Support multiple bloom filters

[hardbyte]

It may make sense to calculate multiple CLKs using different field sets for improved matching, blocking, matching with orgs who only have a subset of the fields, and most importantly for automated methods to determine the threshold.

Example

If the possible fields are Name, Phone, Address, Email we might calculate CLKs for:

• Name, Phone, Address, Email
• Name, Phone, Address
• Name, Address, Email
• Phone, Address, Email
• Name, Phone
• Name, Address
• Name, Email
• Phone, Address
• Phone, Email
• etc

Need to investigate further and determine what changes would be required in anonlink to support matching datasets which contain sets of different clks.

Aha! Link: https://csiro.aha.io/features/ANONLINK-78