base-camp-luftdaten / data Goto Github PK

For the frontend, we need an endpoint which gives us an average for an entire day/hour(?) per sensor.

Ideally, the frontend requests the data for a single day/hour and the endpoint returns a list of sensors which include the average measurements for that sensor grouped into 5min intervals.

For example, the result could be a map <sensor_id> --> list of 288 values (1440 minutes per day / 5 minutes = 288)

{
  "92": [ { "p1": 4.2, "p2": 3.5 }, { "p1": 5.2, "p2": 3.3 }, "286 more...." ]
}

We need to add a continuous field (boolean) to the response of the /bySensor endpoint.
Since the response is currently just a JSON array, we should change it to an object which could look like this (where the measurements field is the array of the current response):

{
  "continuous": true,
  "measurements": []
}

"Continuous" can be calculated by looking 7 days into the past from the requested timestamp (for the requested sensor) and checking whether we always have a measurement for every 1 hour step. If there hasn't been a measurement +-3 hours a timestamp, the result isn't continuous anymore.

Currently, the /bySensor endpoint returns all measurements +- 10 minutes for a certain timestamp. This doesn't respect the cases where the timestamp is in the first/last 10 minutes of a day yet.

Since the measurements are grouped by day and sensor in the database, we would need to include an additional query in these cases.

See the the comment in

Checking the cases should be as simple as:

if (((long) timestamp * 1000) - calendar.getTimeInMillis() < threshold * 1000) {
  // we need to query the previous day
}

We need an endpoint to read all available sensors. A sensor is uniquely identified by ID & Long/Lat

For example: /api/measurements/bySensor?sensor=17287&timestamp=1560034800 which would return all measurements by that specific sensor -10min and +10min from the 1560034800 timestamp which represents Saturday, 08-Jun-19 23:00:00 UTC in epoch seconds.