feat(python): allow list/tuple lit values

[alexander-beedie]

Closes: #6608 and #7862.
Also references: #6596.

This has come up a few times, and it appears to be a straightforward "quality of life" addition; allows for 1D/2D list/tuple literals to initialise lit like any other literals:

• lit( [1,2,3] ) → lit( Series([1,2,3]) )
• lit( [[1,2],[2,3]] ) → lit( Series([[1,2],[2,3]]) )

Example:

df = pl.DataFrame( {'a': [3,4]} )
df.select(
    result = pl.lit( [1,2] ).append( pl.col('a') )
)
# shape: (4, 1)
# ┌────────┐
# │ result │
# │ ---    │
# │ i64    │
# ╞════════╡
# │ 1      │
# │ 2      │
# │ 3      │
# │ 4      │
# └────────┘

[m-legrand]

Given that list[...] is a valid column type, this breaks the intuitive consistency of the api:

df = pl.DataFrame({x: [1, 2], y=[[1], [2]]})
df.filter(pl.col("x") == pl.lit(1))  # works
df.filter(pl.col("y") == pl.lit([1]))  # doesn't work

Not sure this straightforward quality of life was worth it.
The consistent, intuitive api is really one of the core value-add of Polars vs e.g. Pandas (along with perf, of course), I'm not sure changes like this are making it any favour :/