Use definition_id instead of station_id

[lilyclements]

We want to use a definition_id instead of a station_id. This requires a few changes:

• Amend the station metadata to include a definition_id option.
• Amend the definition json files to now call in definition_id not station_id
• Amend the epicsawrap functions (annual_rainfall, etc) where we call station_id to automatically call definition_id from the station metadata
• Amend the R-Instat function to update the definition file to definition_id instead of station_id.
• Amend the R-Instat dialog to update the station metadata to include the new definition_id

[lilyclements]

@jkmusyoka I am working on this here so that we call definition_id instead of station_id. This will mean some changes required on the R-Instat dialog (which I will outline to Vitalis once I have sorted the R-code).

This will mean that the user will have to define a definition_id as well as station_id:

• That definition_id will be the name of the outgoing definition data. E.g., it will now be "<definition_id>..json" instead of "<station_id>..json"
• The station_id will still be used to name any exported station summaries.
• definition_id will update in the station metadata to reflect the new definition_id for that station_id. This is achieved through update_metadata_definition_id in PR Adding definitions_id option instead of station_id in definitions #122. It is also now called in our export_r_instat_to_bucket function automatically (PR 122). So while this is reflected in the R Code, it is not yet in R-Instat. This does mean we need to give the definition_id as a new input in R-Instat when PR 122 is merged.

This means that all data sets we use for testing and otherwise need to be defined in the metadata.rds file (for now, I have coded in that if they are not in the metadata, to set definition_id as station_id, so you shouldn't get any bugs from these changes).

Can you outline which data sets we are using for testing so I can add them into the metadata? For now, we can set their definition_id to be their station_id I think, until we update the R-Instat function.

[jkmusyoka]

@lilyclements

At the moment, zambia_eastern_4 is the test data from Zambia. It has 5 stations. And now there is a malawi_test folder which has a test dataset for 2 stations.

There is one more test data from Kenya which I will add this week. I will let you know when it is ready

[jkmusyoka]

@lilyclements while you are on this issue, could you also add more "countries" into the epicsa r system namely "ml_test" for malawi test data and "ke_test "for Kenya test data.

Thanks

[lilyclements]

@jkmusyoka thanks for this. I have done this now in PR 47