[Enhancement] Friendly field names

[beechnut]

At present, you pass in fields that are just the codes listed in the Census reference handbook. This provides reliability, but it's not very Rubyish.

I think it's possible to let people specify field components by text, which are looked up in a YAML file / Hash to construct the field code.

Example:

field = Field.new(:total_population, :total, :white, :total)
field.code
# => 'P0010001'

@client.find( field, county: 'Suffolk', state: 'MA' )
# => [{"P0010001"=>"722023", "name"=>"Suffolk County", "state"=>"25", "county"=>"025"}] 

# or specify the table name directly, which returns the first field for that table
@client.find( :total_population, county: 'Suffolk', state: 'MA' )
# => [{"P0010001"=>"722023", "name"=>"Suffolk County", "state"=>"25", "county"=>"025"}] 

field_hh = Field.new(:household_type, :other_family)
field_hh.code
#=> 'P0180004'

Field Construction

Field codes appear to have a highly structured construction, but the principles behind it are not clearly laid out anywhere in the census docs. From cursory research, they look like:

P001A0001

• Table Type: P
• Table Indicator: 001 (in Census docs, it's just P1. Any table indicator is formatted with a %3d.
• (Optional) Race Indicator: A
• Field Code:

Or, as an object (looks like YAML a little...):

P: Population
  001: Total (Population)
    A: White
      0001: Total