Skip to main content

Midas Wally Package

This is a package for the easy tracking and validation of analytics data. The goal is to allow you to track as much data as you want, so you can focus on solving problems with the data, rather than just storing and collecting it in the first place.

Usage​

To seamlessly integrate into the most common standards of data processing workflows, Midas has pivoted away from a free structured tree of data to that of tables.

1. Initialize Midas​


    Midas.init()
    Midas.ProjectId = "abcdef1234567"

2. Define a Storage Solution​

MongoDB​

One solution I personally use is MongoDB Atlas. You can read how to set that up here, remember to enabled the "Data API" under the "Services" menu. MongoDB offers a terabyte of storage at a reasonable price. To get the data out there are also many official MongoDB tools and solutions for that as well.

    local mongoDB = Midas.StorageProviders.MongoDB.new(
        "api-key-123", 
        "https://us-east-2.aws.data.mongodb-api.com/app/data-abcdef"
    )
    mongoDB.DebugPrintEnabled = RunService:IsStudio()
    Midas:SetOnBatchSaveInvoke(
        function(
            projectId: string,
            dataSetId: string,
            dataTableId: string,
            dataList: { [number]: { [string]: unknown } },
            format: { [string]: DataType },
            onPayloadSizeKnownInvoke:(number) -> ()
        ): boolean
            return mongoDB:InsertMany(
                projectId, 
                dataSetId, 
                dataTableId, 
                dataList, 
                format, 
                onPayloadSizeKnownInvoke
            )
        end
    )

3. Define a Table​

    -- midas accepts the rowData as a variadic type, allowing you to have type safety when recording data
    type RowData = {
        server_id: string,
        session_id: string,
        timestamp: DateTime,
        user_id: number,
        is_premium: boolean,
        friends_in_game: number?,
        pos_x: number?,
        pos_y: number?,
    }

    -- level of organization for datatables
    local dataSet = Midas:CreateDataSet("UserData", "abc123")

    -- a table with rows and columns
    local dataTable = dataSet:CreateDataTable("Session", "def456") :: Midas.DataTable<RowData>
    dataTable:AddColumn("server_id", "String", false)
    dataTable:AddColumn("session_id", "String", false)
    dataTable:AddColumn("timestamp", "Date", false)
    dataTable:AddColumn("user_id", "Int64", false)
    dataTable:AddColumn("friends_in_game", "Int32", true)
    dataTable:AddColumn("is_premium", "boolean", false)
    dataTable:AddColumn("pos_x", "Double", true)
    dataTable:AddColumn("pos_z", "Double", true)

4. Add Rows​

    local playerPosition: Vector3?

    dataTable:AddRow({
        server_id = game.JobId,
        session_id = "abc-123",
        timestamp = DateTime.now(),
        user_id = 123456,
        is_premium = true,
        friends_in_game = nil,
        pos_x = if playerPosition then playerPosition.X else nil,
        pos_z = if playerPosition then playerPosition.Z else nil,
    })

5. Post​

Manual​

If you want to post all tables at once you can

    Midas:Post(50, 400, 1, false)

Otherwise if you want to post a specific table that is available as well.

    dataTable:Post(50, 400)

Automated​

If you just want to forget about posting, you can tell Midas to try to manage it for you.

    Midas:Automate(RunService:IsStudio())