unique vs dropUnique inconsistency when using predicate

[atlanteh]

Environment

Knex version: 3.0.1
Database + version: postgres
OS: all

Bug

1. Explain what kind of behaviour you are getting and how you think it should do
   When creating and dropping a unique constraint with a predicate the api doesn't behave the same way

2. Error message

3. Reduced test code, for example in https://npm.runkit.com/knex or if it needs real
   database connection to MySQL or PostgreSQL, then single file example which initializes
   needed data and demonstrates the problem.

knex.schema.table('my_table', (table) => {
    table.unique(['my_column'], {predicate: knex.whereRaw('"my_column" IS NOT NULL')})
});

Results in

create unique index "my_table_my_column_unique" on "my_table" ("my_column")
where
  "my_column" IS NOT NULL

And

knex.schema.table('my_table', (table) => {
    table.dropUnique(['my_column'])
})

results in

alter table "my_table"
drop constraint "my_table_my_column_unique"

which fails because there's no constraint, but an index by the name of "my_table_my_column_unique"