Allow flexible input spectral types when comparing to BEAST fits

[petiay]

To build on the handy - compare_spec_type.py - it would be good to allow for greater flexibility in the input spectral types to which BEAST fits are compared. Spectral classification uncertainties or distinguishing characteristics can be taken into account as they may impact the range of stellar properties in the comparison. For example, the code may benefit from incorporating:

• Temperature sequence uncertainties: The code can allow the BEAST fits for log(Teff) to be compared to a coarse spectral designation when there is no precise classification of the stellar spectrum in existing literature. Such stars may be listed as, say, 'B1-5', to indicate that the star can be early or late B-type, where the uncertainty may be due to reduced metallicity or low luminosity (see Sec. 3.2.1 in Evans et al. (2014)).
• Luminosity class uncertainties: In low or moderate S/N regimes it may be hard to reliably assign a luminosity class due to difficulties in determining the absolute magnitude. The luminosity class is then given in parentheses. (It is not immediate clear how/if this uncertainties can be quantified).
• The effects of qualifiers: Some luminosity type classifications may include qualifiers (such as Ia, Ib, or Iab to indicate the luminosity of supergiants), and spectral types may include other qualifiers (say, 'e' indicating emission lines are present).
• Other peculiarities: The published spectral type may include slashes, dashes, colons, etc. to indicate other types of uncertainties. A dash '-' indicates the star is between spectral types, and a slash '/' is used when the star is either of two spectral types.