Feat: MongoDB Adapter

[kodumbeats]

This PR adds a MongoDB database adapter

[kodumbeats]

@eldadfux I could use your advice going forward:

When fetching documents, Mongo client SDK returns everything in a MongoDB\Model\BSONDocument object:
https://docs.mongodb.com/php-library/v1.2/reference/bson/#phpclass.MongoDB\Model\BSONDocument

What'd be the best approach to parse these objects and get back the PHP array? I've found I can go one level deep with iterator_to_array, but the array contains BSON models as well (MongoDB\Model\BSONArray and MongoDB\BSON\ObjectId) - looks like this:

array(10) {
  ["_id"]=>
  object(MongoDB\BSON\ObjectId)#167 (1) {
    ["oid"]=>
    string(24) "60945e59acf8b169f83ce597"
  }
  ["+id"]=>
  string(9) "documents"
  ["+read"]=>
  object(MongoDB\Model\BSONArray)#164 (1) {
    ["storage":"ArrayObject":private]=>
    array(1) {
      [0]=>
      string(1) "*"
    }
  }
  ["+write"]=>
  object(MongoDB\Model\BSONArray)#169 (1) {
    ["storage":"ArrayObject":private]=>
    array(1) {
      [0]=>
      string(1) "*"
    }
  }
  ["name"]=>
  string(9) "documents"
  ["attributes"]=>
  string(2) "[]"
  ["indexes"]=>
  string(2) "[]"
  ["attributesInQueue"]=>
  string(2) "[]"
  ["indexesInQueue"]=>
  string(2) "[]"
  ["+collection"]=>
  string(11) "collections"
}

[TorstenDittmann]

@eldadfux I could use your advice going forward:

When fetching documents, Mongo client SDK returns everything in a MongoDB\Model\BSONDocument object:
https://docs.mongodb.com/php-library/v1.2/reference/bson/#phpclass.MongoDB\Model\BSONDocument

What'd be the best approach to parse these objects and get back the PHP array? I've found I can go one level deep with iterator_to_array, but the array contains BSON models as well (MongoDB\Model\BSONArray and MongoDB\BSON\ObjectId) - looks like this:

array(10) {
  ["_id"]=>
  object(MongoDB\BSON\ObjectId)#167 (1) {
    ["oid"]=>
    string(24) "60945e59acf8b169f83ce597"
  }
  ["+id"]=>
  string(9) "documents"
  ["+read"]=>
  object(MongoDB\Model\BSONArray)#164 (1) {
    ["storage":"ArrayObject":private]=>
    array(1) {
      [0]=>
      string(1) "*"
    }
  }
  ["+write"]=>
  object(MongoDB\Model\BSONArray)#169 (1) {
    ["storage":"ArrayObject":private]=>
    array(1) {
      [0]=>
      string(1) "*"
    }
  }
  ["name"]=>
  string(9) "documents"
  ["attributes"]=>
  string(2) "[]"
  ["indexes"]=>
  string(2) "[]"
  ["attributesInQueue"]=>
  string(2) "[]"
  ["indexesInQueue"]=>
  string(2) "[]"
  ["+collection"]=>
  string(11) "collections"
}

https://stackoverflow.com/a/61972048

Did you try this?

[eldadfux]

We need to make sure this limits the number of documents the DB scans and not just the number of results returned (which is always 1 anyway).

[kodumbeats]

I'm not sure on this one:
According to the mongo shell docs:

Field	Type	Description
limit	integer	Optional. The maximum number of documents to count.

But for the PHP lib implementation of the same method:

Field	Type	Description
limit	integer	Optional. The maximum number of matching documents to return.

I think all implementations just limit the number of documents passed to the next step in the pipeline:
https://docs.mongodb.com/manual/reference/operator/aggregation/limit/

[eldadfux]

Maybe we can raise an issue on the Mongo SDK repo and ask. I feel like this is just a doc mistake

[kodumbeats]

Yep, limit was the wrong setting. After more research, I have good news and bad news 😐

Good news: I found the exact setting we need for our desired count() and find() methods: maxScan:

maxScan Optional. Maximum number of documents or index keys to scan when executing the query.

Bad news: it was deprecated in Mongo 4.0 in favor of time-based limits:

MongoDB deprecates the option maxScan for the find command and the mongo shell helper cursor.maxScan(). Instead, use maxTimeMS option or the helper cursor.maxTimeMS().

https://docs.mongodb.com/manual/release-notes/4.0-compatibility/#maxscan-option

Since the behavior was specifically deprecated by the Mongo team, I'm not sure we'll be able to hack our way around it.