Multiply polars columns of number type with object type (which supports __mul__)

I have the following code:
import polars as pl

class Summary:
def __init__(self, value: float, origin: str):
self.value = value
self.origin = origin

def __repr__(self) -> str:
return f'Summary({self.value},{self.origin})'

def __mul__(self, x: float | int) -> 'Summary':
return Summary(self.value * x, self.origin)

def __rmul__(self, x: float | int) -> 'Summary':
return self * x

mapping = {
'CASH': Summary( 1, 'E'),
'ITEM': Summary(-9, 'A'),
'CHECK': Summary(46, 'A'),
}

df = pl.DataFrame({'quantity': [7, 4, 10], 'type': mapping.keys(), 'summary': mapping.values()})



Variable df looks like this
shape: (3, 3)
┌──────────┬───────┬───────────────┐
│ quantity ┆ type ┆ summary │
│ --- ┆ --- ┆ --- │
│ i64 ┆ str ┆ object │
╞══════════╪═══════╪═══════════════╡
│ 7 ┆ CASH ┆ Summary(1,E) │
│ 4 ┆ ITEM ┆ Summary(-9,A) │
│ 10 ┆ CHECK ┆ Summary(46,A) │
└──────────┴───────┴───────────────┘



So the summary column contains a Summary class object, which supports multiplication. I now want to multiply this column with the quantity column. However
df.with_columns(pl.col('quantity').mul(pl.col('summary')).alias('qty_summary'))



is failing with SchemaError: failed to determine supertype of i64 and object.


Is there a way to multiply these columns?