[BUG](api): HNSW parameter validators accept bool values due to isinstance(p, int)

## Problem

In `chromadb/segment/impl/vector/hnsw_params.py`, the parameter validators use `isinstance(p, int)` to check that HNSW parameters are integers:

```python
param_validators: Dict[str, Validator] = {
"hnsw:construction_ef": lambda p: isinstance(p, int),
"hnsw:search_ef": lambda p: isinstance(p, int),
"hnsw:M": lambda p: isinstance(p, int),
"hnsw:num_threads": lambda p: isinstance(p, int),
"hnsw:resize_factor": lambda p: isinstance(p, (int, float)),
}
```

Since `bool` is a subclass of `int` in Python, `isinstance(True, int)` returns `True`. This means:
- `True` (which equals `1`) passes validation for `construction_ef`, `search_ef`, `M`, `num_threads`
- `False` (which equals `0`) also passes
- For `resize_factor`, `isinstance(True, (int, float))` is `True`

These are semantically invalid values for HNSW parameters. Setting `construction_ef=True` (1) or `M=True` (1) would produce a nearly unusable index.

## Proposed Fix

Add `and not isinstance(p, bool)` to the integer validators, or use `type(p) is int` instead of `isinstance(p, int)`.

## Affected Files

- `chromadb/segment/impl/vector/hnsw_params.py` (lines 12-16, 21-22)

Start in chromadb/segment/impl/vector/hnsw_params.py and inspect the param_validators entries for the HNSW parameters. Verify the existing validation behavior for bool, int, and float inputs, then update the validators so semantically invalid bool values are rejected while valid numeric types retain their intended behavior.

Written by the indexing model from the issue text.