Full string match regex function doesn't prune row group

### What happens?

See repro for details.

### To Reproduce

```sql
memory D CREATE OR REPLACE TABLE r3 AS
SELECT
i::INTEGER AS id,
('k' || lpad(i::VARCHAR, 8, '0')) AS s
FROM range(0, 1000000) t(i)
ORDER BY id;

-- should scan only one row group
memory D EXPLAIN ANALYZE SELECT count(*) FROM r3 WHERE replace(s, 'z', 'z') = 'k00900000';
╭─ Summary ───────────╮
│ Total Time: 0.0107s │
╰─────────────────────╯
╭─ Ungrouped Aggregate ────────────────╮
│ 1 row 0µs │
╰───────────────────┒┎─────────────────╯
╭─ Table Scan ──────┚┖─────────────────╮
│ Table: memory.main.r3 │
│ Type: Sequential Scan │
│ Filters: │
│ "replace"(s, 'z', 'z') = 'k00900000' │
│ Row Groups Scanned: 9 / 9 │
│ 1 row 50.0ms │
╰──────────────────────────────────────╯

-- should scan only one row group
memory D EXPLAIN ANALYZE SELECT count(*) FROM r3 WHERE regexp_full_match(s, 'k00900000');
╭─ Summary ───────────╮
│ Total Time: 0.0101s │
╰─────────────────────╯
╭─ Ungrouped Aggregate ─────────────╮
│ 1 row 0µs │
╰─────────────────┒┎────────────────╯
╭─ Table Scan ────┚┖────────────────╮
│ Table: memory.main.r3 │
│ Type: Sequential Scan │
│ Filters: │
│ regexp_full_match(s, 'k00900000') │
│ Row Groups Scanned: 9 / 9 │
│ 1 row 40.0ms │
╰───────────────────────────────────╯
```

### OS:

macos

### DuckDB Version:

abb2fb974537a975ce03dd23266056606b1a0e0b

### DuckDB Client:

SQL

### Hardware:

_No response_

### Full Name:

dentiny

### Affiliation:

N/A

### Did you include all relevant configuration (e.g., CPU architecture, Linux distribution) to reproduce the issue?

- [x] Yes, I have

### Did you include all code required to reproduce the issue?

- [x] Yes, I have

### Did you include all relevant data sets for reproducing the issue?

Yes

Start by running the supplied SQL reproduction in DuckDB and compare EXPLAIN ANALYZE output for replace and regexp_full_match. Investigate the row-group pruning path used by these predicates. Done means the equivalent full-string match query scans only the matching row group instead of 9 / 9 while returning the expected count.

Written by the indexing model from the issue text.