Rename un/merge, add unexplode

[martindurant]

No description provided.

[martindurant]

@jpivarski : my implementation of the opposite of explode. We could still call this "unexplode", maybe.

In [1]: import polars as pl

In [2]: import akimbo.polars

In [3]: df = pl.DataFrame({"x": [1, 1, 1, 2, 1, 3, 3, 1], 'y':  [1, 1, 1, 2, 1, 3, 3, 1], 'z':  [1, 1, 1, 2, 1, 3, 3, 2]})

In [4]: df.ak.group_lists("x")
Out[4]:
shape: (3, 2)
┌─────┬─────────────────────────┐
│ x   ┆ grouped                 │
│ --- ┆ ---                     │
│ i64 ┆ list[struct[2]]         │
╞═════╪═════════════════════════╡
│ 1   ┆ [{1,1}, {1,1}, … {1,2}] │
│ 2   ┆ [{2,2}]                 │
│ 3   ┆ [{3,3}, {3,3}]          │
└─────┴─────────────────────────┘

In [5]: df.ak.group_lists("x", "y")
Out[5]:
shape: (3, 3)
┌─────┬─────┬───────────────────┐
│ x   ┆ y   ┆ grouped           │
│ --- ┆ --- ┆ ---               │
│ i64 ┆ i64 ┆ list[struct[1]]   │
╞═════╪═════╪═══════════════════╡
│ 1   ┆ 1   ┆ [{1}, {1}, … {2}] │
│ 2   ┆ 2   ┆ [{2}]             │
│ 3   ┆ 3   ┆ [{3}, {3}]        │
└─────┴─────┴───────────────────┘

[martindurant]

A better algorithm for unexplode can probably be found in awkward with or without numba - but this will do for now.