我有一个文件集，内容如下：

{
  gameId: '0001A',
  score: 40,
  name: 'Bob',
  city: 'London'
}

我正在尝试对我的文档运行聚合，它将为每个gameId输出以下视图：

{
  cities: [
    London: {
      totalScore: 500 // sum of the scores for all documents that have a city of London
      people: [
        'Bob',
        'Anna',
        'Sally',
        'Sue'
      ],
      peopleCount: 4 // a count of all people who also have the city of London
    },
    Barcelona: {
      totalScore: 400 // sum of the scores for all documents that have a city of Barcelona
      people: [
        'Tim',
        'Tina',
        'Amir'
      ], // names of all people who also have the city of Barcelona
      peopleCount: 3 // count of how many names appear
    },
  ]

我尝试过在聚合管道中使用$facet$和$bucket来实现这一点，但是这似乎不符合要求，因为$bucket/$bucketAuto似乎分别需要范围或桶的数量，然后$bucketAuto在对象中设置min和max的值。
现在，我可以像这样直接地对总人数、姓名和分数进行分组：

$group: {
  _id: '$gameId',
  totalScore: {
    $sum: '$score'
 },  
  uniqueClients: {
    $addToSet: '$name'
  }
},
$addFields: {
  uniqueClientCount: {
  $size: '$uniqueClients'
 }
}

我该怎么按城市细分呢？