Expr.replace and Expr.replace_strict set "NO_DEFAULT" as a value

[henryharbeck]

Checks

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of Polars.

Reproducible example

df = pl.DataFrame({"a": ["a", "b", "c"]})
df.select(pl.col("a").replace("b"))
# shape: (3, 1)
# ┌────────────┐
# │ a          │
# │ ---        │
# │ str        │
# ╞════════════╡
# │ a          │
# │ NO_DEFAULT │
# │ c          │
# └────────────┘


df_2 = pl.DataFrame({"a": ["b", "b", "b"]})
df_2.select(pl.col("a").replace_strict("b"))
# shape: (3, 1)
# ┌────────────┐
# │ a          │
# │ ---        │
# │ str        │
# ╞════════════╡
# │ NO_DEFAULT │
# │ NO_DEFAULT │
# │ NO_DEFAULT │
# └────────────┘

Log output

No response

Issue description

NO_DEFAULT should not be set as a string value

Expected behavior

Raise that new must be specified unless old is a mapping

Installed versions

main branch

[henryharbeck]

I am keen to pick this up, as I noticed this while working on a PR for #17220