[Data] Improve readability of `RefBundle.__repr__`

[bveeramani]

Here's what it looks like right now for pandas blocks for example:

RefBundle(1 blocks,
  12 rows,
  schema=id: int64
-- schema metadata --
pandas: '{"index_columns": [], "column_indexes": [], "columns": [{"name":' + 176,
  owns_blocks=False,
  blocks=(
    0: 14 rows, 240 bytes, slice=BlockSlice(start_offset=0, end_offset=12)
  )
)

Why
You see reprs during interactive debugging. Janky reprs make it harder to debug.

Constraints:

• The implementation should work well with both pandas and Arrow blocks
• The implementation should be reasonably simple -- avoid anything too complicated

[bveeramani]

(@pushpavanthar might work on this)

[Rob12312368]

Would love to take a stab if still available. So bascially implementing a better repr so the output is better looking?

[pushpavanthar]

@bveeramani yes, I'll work on this.