MongoDB text search and correct textScore?

Question

MongoDB

There is a collection (with a text index) that looks something like this:

[{id: 1, "items" : [ "milk", "bread", "fish" ]},
{id: 2, "items" : [ "milk", "milk", "milk" ]},
{id: 3, "items" : [ "cheese", "salt" ]},
{id: 4, "items" : [ "fish", "bread"]}]

We need to find all documents that have at least one element in the items array that contains the words "milk", "bread", and "fish". And sort the result, but not by the total number of occurrences of any search words (as the code below does), but by the number of simultaneous occurrences of the search words.

coll.find({$text:{"$search": "milk bread fish"}}, {score:{$meta:"textScore"}}).sort({score:{$meta:"textScore"}})

This code sorts the result like this:
[{id: 2, "items" : [ "milk", "milk", "milk" ], "score": 3.3},
{id: 1, "items" : [ "milk ", "bread", "fish" ], "score": 3.3},
{id: 4, "items" : [ "fish", "bread"], "score": 2.2}]
And I need:
[{ id: 1, "items" : [ "milk", "bread", "fish" ], "score" : 3},
{id: 4, "items" : [ "fish", "bread"], "score ": 2},
{id: 2, "items" : [ "milk", "milk", "milk" ], "score": 1}]