How to execute this druid query with python script (code included)?

Question

This is my SQL query (it takes the average "value" over an hour for a specific device/metric combination) for a certain time period.

SELECT TIME_FLOOR(__time, 'PT1h') AS "__time_time_floor",
       AVG("value"), COUNT(*) AS "Count"
FROM "database"
WHERE "__time" >= CURRENT_TIMESTAMP - INTERVAL '1' DAY AND "device" = 'device_1'AND
      "metric"='metric_1'
GROUP BY 1
ORDER BY "__time_time_floor" DESC;

I'd like to use pydruid's "groupby" function to give run the above from a python script. I tried with the code below but I'm unsure how to input the "Avg" function in here. Anyone know if this is possible? Thanks! (Unless there's a more efficient way than using the groupby function)

 client = PyDruid('datasource website', 'druid/v2')  
 ts = client.groupby(
        datasource='dataset',
        granularity='hour',
        intervals='/'.join(time_range),
        dimensions=["__time", "value"],
        filter=((Dimension("device") == "device_1")) &
               (Dimension("metric") == "metric_1"),
        aggregations={"count": doublesum("count")}
    )