Odtnhj

I would like to play with the plumber library by making an app that takes 14 days of historical data and returns an exponential smoothing forecast.

The problem is I am somewhat unfamiliar with passing a lot of data (a parameter with multiple values) to an API. My questions can be summarized as follows:

1. How should I prepare the data in R to be passed to the API?




2. How should the API be prepared in plumber to receive time series data?

Below is some example data and a function which accomplishes what I would like in R.

library(tidyverse)

# data to be passed to API
head(forecast::wineind,14)
#> Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov
#> 1980 15136 16733 20016 17708 18019 19227 22893 23739 21133 22591 26786
#> 1981 15028 17977 
#> Dec
#> 1980 29740
#> 1981

#* Return Forecast Data
#* @list a The first number
#* @get /simple_fcast
function()
 ts() %>% 
 forecast::ets() %>% 
 forecast::forecast()

#> function()
#> ts() %>% 
#> forecast::ets() %>% 
#> forecast::forecast()
#> 

^{Created on 2018-11-14 by the reprex package (v0.2.1)}

The answer is to use "message body" per the documentation here https://www.rplumber.io/docs/routing-and-input.html#

#' @post /user
function(req, id, name)
 list(
 id = id,
 name = name,
 raw = req$postBody
 )

Running curl --data "id=123&name=Jennifer" "http://localhost:8000/user" will return:

 "id": [123],
 "name": ["Jennifer"],
 "raw": ["id=123&name=Jennifer"]

I found that arrays can also be passed to functions.