[Python] python/Jupyter-lab: How to use built-in JSON visualization functionality for a...

IPython/Jupyter has built-in functionality to visualise JSON structures as an interactive tree where you can expand and collapse nodes. I like using it for lists and dictionaries, but now I want to use it for visualising the content of dataclasses also:

import IPython
from dataclasses import dataclass
import dataclasses

@dataclass
class TestDC:
id: str
val: int

_ipy = IPython.get_ipython()
_json_formatter = _ipy.display_formatter.formatters["application/json"]
_json_formatter.for_type(dict, lambda obj: obj)
_json_formatter.for_type(list, lambda obj: obj)
_json_formatter.for_type(TestDC, lambda obj: dataclasses.asdict(obj))

dc1 = TestDC('a', 1)
dc2 = TestDC('b', 2)


In Jupyter-lab, this works fine when visualising the content of a single TestDC object,



but when I try to visualise the content of a list of TestDC objects, like [dc1, dc2], an error occurs: ValueError: Can't clean for JSON: TestDC(id='a', val=1).

I understand that there's no automatic JSON serialisation of dataclasses, but I don't understand why the built-in visualisation functionality is able to process/visualise the content of a single TestDC object and also nested list and dictionary objects, but runs into troubles when processing a list of objects, for which it knows how to visualise them individually...

What am I missing here?

Continue reading...