support `skip_rows` for `CsvFormat`

### Is your feature request related to a problem or challenge?

Some tools generate CSVs that contain some extra data at the top or bottom, such as bank statement exports

```csv
"Account Name : REDACTED"
"Account Number : REDACTED"
"Date Range : 01/01/2023-12/31/2023"
Transaction Date,Posted Date,Category,Debit
2023-01-01,2023-01-01,Payment/Credit,100.0
```

I'd like a way to skip the first `n` rows

### Describe the solution you'd like

add a new option to `CsvFormat` for skipping rows

```rust
CsvFormat::default().with_skip_rows(2)
```

### Describe alternatives you've considered

none that I can think of.

### Additional context

https://github.com/GlareDB/glaredb/issues/2035

Start by locating CsvFormat and the existing CSV reader tests or entry points. Review how CsvFormat options are applied during CSV ingestion, then add coverage for the requested skip_rows behavior, including the example with leading metadata. Done means CsvFormat::default().with_skip_rows(2) skips the first two rows when reading a CSV.

Written by the indexing model from the issue text.