markdown.py

SQLAlchemy. 经典的Python ORM框架。学习必看。

#!/usr/bin/env python

# The following constant specifies the name used in the usage
# statement displayed for python versions lower than 2.3.  (With
# python2.3 and higher the usage statement is generated by optparse
# and uses the actual name of the executable called.)

EXECUTABLE_NAME_FOR_USAGE = "python markdown.py"

SPEED_TEST = 0

"""
====================================================================
IF YOA ARE LOOKING TO EXTEND MARKDOWN, SEE THE "FOOTNOTES" SECTION
====================================================================

Python-Markdown
===============

Converts Markdown to HTML.  Basic usage as a module:

    import markdown
    html = markdown.markdown(your_text_string)

Started by [Manfred Stienstra](http://www.dwerg.net/).  Continued and
maintained  by [Yuri Takhteyev](http://www.freewisdom.org).

Project website: http://www.freewisdom.org/projects/python-markdown
Contact: yuri [at] freewisdom.org

License: GPL 2 (http://www.gnu.org/copyleft/gpl.html) or BSD

Version: 1.5a (July 9, 2006)

For changelog, see end of file
"""

import re, sys, os, random, codecs

# set debug level: 3 none, 2 critical, 1 informative, 0 all
(VERBOSE, INFO, CRITICAL, NONE) = range(4)

MESSAGE_THRESHOLD = CRITICAL

def message(level, text) :
    if level >= MESSAGE_THRESHOLD :
        print text


# --------------- CONSTANTS YOU MIGHT WANT TO MODIFY -----------------

# all tabs will be expanded to up to this many spaces
TAB_LENGTH = 4
ENABLE_ATTRIBUTES = 1
SMART_EMPHASIS = 1

# --------------- CONSTANTS YOU _SHOULD NOT_ HAVE TO CHANGE ----------

# a template for html placeholders
HTML_PLACEHOLDER_PREFIX = "qaodmasdkwaspemas"
HTML_PLACEHOLDER = HTML_PLACEHOLDER_PREFIX + "%dajkqlsmdqpakldnzsdfls"

BLOCK_LEVEL_ELEMENTS = ['p', 'div', 'blockquote', 'pre', 'table',
                        'dl', 'ol', 'ul', 'script', 'noscript',
                        'form', 'fieldset', 'iframe', 'math', 'ins',
                        'del', 'hr', 'hr/', 'style']

def is_block_level (tag) :
    return ( (tag in BLOCK_LEVEL_ELEMENTS) or
             (tag[0] == 'h' and tag[1] in "0123456789") )

"""
======================================================================
========================== NANODOM ===================================
======================================================================

The three classes below implement some of the most basic DOM
methods.  I use this instead of minidom because I need a simpler
functionality and do not want to require additional libraries.

Importantly, NanoDom does not do normalization, which is what we
want. It also adds extra white space when converting DOM to string
"""


class Document :

    def appendChild(self, child) :
        self.documentElement = child
        child.parent = self
        self.entities = {}

    def createElement(self, tag, textNode=None) :
        el = Element(tag)
        el.doc = self
        if textNode :
            el.appendChild(self.createTextNode(textNode))
        return el

    def createTextNode(self, text) :
        node = TextNode(text)
        node.doc = self
        return node

    def createEntityReference(self, entity):
        if entity not in self.entities:
            self.entities[entity] = EntityReference(entity)
        return self.entities[entity]

    def toxml (self) :
        return self.documentElement.toxml()

    def normalizeEntities(self, text) :

        pairs = [ ("&", "&"),
                  ("<", "&lt;"),
                  (">", "&gt;"),
                  ("\"", "&quot;")]


        for old, new in pairs :
            text = text.replace(old, new)
        return text

    def find(self, test) :
        return self.documentElement.find(test)

    def unlink(self) :
        self.documentElement.unlink()
        self.documentElement = None


class Element :

    type = "element"

    def __init__ (self, tag) :

        self.nodeName = tag
        self.attributes = []
        self.attribute_values = {}
        self.childNodes = []

    def unlink(self) :
        for child in self.childNodes :
            if child.type == "element" :
                child.unlink()
        self.childNodes = None

    def setAttribute(self, attr, value) :
        if not attr in self.attributes :
            self.attributes.append(attr)

        self.attribute_values[attr] = value

    def insertChild(self, position, child) :
        self.childNodes.insert(position, child)
        child.parent = self

    def removeChild(self, child) :
        self.childNodes.remove(child)

    def replaceChild(self, oldChild, newChild) :
        position = self.childNodes.index(oldChild)
        self.removeChild(oldChild)
        self.insertChild(position, newChild)

    def appendChild(self, child) :
        self.childNodes.append(child)
        child.parent = self

    def handleAttributes(self) :
        pass

    def find(self, test, depth=0) :
        """ Returns a list of descendants that pass the test function """
        matched_nodes = []
        for child in self.childNodes :
            if test(child) :
                matched_nodes.append(child)
            if child.type == "element" :
                matched_nodes += child.find(test, depth+1)
        return matched_nodes

    def toxml(self):
        if ENABLE_ATTRIBUTES :
            for child in self.childNodes:
                child.handleAttributes()
        buffer = ""
        if self.nodeName in ['h1', 'h2', 'h3', 'h4'] :
            buffer += "\n"
        elif self.nodeName in ['li'] :
            buffer += "\n "
        buffer += "<" + self.nodeName
        for attr in self.attributes :
            value = self.attribute_values[attr]
            value = self.doc.normalizeEntities(value)
            buffer += ' %s="%s"' % (attr, value)
        if self.childNodes or self.nodeName in ['blockquote']:
            buffer += ">"
            for child in self.childNodes :
                buffer += child.toxml()
            if self.nodeName == 'p' :
                buffer += "\n"
            elif self.nodeName == 'li' :
                buffer += "\n "
            buffer += "</%s>" % self.nodeName
        else :
            buffer += "/>"
        if self.nodeName in ['p', 'li', 'ul', 'ol',
                             'h1', 'h2', 'h3', 'h4'] :
            buffer += "\n"

        return buffer


class TextNode :

    type = "text"
    attrRegExp = re.compile(r'\{@([^\}]*)=([^\}]*)}') # {@id=123}

    def __init__ (self, text) :
        self.value = text        

    def attributeCallback(self, match) :
        self.parent.setAttribute(match.group(1), match.group(2))

    def handleAttributes(self) :
        self.value = self.attrRegExp.sub(self.attributeCallback, self.value)

    def toxml(self) :
        text = self.value
        if not text.startswith(HTML_PLACEHOLDER_PREFIX):
            if self.parent.nodeName == "p" :
                text = text.replace("\n", "\n   ")
            elif (self.parent.nodeName == "li"
                  and self.parent.childNodes[0]==self):
                text = "\n     " + text.replace("\n", "\n     ")
        text = self.doc.normalizeEntities(text)
        return text


class EntityReference:

    type = "entity_ref"

    def __init__(self, entity):
        self.entity = entity

    def handleAttributes(self):
        pass

    def toxml(self):
        return "&" + self.entity + ";"


"""
======================================================================
========================== PRE-PROCESSORS ============================
======================================================================

Preprocessors munge source text before we start doing anything too
complicated.

Each preprocessor implements a "run" method that takes a pointer to a list of lines of the document,
modifies it as necessary and returns either the same pointer or a
pointer to a new list.  Preprocessors must extend
markdown.Preprocessor.

"""


class Preprocessor :
    pass


class HeaderPreprocessor (Preprocessor):

    """
       Replaces underlined headers with hashed headers to avoid
       the nead for lookahead later.
    """

    def run (self, lines) :

        i = -1
        while i+1 < len(lines) :
            i = i+1
            if not lines[i].strip() :
                continue

            if lines[i].startswith("#") :
                lines.insert(i+1, "\n")

            if (i+1 <= len(lines)
                  and lines[i+1]
                  and lines[i+1][0] in ['-', '=']) :

                underline = lines[i+1].strip()

                if underline == "="*len(underline) :
                    lines[i] = "# " + lines[i].strip()
                    lines[i+1] = ""
                elif underline == "-"*len(underline) :
                    lines[i] = "## " + lines[i].strip()
                    lines[i+1] = ""

        #for l in lines :
        #    print l.encode('utf8')
        #sys.exit(0)

        return lines

HEADER_PREPROCESSOR = HeaderPreprocessor()

class LinePreprocessor (Preprocessor):
    """Deals with HR lines (needs to be done before processing lists)"""

    def run (self, lines) :
        for i in range(len(lines)) :
            if self._isLine(lines[i]) :
                lines[i] = "<hr />"
        return lines

    def _isLine(self, block) :
        """Determines if a block should be replaced with an <HR>"""
        if block.startswith("    ") : return 0  # a code block
        text = "".join([x for x in block if not x.isspace()])
        if len(text) <= 2 :
            return 0
        for pattern in ['isline1', 'isline2', 'isline3'] :
            m = RE.regExp[pattern].match(text)
            if (m and m.group(1)) :
                return 1
        else:
            return 0

LINE_PREPROCESSOR = LinePreprocessor()


class LineBreaksPreprocessor (Preprocessor):
    """Replaces double spaces at the end of the lines with <br/ >."""

    def run (self, lines) :
        for i in range(len(lines)) :
            if (lines[i].endswith("  ")
                and not RE.regExp['tabbed'].match(lines[i]) ):
                lines[i] += "<br />"
        return lines

LINE_BREAKS_PREPROCESSOR = LineBreaksPreprocessor()


class HtmlBlockPreprocessor (Preprocessor):
    """Removes html blocks from self.lines"""
    
    def _get_left_tag(self, block):
        return block[1:].replace(">", " ", 1).split()[0].lower()


    def _get_right_tag(self, left_tag, block):
        return block.rstrip()[-len(left_tag)-2:-1].lower()

    def _equal_tags(self, left_tag, right_tag):
        if left_tag in ['?', '?php', 'div'] : # handle PHP, etc.
            return True
        if ("/" + left_tag) == right_tag:
            return True
        elif left_tag == right_tag[1:] \
            and right_tag[0] != "<":
            return True
        else:
            return False

    def _is_oneliner(self, tag):
        return (tag in ['hr', 'hr/'])

    
    def run (self, lines) :
        new_blocks = []
        text = "\n".join(lines)
        text = text.split("\n\n")
        
        items = []
        left_tag = ''
        right_tag = ''
        in_tag = False # flag
        
        for block in text:
            if block.startswith("\n") :
                block = block[1:]

            if not in_tag:

                if block.startswith("<"):
                    
                    left_tag = self._get_left_tag(block)
                    right_tag = self._get_right_tag(left_tag, block)

                    if not (is_block_level(left_tag) \
                        or block[1] in ["!", "?", "@", "%"]):
                        new_blocks.append(block)
                        continue

                    if self._is_oneliner(left_tag):
                        new_blocks.append(block.strip())
                        continue
                        
                    if block[1] == "!":
                        # is a comment block
                        left_tag = "--"
                        right_tag = self._get_right_tag(left_tag, block)
                        # keep checking conditions below and maybe just append
                        
                    if block.rstrip().endswith(">") \
                        and self._equal_tags(left_tag, right_tag):
                        new_blocks.append(
                            self.stash.store(block.strip()))