It is impossible to see what device collected a measurement in the data stream

[bardram]

Looking at the data stream, like:

  {
    "id": 5253,
    "data_stream_id": 639,
    "snapshot": {
      "syncPoint": {
        "synchronizedOn": "1970-01-01T00:00:00Z",
        "relativeClockSpeed": 1.0,
        "sensorTimestampAtSyncPoint": 0
      },
      "triggerIds": [
        10,
        14
      ],
      "measurements": [
        {
          "data": {
            "id": "f2307b9b-e401-4014-b2c7-68e05e2ca17c",
            "__type": "dk.cachet.carp.audio",
            "upload": true,
            "filename": "f2307b9b-e401-4014-b2c7-68e05e2ca17c.mp4",
            "metadata": {},
            "mediaType": "audio",
            "endRecordingTime": "2024-10-29T19:30:46.677579Z",
            "startRecordingTime": "2024-10-29T19:30:34.778037Z"
          },
          "sensorEndTime": 1730230246677579,
          "sensorStartTime": 1730230234778037
        },

there is no way to know what device collected the measurement or data item(s).

It is not part of the Trigger since in the protocol, this is stored in the TaskControl.

How should we store this?

Maybe @Whathecode has thought about this?

[Whathecode]

That's by design. To interpret the data, you need to look up the ids referenced (trigger ids) in the study protocol.

[Whathecode]

Actually, deviceRoleName should be there. The above comment is only true to get the full context (as in, know what type of device that is).

But, the data shown here is much different from e.g. a carp core data point: https://github.com/cph-cachet/carp.core-kotlin/blob/develop/carp.data.core%2Fsrc%2FcommonMain%2Fkotlin%2Fdk%2Fcachet%2Fcarp%2Fdata%2Fapplication%2FDataStreamPoint.kt

What am I looking at here?

The "snapshot" looks a bit like a DataStreamSequence. JSON schema: https://github.com/cph-cachet/carp.core-kotlin/blob/develop/rpc%2Fschemas%2Fdata%2FDataStreamSequence.json

That should have a dataStreamId: https://github.com/cph-cachet/carp.core-kotlin/blob/develop/rpc%2Fschemas%2Fdata%2FDataStreamId.json

Data stream id being an integer is definitely not carp core defined infrastructure, and neither is the "snapshot" property.

There's also minor things, like the start and end recording time of audio, which is a clear candidate to use the default timestamps for.

And, as far as I know __type needs to be the first defined field for polymorphic deserialization to work.

[bardram]

Actually, deviceRoleName should be there. The above comment is only true to get the full context (as in, know what type of device that is).

But, the data shown here is much different from e.g. a carp core data point: https://github.com/cph-cachet/carp.core-kotlin/blob/develop/carp.data.core%2Fsrc%2FcommonMain%2Fkotlin%2Fdk%2Fcachet%2Fcarp%2Fdata%2Fapplication%2FDataStreamPoint.kt

What is this DataStreamPoint used for? As far as I can see, this class is not used for anything in the domain model when I read the Kotlin code @Whathecode ? We upload a list of DataStreamBatch which consists of a list of DataStreamSequence which holds a list of Measurement. Nowhere is the DataStreamPoint used.

What am I looking at here?

You are looking at a DataStreamSequence which is part of a list in the DataStreamBatch.

The "snapshot" looks a bit like a DataStreamSequence. JSON schema: https://github.com/cph-cachet/carp.core-kotlin/blob/develop/rpc%2Fschemas%2Fdata%2FDataStreamSequence.json

That should have a dataStreamId: https://github.com/cph-cachet/carp.core-kotlin/blob/develop/rpc%2Fschemas%2Fdata%2FDataStreamId.json

Yes - this DataStreamId seems to be missing in the JSON.

Data stream id being an integer is definitely not carp core defined infrastructure, and neither is the "snapshot" property.

Indeed - I will create and issue for CAWS for @yuanchen233 to look at - there is something very wrong here in the data export.

There are also minor things, like the start and end recording time of audio, which is a clear candidate to use the default timestamps for.

A lot of the data we collect on the phone gets its own timestamp in a UTC format which we save.

And, as far as I know __type needs to be the first defined field for polymorphic deserialization to work.

I hope not - the order of fields in JSON should hopefully not matter - I can't control this in Dart.

[bardram]

Created an issue i CAWS to address these problems.

• The DataStream data downloaded in the Export has wrong JSON format and missing data carp-webservices-spring#178