Fetching contributors…
Cannot retrieve contributors at this time
169 lines (138 sloc) 5.94 KB
Base class for renderers.
import re
import sys
from mistletoe import block_token, span_token
class BaseRenderer(object):
Base class for renderers.
All renderers should ...
* ... define all render functions specified in self.render_map;
* ... be a context manager (by inheriting __enter__ and __exit__);
Custom renderers could ...
* ... add additional tokens into the parsing process by passing custom
tokens to super().__init__();
* ... add additional render functions by appending to self.render_map;
Suppose SomeRenderer inherits BaseRenderer, and fin is the input file.
The syntax looks something like this:
>>> from mistletoe import Document
>>> from some_renderer import SomeRenderer
>>> with SomeRenderer() as renderer:
... rendered = renderer.render(Document(fin))
See mistletoe.html_renderer for an implementation example.
Naming conventions:
* The keys of self.render_map should exactly match the class
name of tokens;
* Render function names should be of form: "render_" + the
"snake-case" form of token's class name.
render_map (dict): maps tokens to their corresponding render functions.
_extras (list): a list of custom tokens to be added to the
parsing process.
_parse_name = re.compile(r"([A-Z][a-z]+|[A-Z]+(?![a-z]))")
def __init__(self, *extras):
self.render_map = {
'Strong': self.render_strong,
'Emphasis': self.render_emphasis,
'InlineCode': self.render_inline_code,
'RawText': self.render_raw_text,
'Strikethrough': self.render_strikethrough,
'Image': self.render_image,
'Link': self.render_link,
'AutoLink': self.render_auto_link,
'EscapeSequence': self.render_escape_sequence,
'Heading': self.render_heading,
'SetextHeading': self.render_heading,
'Quote': self.render_quote,
'Paragraph': self.render_paragraph,
'CodeFence': self.render_block_code,
'BlockCode': self.render_block_code,
'List': self.render_list,
'ListItem': self.render_list_item,
'Table': self.render_table,
'TableRow': self.render_table_row,
'TableCell': self.render_table_cell,
'ThematicBreak': self.render_thematic_break,
'LineBreak': self.render_line_break,
'Document': self.render_document,
self._extras = extras
for token in extras:
if issubclass(token, span_token.SpanToken):
token_module = span_token
token_module = block_token
render_func = getattr(self, self._cls_to_func(token.__name__))
self.render_map[token.__name__] = render_func
self.footnotes = {}
def render(self, token):
Grabs the class name from input token and finds its corresponding
render function.
Basically a janky way to do polymorphism.
token: whose __class__.__name__ is in self.render_map.
return self.render_map[token.__class__.__name__](token)
def render_inner(self, token):
Recursively renders child tokens. Joins the rendered
strings with no space in between.
If newlines / spaces are needed between tokens, add them
in their respective templates, or override this function
in the renderer subclass, so that whitespace won't seem to
appear magically for anyone reading your program.
token: a branch node who has children attribute.
rendered = [self.render(child) for child in token.children]
return ''.join(rendered)
def __enter__(self):
Make renderer classes into context managers.
return self
def __exit__(self, exception_type, exception_val, traceback):
Make renderer classes into context managers.
Reset block_token._token_types and span_token._token_types.
def _cls_to_func(self, cls_name):
snake = '_'.join(map(str.lower, self._parse_name.findall(cls_name)))
return 'render_{}'.format(snake)
def _tokens_from_module(module):
Helper method; takes a module and returns a list of all token classes
specified in module.__all__. Useful when custom tokens are defined in a
separate module.
return [getattr(module, name) for name in module.__all__]
def render_raw_text(self, token):
Default render method for RawText. Simply return token.content.
return token.content
def __getattr__(self, name):
Provides a default render method for all tokens.
Any token without a custom render method will simply be rendered by
If name does not start with 'render_', raise AttributeError as normal,
for less magic during debugging.
This method would only be called if the attribute requested has not
been defined. Defined attributes will not be overridden.
I still think this is heavy wizardry.
Let me know if you would like this method removed.
if not name.startswith('render_'):
msg = '{cls} object has no attribute {name}'.format(cls=type(self).__name__, name=name)
raise AttributeError(msg).with_traceback(sys.exc_info()[2])
return self.render_inner