Plaid sync domain improvements (#2267)

Breaks our Plaid sync process out into more manageable classes. Notably, this moves the sync process to a distinct, 2-step flow:

1. Import stage - we first make API calls and import Plaid data to "mirror" tables
2. Processing stage - read the raw data, apply business rules, build internal domain models and sync balances

This provides several benefits:

- Plaid syncs can now be "replayed" without fetching API data again
- Mirror tables provide better audit and debugging capabilities
- Eliminates the "all or nothing" sync behavior that is currently in place, which is brittle

test/models/plaid_item_test.rb

@@ -5,11 +5,11 @@ class PlaidItemTest < ActiveSupport::TestCase
 
   setup do
     @plaid_item = @syncable = plaid_items(:one)
+    @plaid_provider = mock
+    Provider::Registry.stubs(:plaid_provider_for_region).returns(@plaid_provider)
   end
 
   test "removes plaid item when destroyed" do
-    @plaid_provider = mock
-    @plaid_item.stubs(:plaid_provider).returns(@plaid_provider)
     @plaid_provider.expects(:remove_item).with(@plaid_item.access_token).once
 
     assert_difference "PlaidItem.count", -1 do
@@ -18,8 +18,6 @@ class PlaidItemTest < ActiveSupport::TestCase
   end
 
   test "if plaid item not found, silently continues with deletion" do
-    @plaid_provider = mock
-    @plaid_item.stubs(:plaid_provider).returns(@plaid_provider)
     @plaid_provider.expects(:remove_item).with(@plaid_item.access_token).raises(Plaid::ApiError.new("Item not found"))
 
     assert_difference "PlaidItem.count", -1 do