Modifying columns with custom API

Hi everyone,

I was planning on using my own NLP prediction model (behind a custom API) to make predictions whenever new entries of a certain data source were added to my app. However, I was thinking of a way to achieve this as there is no trivial way of doing this on Appsheet. I was thinking of creating a task that would call webhook whenever data is added to the source. The hook would then do a POST to my API with the input data necessary to predict the label. My custom API would then use the Appsheet API to update the new predicted label of the new added entry.

Unfortunately again, I see absolutely no way to achieve this as:

  • Tasks currently seem to trigger whenever the data source is updated, so not only added, which is what I ultimately want (SOLVED: see update)
  • You have to hardcode parameters in the webhook. This is a problem, I do not know the key nor the input data of the data entry that is being added during the Appsheet app beforehand. This is also means that I could never update the new predicted label of an entry outside the app.

Am I overcomplicating this process or is this just not possible as of yet? An easier way would probably be to put my own predictive model (say model123) into production at the Google Cloud Platform and then use PREDICT(“model123”) on Appsheet. I don’t think this is possible either.

Let me know if I’m missing anything. Thank you!

– UPDATE:

I’ve managed to create a bot that will trigger the process (webhook) whenever new data is added to my data source(event).

2 Likes

Hi Kenan,

You can create a template for the webhook call payload body.
Check out this doc (it also applies to webhook body although the doc said it’s for email task) – Email Templates | AppSheet Help Center

Thanks!

2 Likes