[FEA] Support unique expression (polars)

[beckernick]

The unique expression is useful for things like getting the unique number of {customers, accounts, devices, transactions, etc.} in a dataset.

import polars as pl
from functools import partial
from cudf_polars.callback import execute_with_cudf

ldf = pl.DataFrame({"a": [0,0,1,2]}).lazy()

print(ldf.select(pl.col("a").unique()).collect())
print(ldf.select(pl.col("a").unique()).collect(post_opt_callback=partial(execute_with_cudf, raise_on_fail=True)))
shape: (3, 1)
┌─────┐
│ a   │
│ --- │
│ i64 │
╞═════╡
│ 0   │
│ 1   │
│ 2   │
└─────┘
---------------------------------------------------------------------------
ComputeError                              Traceback (most recent call last)
Cell In[18], line 8
      5 ldf = pl.DataFrame({"a": [0,0,1,2]}).lazy()
      7 print(ldf.select(pl.col("a").unique()).collect())
----> 8 print(ldf.select(pl.col("a").unique()).collect(post_opt_callback=partial(execute_with_cudf, raise_on_fail=True)))

File [/raid/nicholasb/miniconda3/envs/all_cuda-122_arch-x86_64/lib/python3.11/site-packages/polars/lazyframe/frame.py:1942](http://10.117.23.184:8882/lab/tree/raid/raid/nicholasb/miniconda3/envs/all_cuda-122_arch-x86_64/lib/python3.11/site-packages/polars/lazyframe/frame.py#line=1941), in LazyFrame.collect(self, type_coercion, predicate_pushdown, projection_pushdown, simplify_expression, slice_pushdown, comm_subplan_elim, comm_subexpr_elim, cluster_with_columns, no_optimization, streaming, background, _eager, **_kwargs)
   1939 # Only for testing purposes atm.
   1940 callback = _kwargs.get("post_opt_callback")
-> 1942 return wrap_df(ldf.collect(callback))

ComputeError: 'cuda' conversion failed: NotImplementedError: No handler for Expr function node with name='unique'

[lithomas1]

@wence- has this implemented locally I think
wence-@acf0e2f

[wence-]

@wence- has this implemented locally I think wence-@acf0e2f

Yes, thanks for the reminder.