            self.startpos = self.node_code[0].find(">", column) + 1
            self.endpos_pending = False

        is_node_marker = self.node_marker_stack.pop()

        if is_node_marker and not self.clear_http_attributes:
            self.last_position = self.rst.http_map[is_node_marker]
            if is_node_marker != self.last_marker:
                if bwm_file: print >> bwm_file, "Handle endtag:", is_node_marker, self.stack
                mod_http.set_http_attribute(self.rst.http_map[is_node_marker], self.stack)
                self.last_marker = is_node_marker
                #bwm: last_marker is not needed?

        self.stack = self.stack[2]
    #@-node:ekr.20050805162550.47:handle_endtag
    #@+node:ekr.20050805162550.49:feed
    def feed(self, line):

        # g.trace(repr(line))

        self.node_code.append(line)

        HTMLParser.HTMLParser.feed(self, line) # Call the base class's feed().
    #@-node:ekr.20050805162550.49:feed
    #@-others
#@nonl
#@-node:ekr.20050805162550.44:class htmlParserClass (linkAnchorParserClass)
#@+node:ekr.20050805162550.50:class anchor_htmlParserClass (linkAnchorParserClass)
class anchor_htmlParserClass (linkAnchorParserClass):

    '''
    This htmlparser does the first step of relocating: finding all the anchors within the html nodes.

    Each anchor is mapped to a tuple:
        (current_file, position).

    Filters out markers which mark the beginning of the html code for a node.
    '''

    #@    @+others
    #@+node:ekr.20050805162550.51: __init__
    def __init__ (self,rst,p):

        linkAnchorParserClass.__init__(self,rst)

        self.p = p.copy()
        self.anchor_map = rst.anchor_map
    #@-node:ekr.20050805162550.51: __init__
    #@+node:ekr.20050805162550.52:handle_starttag
    def handle_starttag(self, tag, attrs):
        '''
        1. Find out if the current tag is an achor.
        2. If the current tag is an anchor, update the mapping;
             anchor -> (filename, p)
        '''
        if not self.is_anchor(tag, attrs):
            return

        if self.current_file not in self.anchor_map:
            self.anchor_map[self.current_file] = (self.current_file, self.p)
            simple_name = g.os_path_split(self.current_file)[1]
            self.anchor_map[simple_name] = self.anchor_map[self.current_file]
            if bwm_file: print >> bwm_file, "anchor(1): current_file:", self.current_file, "position:", self.p, "Simple name:", simple_name
            # Not sure what to do here, exactly. Do I need to manipulate
            # the pathname?

        for name, value in attrs:
            if name == 'name' or tag == 'span' and name == 'id':
                if not value.startswith(self.node_begin_marker):
                    if bwm_file: print >> bwm_file, "anchor(2):", value, self.p
                    self.anchor_map[value] = (self.current_file, self.p.copy())
    #@-node:ekr.20050805162550.52:handle_starttag
    #@-others
#@nonl
#@-node:ekr.20050805162550.50:class anchor_htmlParserClass (linkAnchorParserClass)
#@+node:ekr.20050805162550.53:class link_htmlParserClass (linkAnchorParserClass)
class link_htmlparserClass (linkAnchorParserClass):

    '''This html parser does the second step of relocating links:
    1. It scans the html code for links.
    2. If there is a link which links to a previously processed file
       then this link is changed so that it now refers to the node.
    '''

    #@    @+others
    #@+node:ekr.20050805162550.54:__init__
    def __init__ (self,rst,p):

        linkAnchorParserClass.__init__(self,rst)

        self.p = p.copy()
        self.anchor_map = rst.anchor_map
        self.replacements = []
    #@nonl
    #@-node:ekr.20050805162550.54:__init__
    #@+node:ekr.20050805162550.55:handle_starttag
    def handle_starttag(self, tag, attrs):
        '''
        1. Find out if the current tag is an achor.
        2. If the current tag is an anchor, update the mapping;
             anchor -> p
            Update the list of replacements for the document.
        '''
        if bwm_file: print >> bwm_file, "Is link?", tag, attrs
        if not self.is_link(tag, attrs):
            return

        marker = self.node_begin_marker
        for name, value in attrs:
            if name == 'href':
                href = value
                href_parts = href.split("#")
                if len(href_parts) == 1:
                    href_a = href_parts[0]
                else:
                    href_a = href_parts[1]
                if bwm_file: print >> bwm_file, "link(1):", name, value, href_a
                if not href_a.startswith(marker):
                    if href_a in self.anchor_map:
                        href_file, href_node = self.anchor_map[href_a]
                        http_node_ref = mod_http.node_reference(href_node)
                        line, column = self.getpos()
                        if bwm_file: print >> bwm_file, "link(2):", line, column, href, href_file, http_node_ref
                        self.replacements.append((line, column, href, href_file, http_node_ref))
    #@nonl
    #@-node:ekr.20050805162550.55:handle_starttag
    #@+node:ekr.20050805162550.56:get_replacements
    def get_replacements(self):

        return self.replacements
    #@nonl
    #@-node:ekr.20050805162550.56:get_replacements
    #@-others
#@nonl
#@-node:ekr.20050805162550.53:class link_htmlParserClass (linkAnchorParserClass)
#@-others
#@nonl
#@-node:ekr.20050805162550.39:html parser classes
#@+node:ekr.20050805162550.8:class rstClass
class rstClass:

    '''A class to write rst markup in Leo outlines.'''

#@+at 
#@nonl
# This plugin optionally stores information for the http plugin.
# 
# Each node can have one additional attribute, with the name 
# rst_http_attributename, which is a list.
# 
# The first three elements are stack of tags, the rest is html code.
# 
# [<tag n start>, <tag n end>, <other stack elements>, <html line 1>, <html 
# line 2>, ...]
# 
# <other stack elements has the same structure:
#     [<tag n-1 start>, <tag n-1 end>, <other stack elements>]
#@-at
#@@c

    #@    @+others
    #@+node:ekr.20050805162550.9: Birth & init
    #@+node:ekr.20050805162550.10: ctor (rstClass)
    def __init__ (self,c):

        global SilverCity

        # g.trace('rst3.py:rstClass',g.callers())

        self.c = c
        #@    << init ivars >>
        #@+node:ekr.20050805162550.11:<< init ivars >>
        self.silverCityWarningGiven = False

        # The options dictionary.
        self.optionsDict = {}
        self.option_prefix = '@rst-option'

        # Formatting...
        self.code_block_string = ''
        self.node_counter = 0
        self.toplevel = 0
        self.topNode = None
        self.use_alternate_code_block = SilverCity is None

        # Http support...
        self.nodeNumber = 0
        # All nodes are numbered so that unique anchors can be generated.

        self.http_map = {} 
        # Keys are named hyperlink targets.  Value are positions.
        # The targets mark the beginning of the html code specific
        # for this position.

        self.anchor_map = {}
        # Maps anchors (generated by this module) to positions

        self.rst3_all = False
        # Set to True by the button which processes all @rst trees.

        # For writing.
        self.defaultEncoding = 'utf-8'
        self.leoDirectivesList = g.globalDirectiveList
        self.encoding = self.defaultEncoding
        self.ext = None # The file extension.
        self.outputFileName = None # The name of the file being written.
        self.outputFile = None # The open file being written.
        self.path = '' # The path from any @path directive.
        self.source = None # The written source as a string.
        #@nonl
        #@-node:ekr.20050805162550.11:<< init ivars >>
        #@nl

        self.createDefaultOptionsDict()
        self.initOptionsFromSettings() # Still needed.
        self.initHeadlineCommands() # Only needs to be done once.
        self.initSingleNodeOptions()
        self.addMenu()
    #@-node:ekr.20050805162550.10: ctor (rstClass)
    #@+node:ekr.20050805162550.12:addMenu
    def addMenu (self):

        c = self.c ; editMenu = c.frame.menu.getMenu('Edit')

        def rst3PluginCallback (event=None):
            self.processTopTree(c.p)

        c.k.registerCommand('write-restructured-text',shortcut=None,
            func=rst3PluginCallback,pane='all',verbose=False)

        table = (
            ("-",None,None),
            ### ("Write Restructed Text","",rst3PluginCallback),
            '&write-restructured-text',
        )

        c.frame.menu.createMenuEntries(editMenu,table,dynamicMenu=True)
    #@nonl
    #@-node:ekr.20050805162550.12:addMenu
    #@+node:ekr.20050813083007:initHeadlineCommands
    def initHeadlineCommands (self):

        '''Init the list of headline commands used by writeHeadline.'''

        self.headlineCommands = [
            self.getOption('code_prefix'),
            self.getOption('doc_only_prefix'),
            self.getOption('default_path_prefix'),
            self.getOption('rst_prefix'),
            self.getOption('ignore_headline_prefix'),
            self.getOption('ignore_headlines_prefix'),
            self.getOption('ignore_node_prefix'),
            self.getOption('ignore_tree_prefix'),
            self.getOption('option_prefix'),
            self.getOption('options_prefix'),
            self.getOption('show_headline_prefix'),
            # # Suggested by Hemanth P.S.: prevent @file nodes from creating headings.
            # self.getOption('keep_at_file_prefix'),
            # self.getOption('strip_at_file_prefix'),
        ]
    #@nonl
    #@-node:ekr.20050813083007:initHeadlineCommands
    #@+node:ekr.20050813085236:initSingleNodeOptions
    def initSingleNodeOptions (self):

        self.singleNodeOptions = [
            'ignore_this_headline',
            'ignore_this_node',
            'ignore_this_tree',
            'preformat_this_node',
            'show_this_headline',
        ]
    #@nonl
    #@-node:ekr.20050813085236:initSingleNodeOptions
    #@+node:ekr.20050808072943:munge
    def munge (self,name):

        '''Convert an option name to the equivalent ivar name.'''

        i = g.choose(name.startswith('rst'),3,0)

        while i < len(name) and name[i].isdigit():
            i += 1

        if i < len(name) and name[i] == '_':
            i += 1

        s = name[i:].lower()
        s = s.replace('-','_')

        return s
    #@nonl
    #@-node:ekr.20050808072943:munge
    #@-node:ekr.20050805162550.9: Birth & init
    #@+node:ekr.20050812122236:options...
    #@+node:ekr.20050808064245:createDefaultOptionsDict
    def createDefaultOptionsDict(self):

        # Warning: changing the names of options changes the names of the corresponding ivars.

        self.defaultOptionsDict = {
            # Http options...
            'rst3_clear_http_attributes':   False,
            'rst3_http_server_support':     False,
            'rst3_http_attributename':      'rst_http_attribute',
            'rst3_node_begin_marker':       'http-node-marker-',
            # Path options...
            'rst3_default_path': None, # New in Leo 4.4a4 # Bug fix: must be None, not ''.
            'rst3_stylesheet_name': 'default.css',
            'rst3_stylesheet_path': None, # Bug fix: must be None, not ''.
            'rst3_publish_argv_for_missing_stylesheets': None,
            # Global options...
            'rst3_code_block_string': '',
            'rst3_number_code_lines': True,
            'rst3_underline_characters': '''#=+*^~"'`-:><_''',
            'rst3_verbose':True,
            'rst3_write_intermediate_file': False, # Used only if generate_rst is True.
            # Mode options...
            'rst3_code_mode': False, # True: generate rst markup from @code and @doc parts.
            'rst3_doc_only_mode': False, # True: generate only from @doc parts.
            'rst3_generate_rst': True, # True: generate rst markup.  False: generate plain text.
            'rst3_generate_rst_header_comment': True,
                # True generate header comment (requires generate_rst option)
            # Formatting options that apply to both code and rst modes....
            'rst3_show_headlines': True,  # Can be set by @rst-no-head headlines.
            'rst3_show_organizer_nodes': True,
            'rst3_show_options_nodes': False,
            'rst3_show_sections': True,
            'rst3_strip_at_file_prefixes': True,
            'rst3_show_doc_parts_in_rst_mode': True,
            # Formatting options that apply only to code mode.
            'rst3_show_doc_parts_as_paragraphs': False,
            'rst3_show_leo_directives': True,
            'rst3_show_markup_doc_parts': False,
            'rst3_show_options_doc_parts': False,
            # *Names* of headline commands...
            'rst3_code_prefix':             '@rst-code',     # Enter code mode.
            'rst3_doc_only_prefix':         '@rst-doc-only', # Enter doc-only mode.
            'rst3_rst_prefix':              '@rst',          # Enter rst mode.
            'rst3_ignore_headline_prefix':  '@rst-no-head',
            'rst3_ignore_headlines_prefix': '@rst-no-headlines',
            'rst3_ignore_node_prefix':      '@rst-ignore-node',
            'rst3_ignore_prefix':           '@rst-ignore',
            'rst3_ignore_tree_prefix':      '@rst-ignore-tree',
            'rst3_option_prefix':           '@rst-option',
            'rst3_options_prefix':          '@rst-options',
            'rst3_preformat_prefix':        '@rst-preformat',
            'rst3_show_headline_prefix':    '@rst-head',
        }
    #@nonl
    #@-node:ekr.20050808064245:createDefaultOptionsDict
    #@+node:ekr.20050812120933:dumpSettings (debugging)
    def dumpSettings (self):

        d = self.optionsDict
        keys = d.keys() ; keys.sort()

        g.pr('present settings...')
        for key in keys:
            g.pr('%20s %s' % (key,d.get(key)))
    #@nonl
    #@-node:ekr.20050812120933:dumpSettings (debugging)
    #@+node:ekr.20050814134351:getOption
    def getOption (self,name):

        bwm = False
        if bwm:
            g.trace("bwm: getOption self:%s, name:%s, value:%s" % (
                self, name, self.optionsDict.get(name)))

        return self.optionsDict.get(name)
    #@nonl
    #@-node:ekr.20050814134351:getOption
    #@+node:ekr.20071015110830:initCodeBlockString
    def initCodeBlockString(self,p):

        # New in Leo 4.4.4: do this here, not in initWrite:
        c = self.c
        d = c.scanAllDirectives(p)
        language = d.get('language')
        if language is None: language = 'python'
        else: language = language.lower()
        syntax = SilverCity is not None

        # g.trace('language',language,'language.title()',language.title(),p.h)

        s = self.getOption('code_block_string')
        if s:
            self.code_block_string = s.replace('\\n','\n')
        elif syntax and language in ('python','ruby','perl','c'):
            self.code_block_string = '**code**:\n\n.. code-block:: %s\n' % language.title()
        else:
            self.code_block_string = '**code**:\n\n.. class:: code\n..\n\n::\n'
    #@-node:ekr.20071015110830:initCodeBlockString
    #@+node:ekr.20050807120331.1:preprocessTree & helpers
    def preprocessTree (self,root):

        self.tnodeOptionDict = {}

        # Bug fix 12/4/05: must preprocess parents too.
        for p in root.parents():
            self.preprocessNode(p)

        for p in root.self_and_subtree():
            self.preprocessNode(p)

        if 0:
            g.trace(root.h)
            for key in self.tnodeOptionDict.keys():
                g.trace(key)
                g.printDict(self.tnodeOptionDict.get(key))
    #@nonl
    #@+node:ekr.20051204070141:preprocessNode
    def preprocessNode (self,p):

        d = self.tnodeOptionDict.get(p.v)
        if d is None:
            d = self.scanNodeForOptions(p)
            self.tnodeOptionDict [p.v] = d
    #@nonl
    #@-node:ekr.20051204070141:preprocessNode
    #@+node:ekr.20050808072943.1:parseOptionLine
    def parseOptionLine (self,s):

        '''Parse a line containing name=val and return (name,value) or None.

        If no value is found, default to True.'''

        s = s.strip()
        if s.endswith(','): s = s[:-1]
        # Get name.  Names may contain '-' and '_'.
        i = g.skip_id(s,0,chars='-_')
        name = s [:i]
        if not name: return None
        j = g.skip_ws(s,i)
        if g.match(s,j,'='):
            val = s [j+1:].strip()
            # g.trace(val)
            return name,val
        else:
            # g.trace('*True')
            return name,'True'
    #@nonl
    #@-node:ekr.20050808072943.1:parseOptionLine
    #@+node:ekr.20050808070018.2:scanForOptionDocParts
    def scanForOptionDocParts (self,p,s):

        '''Return a dictionary containing all options from rstoptionsdocpartsp. import 
        Multiple @rst-options doc parts are allowed: this code aggregates all options.
        '''

        d = {} ; n = 0 ; lines = g.splitLines(s)
        while n < len(lines):
            line = lines[n] ; n += 1
            if line.startswith('@'):
                i = g.skip_ws(line,1)
                for kind in ('@rst-options','@rst-option'):
                    if g.match_word(line,i,kind):
                        # Allow options on the same line.
                        line = line[i+len(kind):]
                        d.update(self.scanOption(p,line))
                        # Add options until the end of the doc part.
                        while n < len(lines):
                            line = lines[n] ; n += 1 ; found = False
                            for stop in ('@c','@code', '@'):
                                if g.match_word(line,0,stop):
                                    found = True ; break
                            if found:
                                break
                            else:
                                d.update(self.scanOption(p,line))
                        break
        return d
    #@nonl
    #@-node:ekr.20050808070018.2:scanForOptionDocParts
    #@+node:ekr.20050811173750:scanHeadlineForOptions
    def scanHeadlineForOptions (self,p):

        '''Return a dictionary containing the options implied by p's headline.'''

        h = p.h.strip()

        if p == self.topNode:
            return {} # Don't mess with the root node.
        elif g.match_word(h,0,self.getOption('option_prefix')): # '@rst-option'
            s = h [len(self.option_prefix):]
            return self.scanOption(p,s)
        elif g.match_word(h,0,self.getOption('options_prefix')): # '@rst-options'
            return self.scanOptions(p,p.b)
        else:
            # Careful: can't use g.match_word because options may have '-' chars.
            i = g.skip_id(h,0,chars='@-')
            word = h[0:i]

            for prefix,ivar,val in (
                ('code_prefix','code_mode',True), # '@rst-code'
                ('doc_mode_prefix','doc_only_mode',True), # @rst-doc-only.
                ('default_path_prefix','default_prefix',''), # '@rst-default-path'
                ('rst_prefix','code_mode',False), # '@rst'
                ('ignore_headline_prefix','ignore_this_headline',True), # '@rst-no-head'
                ('show_headline_prefix','show_this_headline',True), # '@rst-head'  
                ('ignore_headlines_prefix','show_headlines',False), # '@rst-no-headlines'
                ('ignore_prefix','ignore_this_tree',True),      # '@rst-ignore'
                ('ignore_node_prefix','ignore_this_node',True), # '@rst-ignore-node'
                ('ignore_tree_prefix','ignore_this_tree',True), # '@rst-ignore-tree'
                ('preformat_prefix','preformat_this_node',True), # '@rst-preformat
            ):
                prefix = self.getOption(prefix)
                if prefix and word == prefix: # Do _not_ munge this prefix!
                    d = { ivar: val }
                    if ivar != 'code_mode':
                        d ['code_mode'] = False # Enter rst mode.
                        d ['doc_only_mode'] = False
                    # Special case: Treat a bare @rst like @rst-no-head
                    if h == self.getOption('rst_prefix'):
                        d ['ignore_this_headline'] = True
                    # g.trace(repr(h),repr(prefix),ivar,d)
                    return d

            if h.startswith('@rst'):
                g.trace('word',word,'rst_prefix',self.getOption('rst_prefix'))
                g.trace('unknown kind of @rst headline',p.h)

            return {}
    #@-node:ekr.20050811173750:scanHeadlineForOptions
    #@+node:ekr.20050807120331.2:scanNodeForOptions
    def scanNodeForOptions (self,p):

        '''Return a dictionary containing all the option-name:value entries in p.

        Such entries may arise from rstoptionrstoptionstheheadline import 
        or from rstoptionsdocparts. import 

        h = p.h

        d = self.scanHeadlineForOptions(p)

        d2 = self.scanForOptionDocParts(p,p.b)

        # A fine point: body options over-ride headline options.
        d.update(d2)

        return d
    #@nonl
    #@-node:ekr.20050807120331.2:scanNodeForOptions
    #@+node:ekr.20050808070018:scanOption
    def scanOption (self,p,s):

        '''Return { name:val } if s is a line of the form name=val.
        Otherwise return {}'''

        if not s.strip() or s.strip().startswith('..'): return {}

        data = self.parseOptionLine(s)

        if data:
            name,val = data
            fullName = 'rst3_' + self.munge(name)
            if fullName in self.defaultOptionsDict.keys():
                if   val.lower() == 'true': val = True
                elif val.lower() == 'false': val = False
                # g.trace('%24s %8s %s' % (self.munge(name),val,p.h))
                return { self.munge(name): val }
            else:
                g.es_print('ignoring unknown option: %s' % (name),color='red')
                return {}
        else:
            g.trace(repr(s))
            s2 = 'bad rst3 option in %s: %s' % (p.h,s)
            g.es_print(s2,color='red')
            return {}
    #@nonl
    #@-node:ekr.20050808070018:scanOption
    #@+node:ekr.20050808070018.1:scanOptions
    def scanOptions (self,p,s):

        '''Return a dictionary containing all the options in s.'''

        d = {}

        for line in g.splitLines(s):
            d2 = self.scanOption(p,line)
            if d2: d.update(d2)

        return d
    #@nonl
    #@-node:ekr.20050808070018.1:scanOptions
    #@-node:ekr.20050807120331.1:preprocessTree & helpers
    #@+node:ekr.20050808142313.28:scanAllOptions & helpers
    # Once an option is seen, no other related options in ancestor nodes have any effect.

    def scanAllOptions(self,p):

        '''Scan position p and p's ancestors looking for options,
        setting corresponding ivars.
        '''

        self.initOptionsFromSettings() # Must be done on every node.
        self.handleSingleNodeOptions(p)
        seen = self.singleNodeOptions[:] # Suppress inheritance of single-node options.

        # g.trace('-'*20)
        for p in p.self_and_parents():
            d = self.tnodeOptionDict.get(p.v,{})
            # g.trace(p.h,d)
            for key in d.keys():
                ivar = self.munge(key)
                if not ivar in seen:
                    seen.append(ivar)
                    val = d.get(key)
                    self.setOption(key,val,p.h)

        # self.dumpSettings()
        if self.rst3_all:
            self.setOption("generate_rst", True, "rst3_all")
            self.setOption("generate_rst_header_comment",True, "rst3_all")
            self.setOption("http_server_support", True, "rst3_all")
            self.setOption("write_intermediate_file", True, "rst3_all")
    #@+node:ekr.20050805162550.13:initOptionsFromSettings
    def initOptionsFromSettings (self):

        c = self.c ; d = self.defaultOptionsDict
        keys = d.keys() ; keys.sort()

        for key in keys:
            for getter,kind in (
                (c.config.getBool,'@bool'),
                (c.config.getString,'@string'),
                (d.get,'default'),
            ):
                val = getter(key)
                if kind == 'default' or val is not None:
                    self.setOption(key,val,'initOptionsFromSettings')
                    break
        # Special case.
        if self.getOption('http_server_support') and not mod_http:
            g.es('No http_server_support: can not import mod_http plugin',color='red')
            self.setOption('http_server_support',False)
    #@-node:ekr.20050805162550.13:initOptionsFromSettings
    #@+node:ekr.20050810103731:handleSingleNodeOptions
    def handleSingleNodeOptions (self,p):

        '''Init the settings of single-node options from thetnodeOptionsDict. import 

        All such options default to False.'''

        d = self.tnodeOptionDict.get(p.v, {} )

        for ivar in self.singleNodeOptions:
            val = d.get(ivar,False)
            #g.trace('%24s %8s %s' % (ivar,val,p.h))
            self.setOption(ivar,val,p.h)

    #@-node:ekr.20050810103731:handleSingleNodeOptions
    #@-node:ekr.20050808142313.28:scanAllOptions & helpers
    #@+node:ekr.20050811135526:setOption
    def setOption (self,name,val,tag):

        ivar = self.munge(name)

        bwm = False
        if bwm:
            if not self.optionsDict.has_key(ivar):
                g.trace('init %24s %20s %s %s' % (ivar, val, tag, self))
            elif self.optionsDict.get(ivar) != val:
                g.trace('set  %24s %20s %s %s' % (ivar, val, tag, self))

        self.optionsDict [ivar] = val
    #@nonl
    #@-node:ekr.20050811135526:setOption
    #@-node:ekr.20050812122236:options...
    #@+node:ekr.20050809074827:write methods
    #@+node:ekr.20050809082854: Top-level write code
    #@+node:ekr.20050809075309:initWrite
    def initWrite (self,p):

        self.initOptionsFromSettings() # Still needed.

        # Set the encoding from any parent @encoding directive.
        # This can be overridden by @rst-option encoding=whatever.
        c = self.c
        d = c.scanAllDirectives(p)
        self.encoding = d.get('encoding') or self.defaultEncoding
        self.path = d.get('path') or ''

        # g.trace('path:',self.path)
    #@-node:ekr.20050809075309:initWrite
    #@+node:ekr.20050809080925:writeNormalTree
    def writeNormalTree (self,p,toString=False):

        self.initWrite(p)

        # Always write to a string first.
        self.outputFile = StringIO()
        self.writeTree(p)
        self.source = self.stringOutput = self.outputFile.getvalue()

        # Copy to a file if requested.
        if not toString:
            # Comput the output file name *after* calling writeTree.
            self.outputFileName = self.computeOutputFileName(self.outputFileName)
            self.outputFile = open(self.outputFileName,'w')
            self.outputFile.write(self.stringOutput)
            self.outputFile.close()

        return True
    #@-node:ekr.20050809080925:writeNormalTree
    #@+node:ekr.20051121102358:processTopTree
    def processTopTree (self,p,justOneFile=False):

        c = self.c ; current = p.copy()

        for p in current.self_and_parents():
            h = p.h
            if h.startswith('@rst') and not h.startswith('@rst-'):
                self.processTree(p,ext=None,toString=False,justOneFile=justOneFile)
                break
        else:
            self.processTree(current,ext=None,toString=False,justOneFile=justOneFile)

        g.es_print('done',color='blue')
    #@nonl
    #@-node:ekr.20051121102358:processTopTree
    #@+node:ekr.20050805162550.17:processTree
    def processTree(self,p,ext,toString,justOneFile):

        '''Process all @rst nodes in a tree.'''

        self.preprocessTree(p)
        found = False ; self.stringOutput = ''
        p = p.copy() ; after= p.nodeAfterTree()
        while p and p != after:
            h = p.h.strip()
            if g.match_word(h,0,"@rst"):
                self.outputFileName = h[4:].strip()
                if (
                    (self.outputFileName and self.outputFileName[0] != '-') or
                    (toString and not self.outputFileName)
                ):
                    found = True
                    self.toplevel = p.level() # Define toplevel separately for each rst file.
                    if toString:
                        self.ext = ext
                    else:
                        self.ext = g.os_path_splitext(self.outputFileName)[1].lower()
                    # g.trace('ext',self.ext,self.outputFileName)
                    if self.ext in ('.htm','.html','.tex','.pdf'):
                        ok = self.writeSpecialTree(p,toString=toString,justOneFile=justOneFile)
                    else:
                        ok = self.writeNormalTree(p,toString=toString)
                    self.scanAllOptions(p) # Restore the top-level verbose setting.
                    if toString:
                        return p.copy(),self.stringOutput
                    else:
                        if ok: self.report(self.outputFileName)
                    p.moveToNodeAfterTree()
                else:
                    p.moveToThreadNext()
            else: p.moveToThreadNext()
        if not found:
            g.es('No @rst nodes in selected tree',color='blue')
        return None,None
    #@-node:ekr.20050805162550.17:processTree
    #@+node:ekr.20050805162550.21:writeSpecialTree
    def writeSpecialTree (self,p,toString,justOneFile):

        c = self.c
        isHtml = self.ext in ('.html','.htm')
        if isHtml and not SilverCity:
            if not self.silverCityWarningGiven:
                self.silverCityWarningGiven = True
                g.es('SilverCity not present so no syntax highlighting')

        self.initWrite(p)
            # was ,encoding=g.choose(isHtml,'utf-8','iso-8859-1'))
        self.outputFile = StringIO()
        self.writeTree(p)
        self.source = self.outputFile.getvalue()
        self.outputFile = None

        if not toString:
            # Compute this here for use by intermediate file.
            self.outputFileName = self.computeOutputFileName(self.outputFileName)

            # Create the directory if it doesn't exist.
            theDir, junk = g.os_path_split(self.outputFileName)
            theDir = c.os_path_finalize(theDir)
            if not g.os_path_exists(theDir):
                ok = g.makeAllNonExistentDirectories(theDir,c=c,force=False)
                if not ok:
                    g.es_print('did not create:',theDir,color='red')
                    return False

            # if not os.access(theDir,os.F_OK):
                # os.mkdir(theDir)

            if self.getOption('write_intermediate_file'):
                name = self.outputFileName + '.txt'
                f = open(name,'w')
                f.write(self.source)
                f.close()
                self.report(name)

        try:
            output = self.writeToDocutils(self.source)
            ok = True
        except Exception:
            g.pr('Exception in docutils')
            g.es_exception()
            ok = False

        if ok:
            if isHtml:
                import re
                idxTitle = output.find('<title></title>')
                if idxTitle > -1:
                    m = re.search('<h1>([^<]*)</h1>', output)
                    if not m:
                        m = re.search('<h1><[^>]+>([^<]*)</a></h1>', output)
                    if m:
                        output = output.replace(
                            '<title></title>',
                            '<title>%s</title>' % m.group(1)
                        )


            if toString:
                self.stringOutput = output
            else:
                # Write the file to the directory containing the .leo file.
                f = open(self.outputFileName,'w')
                f.write(output)
                f.close()
                self.http_endTree(self.outputFileName, p, justOneFile=justOneFile)

        return ok
    #@-node:ekr.20050805162550.21:writeSpecialTree
    #@+node:ekr.20050809082854.1:writeToDocutils (sets argv) & helper
    def writeToDocutils (self,s):

        '''Send s to docutils using the writer implied by self.ext and return the result.'''

        openDirectory = self.c.frame.openDirectory
        overrides = {'output_encoding': self.encoding }

        # Compute the args list if the stylesheet path does not exist.
        styleSheetArgsDict = self.handleMissingStyleSheetArgs()

        for ext,writer in (
            ('.html','html'),
            ('.htm','html'),
            ('.tex','latex'),
            ('.pdf','leo_pdf'),
        ):
            if self.ext == ext:
                break
        else:
            g.es_print('unknown docutils extension: %s' % (self.ext),color='red')
            return ''

        # Make the stylesheet path relative to the directory containing the output file.
        rel_stylesheet_path = self.getOption('stylesheet_path') or ''

        # New in Leo 4.5: The rel_stylesheet_path is relative to the open directory.
        stylesheet_path = g.os_path_finalize_join(
            self.c.frame.openDirectory,rel_stylesheet_path)

        path = g.os_path_finalize_join(
            stylesheet_path,self.getOption('stylesheet_name'))

        res = ""
        if g.os_path_exists(path):
            if self.ext != '.pdf':
                overrides['stylesheet'] = path
                overrides['stylesheet_path'] = None
        elif styleSheetArgsDict:
            g.es_print('using publish_argv_for_missing_stylesheets',
                styleSheetArgsDict)
            overrides.update(styleSheetArgsDict)
                # MWC add args to settings
        elif rel_stylesheet_path == stylesheet_path:
            g.es_print('stylesheet not found: %s' % (path),color='red')
        else:
            g.es_print('stylesheet not found\n',path,color='red')
            if self.path:g.es_print('@path:', self.path)
            g.es_print('open path:',self.c.frame.openDirectory)
            if rel_stylesheet_path:
                g.es_print('relative path:', rel_stylesheet_path)
        try:
            # All paths now come through here.
            res = docutils.core.publish_string(source=s,
                    reader_name='standalone',
                    parser_name='restructuredtext',
                    writer_name=writer,
                    settings_overrides=overrides)
        except docutils.ApplicationError as error:
            g.es_print('Error (%s): %s' % (error.__class__.__name__, error))
        return res
    #@+node:ekr.20090428082801.64:handleMissingStyleSheetArgs
    def handleMissingStyleSheetArgs (self,s=None):

        '''Parse the publish_argv_for_missing_stylesheets option,
        returning a dict containing the parsed args.'''

        d = {}
        if not s:
            s = self.getOption('publish_argv_for_missing_stylesheets')
        if not s: return d

        args = s.strip()
        if args.find(',') == -1:
            args = [args]
        else:
            args = args.split(',')

        for arg in args:
            data = arg.split('=')
            if len(data) == 1:
                key = data[0]
                d[str(key)] = ""
            elif len(data) == 2:
                key,value = data
                d[str(key)] = str(value)
            else:
                g.es_print('bad option: %s' % s,color='red')
                break

        return d
    #@-node:ekr.20090428082801.64:handleMissingStyleSheetArgs
    #@-node:ekr.20050809082854.1:writeToDocutils (sets argv) & helper
    #@+node:ekr.20060525102337:writeNodeToString (New in 4.4.1)
    def writeNodeToString (self,p=None,ext=None):

        '''Scan p's tree (defaults to presently selected tree) looking for @rst nodes.
        Convert the first node found to an ouput of the type specified by ext.

        The @rst may or may not be followed by a filename; the filename is *ignored*,
        and its type does not affect ext or the output generated in any way.

        ext should start with a period:  .html, .tex or None (specifies rst output).

        Returns p, s, where p is the position of the @rst node and s is the converted text.'''

        c = self.c ; current = p or c.p

        for p in current.self_and_parents():
            if p.h.startswith('@rst'):
                return self.processTree(p,ext=ext,toString=True,justOneFile=True)
        else:
            return self.processTree(current,ext=ext,toString=True,justOneFile=True)
    #@nonl
    #@-node:ekr.20060525102337:writeNodeToString (New in 4.4.1)
    #@-node:ekr.20050809082854: Top-level write code
    #@+node:ekr.20050811154552:getDocPart
    def getDocPart (self,lines,n):

        # g.trace('n',n,repr(''.join(lines)))

        result = []
        #@    << Append whatever follows @doc or @space to result >>
        #@+node:ekr.20060610104435:<< Append whatever follows @doc or @space to result >>
        if n > 0:
            line = lines[n-1]
            if line.startswith('@doc'):
                s = line[4:].lstrip()
            elif line.startswith('@'):
                s = line[1:].lstrip()
            else:
                s = ''

            # New in Leo 4.4.4: remove these special tags.
            for tag in ('@rst-options','@rst-option','@rst-markup'):
                if g.match_word(s,0,tag):
                    s = s[len(tag):].strip()

            if s.strip():
                result.append(s)
        #@-node:ekr.20060610104435:<< Append whatever follows @doc or @space to result >>
        #@nl
        while n < len(lines):
            s = lines [n] ; n += 1
            if g.match_word(s,0,'@code') or g.match_word(s,0,'@c'):
                break
            result.append(s)
        return n, result
    #@nonl
    #@-node:ekr.20050811154552:getDocPart
    #@+node:ekr.20050811102607:skip_literal_block
    def skip_literal_block (self,lines,n):

        s = lines[n] ; result = [s] ; n += 1
        indent = g.skip_ws(s,0)

        # Skip lines until a non-blank line is found with same or less indent.
        while n < len(lines):
            s = lines[n]
            indent2 = g.skip_ws(s,0)
            if s and not s.isspace() and indent2 <= indent:
                break # We will rescan lines [n]
            n += 1
            result.append(s)

        # g.printList(result,tag='literal block')
        return n, result
    #@nonl
    #@-node:ekr.20050811102607:skip_literal_block
    #@+node:ekr.20050811101550.1:writeBody & helpers
    def writeBody (self,p):

        # remove trailing cruft and split into lines.
        lines = p.b.rstrip().split('\n') 

        if self.getOption('code_mode'):
            if not self.getOption('show_options_doc_parts'):
                lines = self.handleSpecialDocParts(lines,'@rst-options',
                    retainContents=False)
            if not self.getOption('show_markup_doc_parts'):
                lines = self.handleSpecialDocParts(lines,'@rst-markup',
                    retainContents=False)
            if not self.getOption('show_leo_directives'):
                lines = self.removeLeoDirectives(lines)
            lines = self.handleCodeMode(lines)
        elif self.getOption('doc_only_mode'):
            # New in version 1.15
            lines = self.handleDocOnlyMode(p,lines)
        else:
            lines = self.handleSpecialDocParts(lines,'@rst-options',
                retainContents=False)
            lines = self.handleSpecialDocParts(lines,'@rst-markup',
                retainContents=self.getOption('generate_rst'))
            if self.getOption('show_doc_parts_in_rst_mode') is True:
                pass  # original behaviour, treat as plain text
            elif self.getOption('show_doc_parts_in_rst_mode'):
                # use value as class for content
                lines = self.handleSpecialDocParts(lines,None,
                    retainContents=True, asClass=self.getOption('show_doc_parts_in_rst_mode'))
            else:  # option evaluates to false, cut them out
                lines = self.handleSpecialDocParts(lines,None,
                    retainContents=False)
            lines = self.removeLeoDirectives(lines)
            if self.getOption('generate_rst') and self.getOption('use_alternate_code_block'):
                lines = self.replaceCodeBlockDirectives(lines)

        s = '\n'.join(lines).strip()
        if s:
            self.write('%s\n\n' % s)
    #@+node:ekr.20050811150541:handleCodeMode & helper
    def handleCodeMode (self,lines):

        '''Handle the preprocessed body text in code mode as follows:

        - Blank lines are copied after being cleaned.
        - @ @rst-markup lines get copied as is.
        - Everything else gets put into a code-block directive.'''

        result = [] ; n = 0 ; code = []
        while n < len(lines):
            s = lines [n] ; n += 1
            if (
                self.isSpecialDocPart(s,'@rst-markup') or
                (self.getOption('show_doc_parts_as_paragraphs') and self.isSpecialDocPart(s,None))
            ):
                if code:
                    self.finishCodePart(result,code)
                    code = []
                result.append('')
                n, lines2 = self.getDocPart(lines,n)
                result.extend(lines2)
            elif not s.strip() and not code:
                pass # Ignore blank lines before the first code block.
            elif not code: # Start the code block.
                result.append('')
                result.append(self.code_block_string)
                code.append(s)
            else: # Continue the code block.
                code.append(s)

        if code:
            self.finishCodePart(result,code)
            code = []
        return self.rstripList(result)
    #@nonl
    #@+node:ekr.20050811152104:formatCodeModeLine
    def formatCodeModeLine (self,s,n,numberOption):

        if not s.strip(): s = ''

        if numberOption:
            return '\t%d: %s' % (n,s)
        else:
            return '\t%s' % s
    #@nonl
    #@-node:ekr.20050811152104:formatCodeModeLine
    #@+node:ekr.20050813155021:rstripList
    def rstripList (self,theList):

        '''Removed trailing blank lines from theList.'''

        s = '\n'.join(theList).rstrip()
        return s.split('\n')
    #@nonl
    #@-node:ekr.20050813155021:rstripList
    #@+node:ekr.20050813160208:finishCodePart
    def finishCodePart (self,result,code):

        numberOption = self.getOption('number_code_lines')
        code = self.rstripList(code)
        i = 0
        for line in code:
            i += 1
            result.append(self.formatCodeModeLine(line,i,numberOption))
    #@nonl
    #@-node:ekr.20050813160208:finishCodePart
    #@-node:ekr.20050811150541:handleCodeMode & helper
    #@+node:ekr.20060608094815:handleDocOnlyMode
    def handleDocOnlyMode (self,p,lines):

        '''Handle the preprocessed body text in doc_only mode as follows:

        - Blank lines are copied after being cleaned.
        - @ @rst-markup lines get copied as is.
        - All doc parts get copied.
        - All code parts are ignored.'''

        ignore              = self.getOption('ignore_this_headline')
        showHeadlines       = self.getOption('show_headlines')
        showThisHeadline    = self.getOption('show_this_headline')
        showOrganizers      = self.getOption('show_organizer_nodes')

        result = [] ; n = 0
        while n < len(lines):
            s = lines [n] ; n += 1
            if self.isSpecialDocPart(s,'@rst-options'):
                n, lines2 = self.getDocPart(lines,n) # ignore.
            elif self.isAnyDocPart(s):
                # Handle any other doc part, including @rst-markup.
                n, lines2 = self.getDocPart(lines,n)
                if lines2: result.extend(lines2)
        if not result: result = []
        if showHeadlines:
            if result or showThisHeadline or showOrganizers or p == self.topNode:
                # g.trace(len(result),p.h)
                self.writeHeadlineHelper(p)
        return result
    #@nonl
    #@-node:ekr.20060608094815:handleDocOnlyMode
    #@+node:ekr.20060608094815.1:isAnyDocPart
    def isAnyDocPart (self,s):

        if s.startswith('@doc'):
            return True
        elif not s.startswith('@'):
            return False
        else:
            return len(s) == 1 or s[1].isspace()
    #@nonl
    #@-node:ekr.20060608094815.1:isAnyDocPart
    #@+node:ekr.20050811153208:isSpecialDocPart
    def isSpecialDocPart (self,s,kind):

        '''Return True if s is a special doc part of the indicated kind.

        If kind is None, return True if s is any doc part.'''

        if s.startswith('@') and len(s) > 1 and s[1].isspace():
            if kind:
                i = g.skip_ws(s,1)
                result = g.match_word(s,i,kind)
            else:
                result = True
        elif not kind:
            result = g.match_word(s,0,'@doc') or g.match_word(s,0,'@')
        else:
            result = False

        return result
    #@nonl
    #@-node:ekr.20050811153208:isSpecialDocPart
    #@+node:ekr.20050811163802:isAnySpecialDocPart
    def isAnySpecialDocPart (self,s):

        for kind in (
            '@rst-markup',
            '@rst-option',
            '@rst-options',
        ):
            if self.isSpecialDocPart(s,kind):
                return True

        return False
    #@-node:ekr.20050811163802:isAnySpecialDocPart
    #@+node:ekr.20050811105438:removeLeoDirectives
    def removeLeoDirectives (self,lines):

        '''Remove all Leo directives, except within literal blocks.'''

        n = 0 ; result = []
        while n < len(lines):
            s = lines [n] ; n += 1
            if s.strip().endswith('::'):
                n, lit = self.skip_literal_block(lines,n-1)
                result.extend(lit)
            elif s.startswith('@') and not self.isAnySpecialDocPart(s):
                for key in self.leoDirectivesList:
                    if g.match_word(s,0,key):
                        # g.trace('removing %s' % s)
                        break
                else:
                    result.append(s)
            else:
                result.append(s)

        return result
    #@nonl
    #@-node:ekr.20050811105438:removeLeoDirectives
    #@+node:ekr.20050811105438.1:handleSpecialDocParts
    def handleSpecialDocParts (self,lines,kind,retainContents,asClass=None):

        # g.trace(kind,g.listToString(lines))

        result = [] ; n = 0
        while n < len(lines):
            s = lines [n] ; n += 1
            if s.strip().endswith('::'):
                n, lit = self.skip_literal_block(lines,n-1)
                result.extend(lit)
            elif self.isSpecialDocPart(s,kind):
                n, lines2 = self.getDocPart(lines,n)
                if retainContents:
                    result.extend([''])
                    if asClass:
                        result.extend(['.. container:: '+asClass, ''])
                        if 'literal' in asClass.split():
                            result.extend(['  ::', ''])
                        for l2 in lines2: result.append('    '+l2)
                    else:
                        result.extend(lines2)
                    result.extend([''])
            else:
                result.append(s)

        return result
    #@-node:ekr.20050811105438.1:handleSpecialDocParts
    #@+node:ekr.20050805162550.30:replaceCodeBlockDirectives
    def replaceCodeBlockDirectives (self,lines):

        '''Replace code-block directive, but not in literal blocks.'''

        n = 0 ; result = []
        while n < len(lines):
            s = lines [n] ; n += 1
            if s.strip().endswith('::'):
                n, lit = self.skip_literal_block(lines,n-1)
                result.extend(lit)
            else:
                i = g.skip_ws(s,0)
                if g.match(s,i,'..'):
                    i = g.skip_ws(s,i+2)
                    if g.match_word(s,i,'code-block'):
                        if 1: # Create a literal block to hold the code.
                            result.append('::\n')
                        else: # This 'annotated' literal block is confusing.
                            result.append('%s code::\n' % s[i+len('code-block'):])
                    else:
                        result.append(s)
                else:
                    result.append(s)

        return result
    #@nonl
    #@-node:ekr.20050805162550.30:replaceCodeBlockDirectives
    #@-node:ekr.20050811101550.1:writeBody & helpers
    #@+node:ekr.20050805162550.26:writeHeadline & helper
    def writeHeadline (self,p):

        '''Generate an rST section if options permit it.
        Remove headline commands from theheadlinefirst import 
        and never generate an rST section for @rst-option and @rst-options.'''

        docOnly             =  self.getOption('doc_only_mode')
        ignore              = self.getOption('ignore_this_headline')
        showHeadlines       = self.getOption('show_headlines')
        showThisHeadline    = self.getOption('show_this_headline')
        showOrganizers      = self.getOption('show_organizer_nodes')

        if (
            p == self.topNode or
            ignore or
            docOnly or # handleDocOnlyMode handles this.
            not showHeadlines and not showThisHeadline or
            # docOnly and not showOrganizers and not thisHeadline or
            not p.h.strip() and not showOrganizers or
            not p.b.strip() and not showOrganizers
        ):
            return

        self.writeHeadlineHelper(p)
    #@nonl
    #@+node:ekr.20060608102001:writeHeadlineHelper
    def writeHeadlineHelper (self,p):

        h = p.h.strip()

        # Remove any headline command before writing the 
        i = g.skip_id(h,0,chars='@-')
        word = h [:i]
        if word:
            # Never generate a section for @rst-option or @rst-options.
            if word in (self.getOption('option_prefix'),self.getOption('options_prefix')):
                return
            # Remove all other headline commands from the headline.
            for prefix in self.headlineCommands:
                if word == prefix:
                    h = h [len(word):].strip()
                    break

            # New in Leo 4.4.4.
            if word.startswith('@'):
                if self.getOption('strip_at_file_prefixes'):
                    for s in ('@auto','@file','@nosent','@thin',):
                        if g.match_word(word,0,s):
                            h = h [len(s):].strip()

        if not h.strip(): return

        if self.getOption('show_sections'):
            if self.getOption('generate_rst'):
                self.write('%s\n%s\n' % (h,self.underline(h,p)))
            else:
                self.write('\n%s\n' % h)
        else:
            self.write('\n**%s**\n\n' % h.replace('*',''))
    #@-node:ekr.20060608102001:writeHeadlineHelper
    #@-node:ekr.20050805162550.26:writeHeadline & helper
    #@+node:ekr.20050810083057:writeNode
    def writeNode (self,p):

        '''Format a node according to the options presently in effect.'''

        self.initCodeBlockString(p)
        self.scanAllOptions(p)

        if 0:
            g.trace('%24s code_mode %s' % (p.h,self.getOption('code_mode')))

        h = p.h.strip()

        if self.getOption('preformat_this_node'):
            self.http_addNodeMarker(p)
            self.writePreformat(p)
            p.moveToThreadNext()
        elif self.getOption('ignore_this_tree'):
            p.moveToNodeAfterTree()
        elif self.getOption('ignore_this_node'):
            p.moveToThreadNext()
        elif g.match_word(h,0,'@rst-options') and not self.getOption('show_options_nodes'):
            p.moveToThreadNext()
        else:
            self.http_addNodeMarker(p)
            self.writeHeadline(p)
            self.writeBody(p)
            p.moveToThreadNext()
    #@-node:ekr.20050810083057:writeNode
    #@+node:ekr.20071115061253:writePreformat
    def writePreformat (self,p):

        '''Write p's body text lines as if preformatted.

         ::

            line 1
            line 2 etc.
        '''

        # g.trace(p.h,g.callers())

        lines = p.b.split('\n')
        lines = [' '*4 + z for z in lines]
        lines.insert(0,'::\n')

        s = '\n'.join(lines)
        if s.strip():
            self.write('%s\n\n' % s)
    #@-node:ekr.20071115061253:writePreformat
    #@+node:ekr.20050805162550.23:writeTree
    def writeTree(self,p):

        '''Write p's tree to self.outputFile.'''

        self.scanAllOptions(p)

        # g.trace(self.getOption('generate_rst_header_comment'))

        if self.getOption('generate_rst'):
            if self.getOption('generate_rst_header_comment'):
                self.write(self.rstComment(
                    'rst3: filename: %s\n\n' % self.outputFileName))

        # We can't use an iterator because we may skip parts of the tree.
        p = p.copy() # Only one copy is needed for traversal.
        self.topNode = p.copy() # Indicate the top of this tree.
        after = p.nodeAfterTree()

        while p and p != after:
            self.writeNode(p)
    #@-node:ekr.20050805162550.23:writeTree
    #@-node:ekr.20050809074827:write methods
    #@+node:ekr.20050810083314:Utils
    #@+node:ekr.20051202070028:computeOutputFileName
    def computeOutputFileName (self,fileName):

        openDirectory = self.c.frame.openDirectory
        default_path = self.getOption('default_path')

        # if default_path:
            # default_path = g.os_path_finalize(default_path)

        # g.trace('default_path',default_path,'fileName',fileName)

        if default_path:
            path = g.os_path_finalize_join(self.path,default_path,fileName)
        elif self.path:
            path = g.os_path_finalize_join(self.path,fileName)
        elif openDirectory:
            path = g.os_path_finalize_join(self.path,openDirectory,fileName)
        else:
            path = g.os_path_finalize_join(fileName)

        return path
    #@nonl
    #@-node:ekr.20051202070028:computeOutputFileName
    #@+node:ekr.20050805162550.16:encode
    def encode (self,s):

        return g.toEncodedString(s,encoding=self.encoding,reportErrors=True)
    #@nonl
    #@-node:ekr.20050805162550.16:encode
    #@+node:ekr.20050805162550.20:report
    def report (self,name):

        if self.getOption('verbose'):

            name = g.os_path_finalize(name)

            g.es_print('wrote: %s' % (name),color="blue")
    #@nonl
    #@-node:ekr.20050805162550.20:report
    #@+node:ekr.20050810083856:rstComment
    def rstComment (self,s):

        return '.. %s' % s
    #@nonl
    #@-node:ekr.20050810083856:rstComment
    #@+node:ekr.20050805162550.19:underline
    def underline (self,s,p):

        '''Return the underlining string to be used at the given level for string s.'''

        u = self.getOption('underline_characters') #  '''#=+*^~"'`-:><_'''

        level = max(0,p.level()-self.toplevel)
        level = min(level+1,len(u)-1) # Reserve the first character for explicit titles.

        ch = u [level]

        # g.trace(self.toplevel,p.level(),level,repr(ch),p.h)

        n = max(4,len(s))

        return ch * n + '\n'
    #@nonl
    #@-node:ekr.20050805162550.19:underline
    #@+node:ekr.20050809080031:write
    def write (self,s):

        s = self.encode(s)

        self.outputFile.write(s)
    #@nonl
    #@-node:ekr.20050809080031:write
    #@-node:ekr.20050810083314:Utils
    #@+node:ekr.20050805162550.33:Support for http plugin
    #@+node:ekr.20050815091008.1:http_addNodeMarker
    def http_addNodeMarker (self,p):

        if (
            self.getOption('http_server_support') and
            self.getOption('generate_rst')
        ):
            self.nodeNumber += 1
            anchorname = "%s%s" % (self.getOption('node_begin_marker'),self.nodeNumber)
            s = "\n\n.. _%s:\n\n" % anchorname
            self.write(s)
            self.http_map [anchorname] = p.copy()
            if bwm_file: print >> bwm_file, "addNodeMarker", anchorname, p
    #@nonl
    #@-node:ekr.20050815091008.1:http_addNodeMarker
    #@+node:ekr.20050805162550.34:http_endTree & helpers
    # Was http_support_main

    def http_endTree (self,filename,p,justOneFile):

        '''Do end-of-tree processing to support the http plugin.'''

        if (
            self.getOption('http_server_support') and
            self.getOption('generate_rst')
        ):
            self.set_initial_http_attributes(filename)
            self.find_anchors(p)
            if justOneFile:
                self.relocate_references(p.self_and_subtree)

            g.es_print('html updated for http plugin',color="blue")

            if self.getOption('clear_http_attributes'):
                g.es_print("http attributes cleared")
    #@nonl
    #@+node:ekr.20050805162550.36:set_initial_http_attributes
    def set_initial_http_attributes (self,filename):

        f = open(filename)
        parser = htmlParserClass(self)

        for line in f.readlines():
            parser.feed(line)

        f.close()
    #@nonl
    #@-node:ekr.20050805162550.36:set_initial_http_attributes
    #@+node:ekr.20050805162550.38:find_anchors
    def find_anchors (self, p):

        '''Find the anchors in all the nodes.'''

        for p1, attrs in self.http_attribute_iter(p):
            html = mod_http.reconstruct_html_from_attrs(attrs)
            # g.trace(pprint.pprint(html))
            parser = anchor_htmlParserClass(self, p1)
            for line in html:
                try:
                    parser.feed(line)
                # bwm: changed to unicode(line)
                except:
                    line = ''.join([ch for ch in line if ord(ch) <= 127])
                    # filter out non-ascii characters.
                    # bwm: not quite sure what's going on here.
                    parser.feed(line)        
        # g.trace(g.dictToString(self.anchor_map,tag='anchor_map'))
    #@nonl
    #@-node:ekr.20050805162550.38:find_anchors
    #@+node:ekr.20050805162550.37:relocate_references
    #@+at 
    #@nonl
    # Relocate references here if we are only running for one file.
    # 
    # Otherwise we must postpone the relocation until we have processed all 
    # files.
    #@-at
    #@@c

    def relocate_references (self, iterator_generator):

        for p in iterator_generator():
            attr = mod_http.get_http_attribute(p)
            if not attr:
                continue
            # g.trace('before',p.h,attr)
            if bwm_file:
                print >> bwm_file
                print >> bwm_file, "relocate_references(1): Position, attr:"
                pprint.pprint((p, attr), bwm_file)
            http_lines = attr [3:]
            parser = link_htmlparserClass(self,p)
            for line in attr [3:]:
                try:
                    parser.feed(line)
                except:
                    line = ''.join([ch for ch in line if ord(ch) <= 127])
                    parser.feed(line)
            replacements = parser.get_replacements()
            replacements.reverse()
            if not replacements:
                continue
            if bwm_file:
                print >> bwm_file, "relocate_references(2): Replacements:"
                pprint.pprint(replacements, bwm_file)
            for line, column, href, href_file, http_node_ref in replacements:
                if bwm_file: 
                    print >> bwm_file, "relocate_references(3): line:", line, "Column:", column, "href:", href, "href_file:", href_file, "http_node_ref:", http_node_ref
                marker_parts = href.split("#")
                if len(marker_parts) == 2:
                    marker = marker_parts [1]
                    replacement = "%s#%s" % (http_node_ref,marker)
                    try:
                        attr [line + 2] = attr [line + 2].replace(
                            'href="%s"' % href,'href="%s"' % replacement)
                    except:
                        g.es("Skipped ", attr[line + 2])
                else:
                    filename = marker_parts [0]
                    try:
                        attr [line + 2] = attr [line + 2].replace(
                            'href="%s"' % href,'href="%s"' % http_node_ref)
                    except:
                        g.es("Skipped", attr[line+2])
        # g.trace('after %s\n\n\n',attr)
    #@nonl
    #@-node:ekr.20050805162550.37:relocate_references
    #@+node:ekr.20050805162550.35:http_attribute_iter
    def http_attribute_iter (self, p):
        """
        Iterator for all the nodes which have html code.
        Look at the descendents of p.
        Used for relocation.
        """

        for p1 in p.self_and_subtree():
            attr = mod_http.get_http_attribute(p1)
            if attr:
                yield (p1.copy(),attr)
    #@nonl
    #@-node:ekr.20050805162550.35:http_attribute_iter
    #@-node:ekr.20050805162550.34:http_endTree & helpers
    #@-node:ekr.20050805162550.33:Support for http plugin
    #@-others
#@nonl
#@-node:ekr.20050805162550.8:class rstClass
#@-others
#@nonl
#@-node:ekr.20050805162550:@thin rst3.py
#@-leo
www.java2java.com | Contact Us
Copyright 2009 - 12 Demo Source and Support. All rights reserved.
All other trademarks are property of their respective owners.