markdown.py

SQLAlchemy. 经典的Python ORM框架。学习必看。

    def detabbed_fn(self, line) :
        """ An auxiliary method to be passed to _findHead """
        m = RE.regExp['tabbed'].match(line)
        if m:
            return m.group(4)
        else :
            return None


    def detectTabbed(self, lines) :

        return self._findHead(lines, self.detabbed_fn,
                              allowBlank = 1)


def print_error(string):
    """Print an error string to stderr"""
    sys.stderr.write(string +'\n')


def dequote(string) :
    """ Removes quotes from around a string """
    if ( ( string.startswith('"') and string.endswith('"'))
         or (string.startswith("'") and string.endswith("'")) ) :
        return string[1:-1]
    else :
        return string

"""
======================================================================
========================== CORE MARKDOWN =============================
======================================================================

This stuff is ugly, so if you are thinking of extending the syntax,
see first if you can do it via pre-processors, post-processors,
inline patterns or a combination of the three.
"""

class CorePatterns :
    """This class is scheduled for removal as part of a refactoring
        effort."""

    patterns = {
        'header':          r'(#*)([^#]*)(#*)', # # A title
        'reference-def' :  r'(\ ?\ ?\ ?)\[([^\]]*)\]:\s*([^ ]*)(.*)',
                           # [Google]: http://www.google.com/
        'containsline':    r'([-]*)$|^([=]*)', # -----, =====, etc.
        'ol':              r'[ ]{0,3}[\d]*\.\s+(.*)', # 1. text
        'ul':              r'[ ]{0,3}[*+-]\s+(.*)', # "* text"
        'isline1':         r'(\**)', # ***
        'isline2':         r'(\-*)', # ---
        'isline3':         r'(\_*)', # ___
        'tabbed':          r'((\t)|(    ))(.*)', # an indented line
        'quoted' :         r'> ?(.*)', # a quoted block ("> ...")
    }

    def __init__ (self) :

        self.regExp = {}
        for key in self.patterns.keys() :
            self.regExp[key] = re.compile("^%s$" % self.patterns[key],
                                          re.DOTALL)

        self.regExp['containsline'] = re.compile(r'^([-]*)$|^([=]*)$', re.M)

RE = CorePatterns()


class Markdown:
    """ Markdown formatter class for creating an html document from
        Markdown text """


    def __init__(self, source=None,
                 extensions=[],
                 extension_configs=None,
                 encoding=None,
                 safe_mode = True):
        """Creates a new Markdown instance.

           @param source: The text in Markdown format.
           @param encoding: The character encoding of <text>. """

        self.safeMode = safe_mode
        self.encoding = encoding
        self.source = source
        self.blockGuru = BlockGuru()
        self.registeredExtensions = []
        self.stripTopLevelTags = 1
        self.docType = ""

        self.preprocessors = [ HEADER_PREPROCESSOR,
                               LINE_PREPROCESSOR,
                               HTML_BLOCK_PREPROCESSOR,
                               LINE_BREAKS_PREPROCESSOR,
                               # A footnote preprocessor will
                               # get inserted here
                               REFERENCE_PREPROCESSOR ]


        self.postprocessors = [] # a footnote postprocessor will get
                                 # inserted later

        self.textPostprocessors = [] # a footnote postprocessor will get
                                     # inserted later                                 

        self.prePatterns = []
        

        self.inlinePatterns = [ DOUBLE_BACKTICK_PATTERN,
                                BACKTICK_PATTERN,
                                ESCAPE_PATTERN,
                                IMAGE_LINK_PATTERN,
                                IMAGE_REFERENCE_PATTERN,
                                REFERENCE_PATTERN,
                                LINK_ANGLED_PATTERN,
                                LINK_PATTERN,
                                AUTOLINK_PATTERN,
                                AUTOMAIL_PATTERN,
                                HTML_PATTERN,
                                ENTITY_PATTERN,
                                NOT_STRONG_PATTERN,
                                STRONG_EM_PATTERN,
                                STRONG_EM_PATTERN_2,
                                STRONG_PATTERN,
                                STRONG_PATTERN_2,
                                EMPHASIS_PATTERN,
                                EMPHASIS_PATTERN_2
                                # The order of the handlers matters!!!
                                ]

        self.registerExtensions(extensions = extensions,
                                configs = extension_configs)

        self.reset()


    def registerExtensions(self, extensions, configs) :

        if not configs :
            configs = {}

        for ext in extensions :

            extension_module_name = "mdx_" + ext

            try :
                module = __import__(extension_module_name)

            except :
                message(CRITICAL,
                        "couldn't load extension %s (looking for %s module)"
                        % (ext, extension_module_name) )
            else :

                if configs.has_key(ext) :
                    configs_for_ext = configs[ext]
                else :
                    configs_for_ext = []
                extension = module.makeExtension(configs_for_ext)    
                extension.extendMarkdown(self, globals())




    def registerExtension(self, extension) :
        """ This gets called by the extension """
        self.registeredExtensions.append(extension)

    def reset(self) :
        """Resets all state variables so that we can start
            with a new text."""
        self.references={}
        self.htmlStash = HtmlStash()

        HTML_BLOCK_PREPROCESSOR.stash = self.htmlStash
        REFERENCE_PREPROCESSOR.references = self.references
        HTML_PATTERN.stash = self.htmlStash
        ENTITY_PATTERN.stash = self.htmlStash
        REFERENCE_PATTERN.references = self.references
        IMAGE_REFERENCE_PATTERN.references = self.references

        for extension in self.registeredExtensions :
            extension.reset()


    def _transform(self):
        """Transforms the Markdown text into a XHTML body document

           @returns: A NanoDom Document """

        # Setup the document

        self.doc = Document()
        self.top_element = self.doc.createElement("span")
        self.top_element.appendChild(self.doc.createTextNode('\n'))
        self.top_element.setAttribute('class', 'markdown')
        self.doc.appendChild(self.top_element)

        # Fixup the source text
        text = self.source.strip()
        text = text.replace("\r\n", "\n").replace("\r", "\n")
        text += "\n\n"
        text = text.expandtabs(TAB_LENGTH)

        # Split into lines and run the preprocessors that will work with
        # self.lines

        self.lines = text.split("\n")

        # Run the pre-processors on the lines
        for prep in self.preprocessors :
            self.lines = prep.run(self.lines)

        # Create a NanoDom tree from the lines and attach it to Document


        buffer = []
        for line in self.lines :
            if line.startswith("#") :
                self._processSection(self.top_element, buffer)
                buffer = [line]
            else :
                buffer.append(line)
        self._processSection(self.top_element, buffer)
        
        #self._processSection(self.top_element, self.lines)

        # Not sure why I put this in but let's leave it for now.
        self.top_element.appendChild(self.doc.createTextNode('\n'))

        # Run the post-processors
        for postprocessor in self.postprocessors :
            postprocessor.run(self.doc)

        return self.doc


    def _processSection(self, parent_elem, lines,
                        inList = 0, looseList = 0) :

        """Process a section of a source document, looking for high
           level structural elements like lists, block quotes, code
           segments, html blocks, etc.  Some those then get stripped
           of their high level markup (e.g. get unindented) and the
           lower-level markup is processed recursively.

           @param parent_elem: A NanoDom element to which the content
                               will be added
           @param lines: a list of lines
           @param inList: a level
           @returns: None"""

        if not lines :
            return

        # Check if this section starts with a list, a blockquote or
        # a code block

        processFn = { 'ul' :     self._processUList,
                      'ol' :     self._processOList,
                      'quoted' : self._processQuote,
                      'tabbed' : self._processCodeBlock }

        for regexp in ['ul', 'ol', 'quoted', 'tabbed'] :
            m = RE.regExp[regexp].match(lines[0])
            if m :
                processFn[regexp](parent_elem, lines, inList)
                return

        # We are NOT looking at one of the high-level structures like
        # lists or blockquotes.  So, it's just a regular paragraph
        # (though perhaps nested inside a list or something else).  If
        # we are NOT inside a list, we just need to look for a blank
        # line to find the end of the block.  If we ARE inside a
        # list, however, we need to consider that a sublist does not
        # need to be separated by a blank line.  Rather, the following
        # markup is legal:
        #
        # * The top level list item
        #
        #     Another paragraph of the list.  This is where we are now.
        #     * Underneath we might have a sublist.
        #

        if inList :

            start, theRest = self._linesUntil(lines, (lambda line:
                             RE.regExp['ul'].match(line)
                             or RE.regExp['ol'].match(line)
                                              or not line.strip()))

            self._processSection(parent_elem, start,
                                 inList - 1, looseList = looseList)
            self._processSection(parent_elem, theRest,
                                 inList - 1, looseList = looseList)


        else : # Ok, so it's just a simple block

            paragraph, theRest = self._linesUntil(lines, lambda line:
                                                 not line.strip())

            if len(paragraph) and paragraph[0].startswith('#') :
                m = RE.regExp['header'].match(paragraph[0])
                if m :
                    level = len(m.group(1))
                    h = self.doc.createElement("h%d" % level)
                    parent_elem.appendChild(h)
                    for item in self._handleInlineWrapper2(m.group(2).strip()) :
                        h.appendChild(item)
                else :
                    message(CRITICAL, "We've got a problem header!")

            elif paragraph :

                list = self._handleInlineWrapper2("\n".join(paragraph))

                if ( parent_elem.nodeName == 'li'
                     and not (looseList or parent_elem.childNodes)):

                    #and not parent_elem.childNodes) :
                    # If this is the first paragraph inside "li", don't
                    # put <p> around it - append the paragraph bits directly
                    # onto parent_elem
                    el = parent_elem
                else :
                    # Otherwise make a "p" element
                    el = self.doc.createElement("p")
                    parent_elem.appendChild(el)

                for item in list :
                    el.appendChild(item)

            if theRest :
                theRest = theRest[1:]  # skip the first (blank) line

            self._processSection(parent_elem, theRest, inList)



    def _processUList(self, parent_elem, lines, inList) :
        self._processList(parent_elem, lines, inList,
                         listexpr='ul', tag = 'ul')

    def _processOList(self, parent_elem, lines, inList) :
        self._processList(parent_elem, lines, inList,
                         listexpr='ol', tag = 'ol')


    def _processList(self, parent_elem, lines, inList, listexpr, tag) :
        """Given a list of document lines starting with a list item,
           finds the end of the list, breaks it up, and recursively
           processes each list item and the remainder of the text file.

           @param parent_elem: A dom element to which the content will be added
           @param lines: a list of lines
           @param inList: a level
           @returns: None"""

        ul = self.doc.createElement(tag)  # ul might actually be '<ol>'
        parent_elem.appendChild(ul)

        looseList = 0

        # Make a list of list items
        items = []
        item = -1

        i = 0  # a counter to keep track of where we are

        for line in lines :

            loose = 0
            if not line.strip() :
                # If we see a blank line, this _might_ be the end of the list
                i += 1
                loose = 1

                # Find the next non-blank line
                for j in range(i, len(lines)) :
                    if lines[j].strip() :
                        next = lines[j]
                        break
                else :
                    # There is no more text => end of the list
                    break

                # Check if the next non-blank line is still a part of the list
                if ( RE.regExp['ul'].match(next) or
                     RE.regExp['ol'].match(next) or 
                     RE.regExp['tabbed'].match(next) ):
                    # get rid of any white space in the line
                    items[item].append(line.strip())
                    looseList = loose or looseList
                    continue
                else :
                    break # found end of the list

            # Now we need to detect list items (at the current level)
            # while also detabing child elements if necessary

            for expr in ['ul', 'ol', 'tabbed']:

                m = RE.regExp[expr].match(line)
                if m :
                    if expr in ['ul', 'ol'] :  # We are looking at a new item
                        if m.group(1) :
                            items.append([m.group(1)])
                            item += 1
                    elif expr == 'tabbed' :  # This line needs to be detabbed
                        items[item].append(m.group(4)) #after the 'tab'

                    i += 1
                    break
            else :
                items[item].append(line)  # Just regular continuation