You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
370 lines
15 KiB
370 lines
15 KiB
"""
|
|
Python Markdown
|
|
|
|
A Python implementation of John Gruber's Markdown.
|
|
|
|
Documentation: https://python-markdown.github.io/
|
|
GitHub: https://github.com/Python-Markdown/markdown/
|
|
PyPI: https://pypi.org/project/Markdown/
|
|
|
|
Started by Manfred Stienstra (http://www.dwerg.net/).
|
|
Maintained for a few years by Yuri Takhteyev (http://www.freewisdom.org).
|
|
Currently maintained by Waylan Limberg (https://github.com/waylan),
|
|
Dmitry Shachnev (https://github.com/mitya57) and Isaac Muse (https://github.com/facelessuser).
|
|
|
|
Copyright 2007-2018 The Python Markdown Project (v. 1.7 and later)
|
|
Copyright 2004, 2005, 2006 Yuri Takhteyev (v. 0.2-1.6b)
|
|
Copyright 2004 Manfred Stienstra (the original version)
|
|
|
|
License: BSD (see LICENSE.md for details).
|
|
|
|
PRE-PROCESSORS
|
|
=============================================================================
|
|
|
|
Preprocessors work on source text before we start doing anything too
|
|
complicated.
|
|
"""
|
|
|
|
from . import util
|
|
import re
|
|
|
|
|
|
def build_preprocessors(md, **kwargs):
|
|
""" Build the default set of preprocessors used by Markdown. """
|
|
preprocessors = util.Registry()
|
|
preprocessors.register(NormalizeWhitespace(md), 'normalize_whitespace', 30)
|
|
preprocessors.register(HtmlBlockPreprocessor(md), 'html_block', 20)
|
|
preprocessors.register(ReferencePreprocessor(md), 'reference', 10)
|
|
return preprocessors
|
|
|
|
|
|
class Preprocessor(util.Processor):
|
|
"""
|
|
Preprocessors are run after the text is broken into lines.
|
|
|
|
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.
|
|
|
|
"""
|
|
def run(self, lines):
|
|
"""
|
|
Each subclass of Preprocessor should override the `run` method, which
|
|
takes the document as a list of strings split by newlines and returns
|
|
the (possibly modified) list of lines.
|
|
|
|
"""
|
|
pass # pragma: no cover
|
|
|
|
|
|
class NormalizeWhitespace(Preprocessor):
|
|
""" Normalize whitespace for consistent parsing. """
|
|
|
|
def run(self, lines):
|
|
source = '\n'.join(lines)
|
|
source = source.replace(util.STX, "").replace(util.ETX, "")
|
|
source = source.replace("\r\n", "\n").replace("\r", "\n") + "\n\n"
|
|
source = source.expandtabs(self.md.tab_length)
|
|
source = re.sub(r'(?<=\n) +\n', '\n', source)
|
|
return source.split('\n')
|
|
|
|
|
|
class HtmlBlockPreprocessor(Preprocessor):
|
|
"""Remove html blocks from the text and store them for later retrieval."""
|
|
|
|
right_tag_patterns = ["</%s>", "%s>"]
|
|
attrs_pattern = r"""
|
|
\s+(?P<attr>[^>"'/= ]+)=(?P<q>['"])(?P<value>.*?)(?P=q) # attr="value"
|
|
| # OR
|
|
\s+(?P<attr1>[^>"'/= ]+)=(?P<value1>[^> ]+) # attr=value
|
|
| # OR
|
|
\s+(?P<attr2>[^>"'/= ]+) # attr
|
|
"""
|
|
left_tag_pattern = r'^\<(?P<tag>[^> ]+)(?P<attrs>(%s)*)\s*\/?\>?' % \
|
|
attrs_pattern
|
|
attrs_re = re.compile(attrs_pattern, re.VERBOSE)
|
|
left_tag_re = re.compile(left_tag_pattern, re.VERBOSE)
|
|
markdown_in_raw = False
|
|
|
|
def _get_left_tag(self, block):
|
|
m = self.left_tag_re.match(block)
|
|
if m:
|
|
tag = m.group('tag')
|
|
raw_attrs = m.group('attrs')
|
|
attrs = {}
|
|
if raw_attrs:
|
|
for ma in self.attrs_re.finditer(raw_attrs):
|
|
if ma.group('attr'):
|
|
if ma.group('value'):
|
|
attrs[ma.group('attr').strip()] = ma.group('value')
|
|
else:
|
|
attrs[ma.group('attr').strip()] = ""
|
|
elif ma.group('attr1'):
|
|
if ma.group('value1'):
|
|
attrs[ma.group('attr1').strip()] = ma.group(
|
|
'value1'
|
|
)
|
|
else:
|
|
attrs[ma.group('attr1').strip()] = ""
|
|
elif ma.group('attr2'):
|
|
attrs[ma.group('attr2').strip()] = ""
|
|
return tag, len(m.group(0)), attrs
|
|
else:
|
|
tag = block[1:].split(">", 1)[0].lower()
|
|
return tag, len(tag)+2, {}
|
|
|
|
def _recursive_tagfind(self, ltag, rtag, start_index, block):
|
|
while 1:
|
|
i = block.find(rtag, start_index)
|
|
if i == -1:
|
|
return -1
|
|
j = block.find(ltag, start_index)
|
|
# if no ltag, or rtag found before another ltag, return index
|
|
if (j > i or j == -1):
|
|
return i + len(rtag)
|
|
# another ltag found before rtag, use end of ltag as starting
|
|
# point and search again
|
|
j = block.find('>', j)
|
|
start_index = self._recursive_tagfind(ltag, rtag, j + 1, block)
|
|
if start_index == -1:
|
|
# HTML potentially malformed- ltag has no corresponding
|
|
# rtag
|
|
return -1
|
|
|
|
def _get_right_tag(self, left_tag, left_index, block):
|
|
for p in self.right_tag_patterns:
|
|
tag = p % left_tag
|
|
i = self._recursive_tagfind(
|
|
"<%s" % left_tag, tag, left_index, block
|
|
)
|
|
if i > 2:
|
|
return tag.lstrip("<").rstrip(">"), i
|
|
return block.rstrip()[-left_index:-1].lower(), len(block)
|
|
|
|
def _equal_tags(self, left_tag, right_tag):
|
|
if left_tag[0] in ['?', '@', '%']: # handle PHP, etc.
|
|
return True
|
|
if ("/" + left_tag) == right_tag:
|
|
return True
|
|
if (right_tag == "--" and left_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 _stringindex_to_listindex(self, stringindex, items):
|
|
"""
|
|
Same effect as concatenating the strings in items,
|
|
finding the character to which stringindex refers in that string,
|
|
and returning the index of the item in which that character resides.
|
|
"""
|
|
items.append('dummy')
|
|
i, count = 0, 0
|
|
while count <= stringindex:
|
|
count += len(items[i])
|
|
i += 1
|
|
return i - 1
|
|
|
|
def _nested_markdown_in_html(self, items):
|
|
"""Find and process html child elements of the given element block."""
|
|
for i, item in enumerate(items):
|
|
if self.left_tag_re.match(item):
|
|
left_tag, left_index, attrs = \
|
|
self._get_left_tag(''.join(items[i:]))
|
|
right_tag, data_index = self._get_right_tag(
|
|
left_tag, left_index, ''.join(items[i:]))
|
|
right_listindex = \
|
|
self._stringindex_to_listindex(data_index, items[i:]) + i
|
|
if 'markdown' in attrs.keys():
|
|
items[i] = items[i][left_index:] # remove opening tag
|
|
placeholder = self.md.htmlStash.store_tag(
|
|
left_tag, attrs, i + 1, right_listindex + 1)
|
|
items.insert(i, placeholder)
|
|
if len(items) - right_listindex <= 1: # last nest, no tail
|
|
right_listindex -= 1
|
|
items[right_listindex] = items[right_listindex][
|
|
:-len(right_tag) - 2] # remove closing tag
|
|
else: # raw html
|
|
if len(items) - right_listindex <= 1: # last element
|
|
right_listindex -= 1
|
|
if right_listindex <= i:
|
|
right_listindex = i + 1
|
|
placeholder = self.md.htmlStash.store('\n\n'.join(
|
|
items[i:right_listindex]))
|
|
del items[i:right_listindex]
|
|
items.insert(i, placeholder)
|
|
return items
|
|
|
|
def run(self, lines):
|
|
text = "\n".join(lines)
|
|
new_blocks = []
|
|
text = text.rsplit("\n\n")
|
|
items = []
|
|
left_tag = ''
|
|
right_tag = ''
|
|
in_tag = False # flag
|
|
|
|
while text:
|
|
block = text[0]
|
|
if block.startswith("\n"):
|
|
block = block[1:]
|
|
text = text[1:]
|
|
|
|
if block.startswith("\n"):
|
|
block = block[1:]
|
|
|
|
if not in_tag:
|
|
if block.startswith("<") and len(block.strip()) > 1:
|
|
|
|
if block[1:4] == "!--":
|
|
# is a comment block
|
|
left_tag, left_index, attrs = "--", 2, {}
|
|
else:
|
|
left_tag, left_index, attrs = self._get_left_tag(block)
|
|
right_tag, data_index = self._get_right_tag(left_tag,
|
|
left_index,
|
|
block)
|
|
# keep checking conditions below and maybe just append
|
|
|
|
if data_index < len(block) and (self.md.is_block_level(left_tag) or left_tag == '--'):
|
|
text.insert(0, block[data_index:])
|
|
block = block[:data_index]
|
|
|
|
if not (self.md.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.rstrip().endswith(">") \
|
|
and self._equal_tags(left_tag, right_tag):
|
|
if self.markdown_in_raw and 'markdown' in attrs.keys():
|
|
block = block[left_index:-len(right_tag) - 2]
|
|
new_blocks.append(self.md.htmlStash.
|
|
store_tag(left_tag, attrs, 0, 2))
|
|
new_blocks.extend([block])
|
|
else:
|
|
new_blocks.append(
|
|
self.md.htmlStash.store(block.strip()))
|
|
continue
|
|
else:
|
|
# if is block level tag and is not complete
|
|
if (not self._equal_tags(left_tag, right_tag)) and \
|
|
(self.md.is_block_level(left_tag) or left_tag == "--"):
|
|
items.append(block.strip())
|
|
in_tag = True
|
|
else:
|
|
new_blocks.append(
|
|
self.md.htmlStash.store(block.strip())
|
|
)
|
|
continue
|
|
|
|
else:
|
|
new_blocks.append(block)
|
|
|
|
else:
|
|
items.append(block)
|
|
|
|
# Need to evaluate all items so we can calculate relative to the left index.
|
|
right_tag, data_index = self._get_right_tag(left_tag, left_index, ''.join(items))
|
|
# Adjust data_index: relative to items -> relative to last block
|
|
prev_block_length = 0
|
|
for item in items[:-1]:
|
|
prev_block_length += len(item)
|
|
data_index -= prev_block_length
|
|
|
|
if self._equal_tags(left_tag, right_tag):
|
|
# if find closing tag
|
|
|
|
if data_index < len(block):
|
|
# we have more text after right_tag
|
|
items[-1] = block[:data_index]
|
|
text.insert(0, block[data_index:])
|
|
|
|
in_tag = False
|
|
if self.markdown_in_raw and 'markdown' in attrs.keys():
|
|
items[0] = items[0][left_index:]
|
|
items[-1] = items[-1][:-len(right_tag) - 2]
|
|
if items[len(items) - 1]: # not a newline/empty string
|
|
right_index = len(items) + 3
|
|
else:
|
|
right_index = len(items) + 2
|
|
new_blocks.append(self.md.htmlStash.store_tag(
|
|
left_tag, attrs, 0, right_index))
|
|
placeholderslen = len(self.md.htmlStash.tag_data)
|
|
new_blocks.extend(
|
|
self._nested_markdown_in_html(items))
|
|
nests = len(self.md.htmlStash.tag_data) - \
|
|
placeholderslen
|
|
self.md.htmlStash.tag_data[-1 - nests][
|
|
'right_index'] += nests - 2
|
|
else:
|
|
new_blocks.append(
|
|
self.md.htmlStash.store('\n\n'.join(items)))
|
|
items = []
|
|
|
|
if items:
|
|
if self.markdown_in_raw and 'markdown' in attrs.keys():
|
|
items[0] = items[0][left_index:]
|
|
items[-1] = items[-1][:-len(right_tag) - 2]
|
|
if items[len(items) - 1]: # not a newline/empty string
|
|
right_index = len(items) + 3
|
|
else:
|
|
right_index = len(items) + 2
|
|
new_blocks.append(
|
|
self.md.htmlStash.store_tag(
|
|
left_tag, attrs, 0, right_index))
|
|
placeholderslen = len(self.md.htmlStash.tag_data)
|
|
new_blocks.extend(self._nested_markdown_in_html(items))
|
|
nests = len(self.md.htmlStash.tag_data) - placeholderslen
|
|
self.md.htmlStash.tag_data[-1 - nests][
|
|
'right_index'] += nests - 2
|
|
else:
|
|
new_blocks.append(
|
|
self.md.htmlStash.store('\n\n'.join(items)))
|
|
new_blocks.append('\n')
|
|
|
|
new_text = "\n\n".join(new_blocks)
|
|
return new_text.split("\n")
|
|
|
|
|
|
class ReferencePreprocessor(Preprocessor):
|
|
""" Remove reference definitions from text and store for later use. """
|
|
|
|
TITLE = r'[ ]*(\"(.*)\"|\'(.*)\'|\((.*)\))[ ]*'
|
|
RE = re.compile(
|
|
r'^[ ]{0,3}\[([^\]]*)\]:\s*([^ ]*)[ ]*(%s)?$' % TITLE, re.DOTALL
|
|
)
|
|
TITLE_RE = re.compile(r'^%s$' % TITLE)
|
|
|
|
def run(self, lines):
|
|
new_text = []
|
|
while lines:
|
|
line = lines.pop(0)
|
|
m = self.RE.match(line)
|
|
if m:
|
|
id = m.group(1).strip().lower()
|
|
link = m.group(2).lstrip('<').rstrip('>')
|
|
t = m.group(5) or m.group(6) or m.group(7)
|
|
if not t:
|
|
# Check next line for title
|
|
tm = self.TITLE_RE.match(lines[0])
|
|
if tm:
|
|
lines.pop(0)
|
|
t = tm.group(2) or tm.group(3) or tm.group(4)
|
|
self.md.references[id] = (link, t)
|
|
# Preserve the line to prevent raw HTML indexing issue.
|
|
# https://github.com/Python-Markdown/markdown/issues/584
|
|
new_text.append('')
|
|
else:
|
|
new_text.append(line)
|
|
|
|
return new_text # + "\n"
|
|
|