BUG: NIST query bugs

Edit: Here is a summary of 3 bugs uncovered by this bug report:

- [ ] `linename` parameter doesn't allow all possible examples allowed by the website. Hoverover claims the following, all of them should be supported and covered by test:
Examples of allowed spectra:
- [ ] Ar I

- [ ] Mg I-IV

- [ ] All spectra

- [ ] Fe I; Si IX,XI; Ni Co-like

- [ ] H-Ar I-II

- [ ] Mg Li-like; Al Li-like-Be-like

- [ ] Sc-Fe K-like-Ca-like

- [ ] 198Hg I

- [ ] `minwav` and `maxwav` are not mandatory on the website, we may want to mirror that behaviour

- [ ] parsing errors from the response should be done and appropriate exception is raised rather than the current `Exception: Result did not contain a table`

--------

Original issue text:

astroquery.nist does not return the table when multiple ions are set in the linename argument

The following code works:

```
import numpy as np
from astroquery.nist import Nist
import astropy.units as u

Nist.query(6500*u.AA, 6550*u.AA, linename="Al",wavelength_type='vac+air')
```

But this does not work:

```
import numpy as np
from astroquery.nist import Nist
import astropy.units as u

Nist.query(6500*u.AA, 6550*u.AA, linename="Al;Mg",wavelength_type='vac+air')
```

Start at the Nist.query entry point and reproduce the single-ion and multi-ion examples from the issue. Check the listed linename forms, behavior when minwav and maxwav are omitted, and how response parsing errors are surfaced; done means these cases are supported, covered by tests, and produce appropriate exceptions.

Written by the indexing model from the issue text.