mongo db. Group documents in array, score them count and sum and add new field

I have two collections:

> Collections with some actions:

[
{
"id": 0,
"action": "qwe",
"timestamp": "2022-11-20T23:59:59",
"user": "test"
},
{
"id": 1,
"action": "zwe",
"timestamp": "2022-11-20T11:22:33",
"user": "test"
},
{
"id": 1,
"action": "qwe",
"timestamp": "2022-11-20T11:22:33",
"user": "test"
}
]

> And collection with score of this actions:

[
{
"action": "qwe",
"score": 5
},
{
"action": "zwe",
"score": 3
},
{
"action": "qwe",
"score": 5
}
]

> I need achieve next results:

[
{
"user": "test",
"actions": [
{
"action":"qwe",
"score":5,
"timestamp":"2022-11-17T12:55:34.804+00:00"
},
{
"action":"zwe",
"score":3,
"timestamp":"2022-11-17T12:55:34.804+00:00"
},
{
"action":"qwe",
"score":5,
"timestamp":"2022-11-18T12:51:11.804+00:00"
}
],
"actions_total": [
{
"action": "qwe",
"count": 2,
"total_score": 10
},
{
"action": "zwe",
"count": 1,
"total_score": 3
}
],
"total_score": 13
}
]

> I write next aggregation:

[
{
$match: {
timestamp: {
$gte: ISODate(
'2022-11-17T00:00:00'
),
$lte: ISODate(
'2022-11-20T23:59:59'
)
}
}
},
{
$lookup: {
from: 'actions_score',
localField: 'action',
foreignField: 'action',
as: 'score'
}
},
{
$unwind: {
path: "$score"
}
},
{
$project: {
'username': true,
'action': true,
'score': '$score.score',
'timestamp': true,
'role_id': true,
'_id': false
}
},
{
$group: {
_id: '$username',
'username': {
$first: '$username'
},
'role_id': {
$first: '$role_id'
},
'actions': {
$addToSet: '$$ROOT'
}
}
},
{
$addFields: {
'total_score': {
"$sum": "$actions.score"
}
}
},
{
$project: {
_id: false,
'actions.username': false,
'actions.role_id': false
}
}
]

> And i get next result:

[
{
"user": "test",
"actions": [
{
"action":"qwe",
"score":5,
"timestamp":"2022-11-17T12:55:34.804+00:00"
},
{
"action":"zwe",
"score":3,
"timestamp":"2022-11-17T12:55:34.804+00:00"
},
{
"action":"qwe",
"score":5,
"timestamp":"2022-11-18T12:51:11.804+00:00"
}
],
"total_score": 13
}
]

> Question:

How i can get this part? How i can grop documents by field actions and get expected results?

"actions_total": [
{
"action": "qwe",
"count": 2,
"total_score": 10
},
{
"action": "zwe",
"count": 1,
"total_score": 3
}
]

Will be grateful for the help!

One option is to use $group twice and then $reduce to flatten the array:

db.actions.aggregate([
{$match: {timestamp: {
$gte: ISODate("2022-11-17T00:00:00Z"),
$lte: ISODate("2022-11-20T23:59:59Z")
}}},
{$lookup: {
from: "actions_score",
localField: "action",
foreignField: "action",
as: "score"
}},
{$project: {
user: 1, _id: 0, action: 1, score: {$first: "$score.score"}, timestamp: 1}},
{$group: {
_id: {user: "$user", action: "$action"},
count: {$sum: 1},
total_score: {$sum: "$score"},
actions: {$addToSet: "$$ROOT"}
}},
{$group: {
_id: "$_id.user",
user: {$first: "$_id.user"},
actions: {$push: "$actions"},
actions_total: {$push: {
action: "$_id.action",
count: "$count",
total_score: "$total_score"
}},
total_score: {$sum: "$total_score"}
}},
{$set: {
actions: {$reduce: {
input: "$actions",
initialValue: [],
in: {$concatArrays: ["$$value", "$$this"]}
}}
}}
])

See how it works on the playground example