Add constants/enums for field values

[Mr0grog]

We have a (semi) standard set of values that are available for each of:

• Race/Ethnicity
• Gender
• Comorbidities/Underlying Conditions
• Transmission Categories

(See the data model docs for more on this. Note that age does not have standardized values like this because the categories are so different across counties that we can’t meaningfully bucket them.)

Rather than having strings that could contain typos for these all over the codebase, we should really have a set of constants or (better yet) enums for them, e.g:

# This part should be in a centralized module somewhere.
from enum import Enum
class RaceEthnicity(Enum):
    african_american = "African_Amer"
    asian = "Asian"
    latinx = "Latinx_or_Hispanic"
    native_american = "Native_Amer"
    pacific_islander = "Pacific_Islander"
    white = "White"
    multiple = "Multiple_Race"
    other = "Other"
    unknown = "Unknown"

# Elsewhere in some scraper code...
# This example from Marin (https://github.com/sfbrigade/data-covid19-sfbayarea/blob/5fb99a51fa89ddc3ffacf65c3ae1c1fda19e75e9/covid19_sfbayarea/data/marin.py#L190-L206)

# Map the county's naming scheme to ours:
mapping = {
    'american indian/alaska native': RaceEthnicity.native_american,
    'asian': RaceEthnicity.asian,
    'black/african american': RaceEthnicity.african_american,
    'hispanic/latinx': RaceEthnicity.latinx,
    'multiracial': RaceEthnicity.multiple,
    'native hawaiian/pacific islander': RaceEthnicity.pacific_islander,
    'unknown': RaceEthnicity.unknown,
    'other': RaceEthnicity.other,
    'white': RaceEthnicity.white,
}
data = get_demographic_totals(api, 'Race')
return {mapping[row['grouping'].lower()]: int(row['cumulative'])
        for row in data}

This way, mypy can check that we haven’t made any typos and our values are consistent across the codebase.