Support vscode details tags

[civvic]

Running this in GitHub or any Jupyter flavor with the usual frontends:

<details open>

- id: 'msg_01J9h95kubV9fAA66i6q9kzD'
- content: [{'text': "Hello Vic, it's nice to meet you. How can I assist you today?", 'type': 'text'}]
- model: 'claude-3-5-sonnet-20240620'
- role: 'assistant'
- stop_reason: 'end_turn'
- stop_sequence: None
- type: 'message'
- usage: {'input_tokens': 19, 'output_tokens': 21}

</details>

you get:

• id: 'msg_01J9h95kubV9fAA66i6q9kzD'
• content: [{'text': "Hello Vic, it's nice to meet you. How can I assist you today?", 'type': 'text'}]
• model: 'claude-3-5-sonnet-20240620'
• role: 'assistant'
• stop_reason: 'end_turn'
• stop_sequence: None
• type: 'message'
• usage: {'input_tokens': 19, 'output_tokens': 21}

As expected. But running the same cell in VSCode 1.91.1 we get an empty usage field:

• id: 'msg_01J9h95kubV9fAA66i6q9kzD'
• content: [{'text': "Hello Vic, it's nice to meet you. How can I assist you today?", 'type': 'text'}]
• model: 'claude-3-5-sonnet-20240620'
• role: 'assistant'
• stop_reason: 'end_turn'
• stop_sequence: None
• type: 'message'
• usage:

with this generated HTML:

<details open="">
<ul>
...
<li>content: [{'text': "Hello Vic, it's nice to meet you. How can I assist you today?", 'type': 'text'}]</li>
...
<li>type: 'message'</li>
<li input_tokens:="" 19,="" output_tokens:="" 21="">usage:</li>
</ul>
</details>

This is VSCode frontend being really weird (a bug, really), The <detail> Details disclosure widget seems well supported by all browsers except VSCode notebooks.

I can open an issue with VSCode, or just add two characters to 00_core.ipnb Message._repr_markdown:

def _repr_markdown_(self:(Message)):
    det = '\n- '.join(f'{k}: `{v}`' for k,v in self.model_dump().items())
    return f"""{contents(self)}

<details>

- {det}

</details>"""

to get this:

• id: 'msg_01J9h95kubV9fAA66i6q9kzD'
• content: [{'text': "Hello Vic, it's nice to meet you. How can I assist you today?", 'type': 'text'}]
• model: 'claude-3-5-sonnet-20240620'
• role: 'assistant'
• stop_reason: 'end_turn'
• stop_sequence: None
• type: 'message'
• usage: {'input_tokens': 19, 'output_tokens': 21}

that IMHO is more legible, equally robust and compatible with this and future versions of VSCode.