Optimize set operations in ExplorationReport (#262)

`set.union` becomes slow significantly when set size large than 100k.
Test with
```python
from tqdm import tqdm
s = set()
for i in tqdm(range(200000)):
    s = s.union([i])
```
in comparison to
```python
from tqdm import tqdm
s = set()
for i in tqdm(range(200000)):
    s.update([i])
```

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **Performance Improvements**
- Enhanced the efficiency of accuracy value aggregation by modifying how
accuracy values are updated, potentially improving performance in
scenarios with large or frequently updated data.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Signed-off-by: zjgemi <[email protected]>

dpgen2/exploration/report/report_adaptive_lower.py

@@ -239,7 +239,7 @@ def record(
                 ii, md_f[ii], md_v[ii]
             )
             self.nframes += add_nframes
-            self.accur = self.accur.union(add_accur)
+            self.accur.update(add_accur)
             self.failed += add_failed
             coll_f += add_f
             coll_v += add_v