issue with season_start_probabilities

[ChrisMarsh82]

It uses values from definition files instead of the passed in start_dates
e.g.
Input:

{
  "country": "zm",
  "station_id": "test_1",
  "start_dates": [
    200,
    220,
    250,
    270,
    300,
    320
  ]
}

Output:

{
  "metadata": {
    "season_start_probabilities": {
      "specified_day": {
        "val1": 50,
        "val2": 100,
        "val3": 150
      }
    }
  },
  "data": [
    {
      "station": "Agades",
      "day": 50,
      "proportion": 0
    },
    {
      "station": "Agades",
      "day": 100,
      "proportion": 0
    },
    {
      "station": "Agades",
      "day": 150,
      "proportion": 0
    }
  ]
}

[lilyclements]

@ChrisMarsh82 look at the data output - that is also giving for the days that are in the metadata. This should work with override = TRUE.

[ChrisMarsh82]

@ChrisMarsh82 look at the data output - that is also giving for the days that are in the metadata. This should work with override = TRUE.

Apologies for this. I have now added this and it works fine. Just checking I have added 'override = TRUE' to all functions that I call. Is this correct?

[lilyclements]

@ChrisMarsh82 if there is already a stored summary output, then yes you need to add override = TRUE. There is stored summaries for the test_1 data only (and for test_3 as well, but that's come from using the R-Instat to bucket functions). So if you were to try it with another station, like station 16, you shouldn't need to specify override = TRUE

[ChrisMarsh82]

@lilyclements This may cause issues. How do we decide to pass it or not? How would we know where the data has come from?

[lilyclements]

@ChrisMarsh82 this is a more global problem then that just having override. We said about storing summaries and that we want to call the summaries.
By default if summaries are stored, it calls them otherwise it creates them.

The override option is just for if summaries are stored and you want to create new summaries. I don’t know how we want to handle this case but I doubt it would be a problem if they don’t want to store the data in the buckets anyway?