Using Blocks Extension to parse YAML and build an HTML element

[joapuiib]

Hello,
I'm trying to specify rubrics in Markdown.

My idea is to create a new Block Extension in order to parse a YAML file with the rubric definition and build a table.

/// rubric
levels:
  1: Excellent
  2: Good
  3: Acceptable
  4: Poor

criteria:
  - title: Clarity
    description: The writing is clear and easy to understand.
    levels:
      1: The writing is excellent.
      2: The writing is good.
      3: The writing is acceptable.
      4: The writing is poor.
  - title: ...
///

As for now, I've managed to define a RubricExtension and a Rubric Block.

class Rubric(Block):
    """Figure captions."""

    NAME = 'rubric'
    CLASSES = ''
    ARGUMENT = None
    OPTIONS = {
        'type': ['', type_html_identifier]
    }

    def __init__(self, length, tracker, md, config):
        """Initialize."""

        super().__init__(length, tracker, md, config)

    def on_markdown(self):
        return "raw"

    def on_create(self, parent):
        """Initialize."""

        el = etree.SubElement(parent, 'table')

        return el


class RubricExtension(BlocksExtension):
    """Rubric Extension."""

    def __init__(self, *args, **kwargs):
        """Initialize."""

        self.config = {}

        super().__init__(*args, **kwargs)

    def extendMarkdownBlocks(self, md, block_mgr):
        """Extend Markdown blocks."""

        config = self.getConfigs()
        block_mgr.register(Rubric, config)


def makeExtension(*args, **kwargs):
    """Return extension."""

    return RubricExtension(*args, **kwargs)

However, I'm having trouble reading the block content. I want to parse it and, then, construct the table.

Thanks in advance!

[facelessuser]

Sorry, I meant to take a look sooner. Let me try to make sure I understand your question.

You want to get all the content before the table is created? If this is true, you can only access the content in the on_end hook. So, as documented, all the content will be stored in the element you gave it. You can then remove the content from he element, process it and stick it back however you like.

Are you running into some kind of technical problem?

[joapuiib]

That's exactly what I did :)

def on_end(self, block):
    """Finalize."""
    self.parse_rubric(block.text)
    block.text = ""

    self.handle_levels(block)
    self.handle_criteria(block)

    return block

def parse_rubric(self, text):
    """Parse rubric."""

    self.data = yaml.safe_load(text)

---

Now, I'm running into a problem when I write an empty rubric Block.

/// rubric
///

def test_empty_rubric():
        extensions = ['rubric']
        extension_configs = {}
        md = markdown.Markdown(extensions=extensions, extension_configs=extension_configs)
    
        markdown_text = R'''
        /// rubric
        ///
        '''
    
        expected_html = R'''
        '''
    
        markdown_text = textwrap.dedent(markdown_text).strip()
        expected_html = textwrap.dedent(expected_html).strip()
    
>       html = md.convert(markdown_text)

tests/test_rubric.py:20: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
self = <rubric.Rubric object at 0x7fe89fbb5520>, text = None, length = 4

    def dedent(self, text, length=None):
        """Dedent raw text."""
    
        if length is None:
            length = self.md.tab_length
    
        min_length = float('inf')
>       for x in RE_INDENT.findall(text):
E       TypeError: expected string or bytes-like object

../../.local/share/hatch/env/virtual/pymdown-rubric-extension/93p7zB9Q/test.py3.8-default/lib/python3.8/site-packages/pymdownx/blocks/block.py:265: TypeError

I don't really understand the internals of how Markdown Extensions and Block Extensions work, so I don't know why this error raises.

I would appreciate a lot if you could give me some guidance.

---

The code is available at https://github.com/joapuiib/pymdown-rubric-extension/blob/fix/empty_rubric/rubric/__init__.py

(I can copy/paste it here if you prefer)

---

Thanks in advance!

[facelessuser]

Fix PR is here: #2587. I will try and get it released today.

[facelessuser]

Fix has been tagged for release in 10.14.3.

[joapuiib]

Hi @facelessuser, thanks for your detailed explanation!
I've read about raw blocks and indenting content, and now it makes more sense to me.

I had chosen raw mode because I wanted to read it as a YAML text, so I didn't want any Markdown processor to change the content, but I haven't really tried other types.

After using it, it seemed to work nicely without indenting. Also, as I did not indent it, I could write Markdown text in the YAML, and it does render it correctly!

Example at https://fpmislata-daw1-projecte.github.io/projecte-daw1/avaluacio/rubriques/ed/:

Rubric definition at: https://github.com/fpmislata-daw1-projecte/projecte-daw1/blob/main/docs/avaluacio/rubriques/_ed/estrategies.yml

...
  levels:
    - Incorporació correcta de branques de característiques a
      la branca de desenvolupament (`develop`) una vegada completades,

So, do you think this approach is good?

---

As for now, the only thing I don't love about it is specifying the rubric in YAML format. It's easy to make errors in the syntax due to indentation.

I've thought of parsing it as some kind of Markdown list syntax to make it easier to write.

[facelessuser]

So, do you think this approach is good?

If you want to parse Markdown, why not use Markdown style constructs instead of YAML?

I've thought of parsing it as some kind of Markdown list syntax to make it easier to write.

I probably would have done something like this. You could use Markdown lists and such in normal block mode and just gather the info from the HTML elements and destroy them and populate your table. I imagine you will run into YAML issues at some point while mixed with Markdown, but maybe I'm wrong 🤷🏻.

Sorry it took so long for me to get back.

[joapuiib]

Thanks for your feedback 😄!

I probably would have done something like this. You could use Markdown lists and such in normal block mode and just gather the info from the HTML elements and destroy them and populate your table.

I'll try to use this approach!