How to get struct datatype similar to value_counts() return in polars?

I am trying to combine two columns row wise and get something similar to what value_counts() returns in polars.

I have tried to use the to_struct but that didn’t give the result that I expected.

data:

import polars as pl

df = pl.DataFrame({
    "tags": ["c", "a", "b", "d"],
    "vals":[4,3,1,1] 
})

df.to_struct

<bound method DataFrame.to_struct of shape: (4, 2)
┌──────┬──────┐
│ tags ┆ vals │
│ ---  ┆ ---  │
│ str  ┆ i64  │
╞══════╪══════╡
│ c    ┆ 4    │
│ a    ┆ 3    │
│ b    ┆ 1    │
│ d    ┆ 1    │

Expected Result:

│ struct[2] │
╞═══════════╡
│ {"c",4}   │
├╌╌╌╌╌╌╌╌╌╌╌┤
│ {"a",3}   │
├╌╌╌╌╌╌╌╌╌╌╌┤
│ {"b",1}   │
├╌╌╌╌╌╌╌╌╌╌╌┤
│ {"d",1}   │

Not sure how can I do that as I have already tried to convert into struct datatype

>Solution :

to_struct is a method that needs to be called with ().