Older versions do not have correct numpy and pandas version constraints

### Solution to issue cannot be found in the documentation.

- [x] I checked the documentation.

### Issue

Not sure if this is the right place to raise this issue; apologies if not.

The conda-forge version constraints for numpy and pandas on some older versions of `bqplot` do not match the constraints on pypi.org. For example, 0.12.37 has

```yaml
dependencies:
- numpy >=1.10.4
- pandas
```

While on PyPI:

```json
"requires_dist": [
"numpy (<2.0.0,>=1.10.4)",
"pandas (<2.0.0,>=1.0.0)"
]
```

Compare the latest, 0.12.45, which has:

```yaml
dependencies:
- numpy >=1.10.4
- pandas >=1.0.0,<3.0.0
```

This means that the resolver thinks older versions of this feedstock are compatible with the latest `pandas` (3.x) and will try to select them in preference to the latest version, causing a cascade of dependency spaghetti, instead of identifying that it needs `pandas<3`.

### Installed packages

```shell
N/A
```

### Environment info

```shell
N/A
```