Complex cell attributes support in Pandoc Jupyter Markdown

[kiwi0fruit]

Sometimes Jupyter needs to have complex cell attributes like in celltags extension (that would be merged into jupyterlab in the near future):

{
    "tags": [
        "test",
        "main"
    ]
}

But there is no info in Creating Jupyter notebooks with pandoc section of Pandoc manual on how to add such cell metadata.

If it's possible to do then this issue is about adding info how to do it to the manual.

If it's not possible to do then this issue is about adding such functionality to Pandoc.

UPD: this actually adds metadata to the output ipynb document:

``` {.python .cell .code dic="{\"main\":\"2\",\"test\":\"1\"}" tags="[\"test\",\"main\"]"}
import pandas as pd
import numpy as np
df = pd.DataFrame(np.random.random(16).reshape(4, 4))
df
```

Title
=====

Text and tbl. 2

and corresponds to the following Cell Metadata:

{
    "tags": [
        "test",
        "main"
    ],
    "dic": {
        "test": "1",
        "main": "2"
    }
}

[kiwi0fruit]

By the way. The way cell metadata is stored can be more user-friendly: store json in yaml. Like in this example:

::: {.cell .markdown .cellmeta}
```yaml
slideshow:
  slide_type: slide
```
# Header

* list
* list
:::

Or

::: {.cell .code .cellmeta}
```yaml
slideshow:
  slide_type: slide
```
```python
print(123)
```
:::

.cellmeta class means that the first yaml code block inside div stores cell metadata.