Tbats: Error in attributes(non.seasonal.model$errors) <- attributes(origy) : invalid time series parameters specified

Question

I'm currently banging my head on my desktop due to the following error, if someone knows how to fix it I'd really appreciate it, or any kind of help.

Trying to adapt the "demo tutorial" of a tbats, to another set of data seems much more complicated than expected...

I'm following this tutorial: https://www.projectpro.io/recipes/what-is-tbats-model-time-series-use-it

But with the following values:

with the command rain_ts <- ts(final_df$value) I obtain MY rain_ts with MY values:

indeed when plotting rain_ts I obtain:

[1] -0.001335113  0.003342246  0.001998668 -0.009308511           NA -0.004026846 -0.004716981 -0.012863913           NA
[10]  0.008916324 -0.001359619  0.004084411 -0.002033898  0.005434783  0.001351351  0.005398111 -0.004026846           NA
[19]  0.006738544  0.000669344  0.006020067           NA  0.004654255

And then trying to apply the crucial step:

model_tbats <- tbats(rain_ts)

I obtain the following error:

I'm not sure to perfectly understand the nature of the error, and I didn't understand or sucess to make any analogy with similar problem I've seen...

Thank you again for any kind of help.