[Bug]: Can't find join_keys between indirectly linked datasets

[gogonzo]

What happened?

join_keys with correct parent-child relationships are not easy to design, especially by the inexperienced developer. Indirectly linked datasets need to have the same parent. IMO this is to strict given that their implicit relationships are "indirect" anyway. Here is an example

jk <- join_keys(
  join_key("customers", keys = "id"),
  join_key("orders", keys = c("id")),
  join_key("products", keys = c("id")),
  join_key("product_components", keys = c("id")),
  join_key("customers", "orders", keys = c(id = "customer_id")),
  join_key("orders", "order_items", keys = c(id = "order_id")),
  join_key("order_items", "products", keys = c(product_id = "id")),  # no common parent
  #join_key("products", "order_items", keys = c(id = "product_id")), # is common parent
  join_key("products", "product_components", keys = c(id = "product_id"))
)

jk["order_items", "product_components"]
# NULL

I propose this change to make it easier for app developer as I couldn't figure out what is the problem with above join keys setup myself.