How to pass dataframe as an argument to mrsdeploy::publishService through curl POST?

Question

I am trying to provide a R function as a web service for which I am using https://blogs.msdn.microsoft.com/mlserver/2018/07/26/dockerizing-r-and-python-web-services/ as a reference.

I changed my function accordingly so that it should accept a dataframe. I want to consume the web service through curl. How do I pass a dataframe to publishService function through curl?

I tried send the dataframe as a JSON so that curl POST should accept it as a string, but it isn't accepting.

library(mrsdeploy)
library(jsonlite)
manualTransmission <- function(hp, wt) { 
     hp <- jsonlite::fromJSON(hp)
     wt <- jsonlite::fromJSON(wt)
     c(hp[1,1],wt[1,1]) 
}
remoteLogin("http://localhost:12800", username = "admin", password = 
"Microsoft@2018", session = FALSE)
api <- publishService("ManualTransmissionService", code = 
manualTransmission, inputs = list(hp = "character", wt = "character"), 
outputs = list(answer = "vector"), v = "1.0.0") 

The command:

curl --header "Content-Type: application/json" --header "Authorization: Bearer <accesstoken>" --request POST --data "{\"hp\": \"[{\"hp_col\": 5}]\",\"wt\": \"[{\"wt_col\": 5}]\"}" http://localhost:12800/api/ManualTransmissionService/1.0.0