2️⃣ R Functions

The R functions used for interacting with Aplos NCA to analyze data are shown and described below.

Get JSON Web Token

This function uses 4 pieces of information to acquire a jwt from Amazon Cognito. These are the client identifier (client_id), username, password, and region. The client identifier and region are available from the web interface under Profile | API Configuration.

get_jwt <- function(client_id, username, password, region){
  headers <- c("Content-Type" = "application/x-amz-json-1.1", 
               "X-Amz-Target" = "AWSCognitoIdentityProviderService.InitiateAuth")
  body <- list(
    AuthFlow="USER_PASSWORD_AUTH",
    AuthParameters=list("USERNAME"=username, "PASSWORD"=password),
    ClientId=client_id,
    method_type = "post"
  )
  
  response <- POST(url = paste0("https://cognito-idp.",region,".amazonaws.com/"),
                   add_headers(.headers = headers),
                   body = toJSON(body, auto_unbox = TRUE))
  
  result <- fromJSON(content(response, "text", encoding = "UTF-8"))
  jwt <- result$AuthenticationResult$IdToken
  return(jwt)
}

Analysis with Aplos NCA

The function requires the following information:

• JSON web token
• Analysis dataset filename (and path)
• Configuration in JSON format
• Meta data in JSON format
• API URL (The URL for the API is available from the web interface under Profile | API Configuration)
• Folder for results file
• Unzip (True/False)

Acquire URL for file upload

The following code requests the secure URL to which you will upload the input data file.

  # Get pre-signed URL
  tenant_id <- decode_jwt(token)$payload$`custom:aplos_user_tenant_id`
  user_id <- decode_jwt(token)$payload$`custom:aplos_user_id`
  filename = str_split_i(input, "/", -1)
  upload_url = paste0(url,"/tenants/",tenant_id,"/users/",user_id,"/nca/files")
  headers <- c("Content-Type" = "application/json",
               "Authorization" = paste0("Bearer ",token))
  body <- list(
    "file_name"=filename,
    method_type = "post"
  )
  
  response <- POST(url = upload_url,
                   add_headers(.headers = headers),
                   body = toJSON(body, auto_unbox = TRUE))
  upload_url_result <- fromJSON(content(response, "text", encoding = "UTF-8"))

Upload input data file

This code uploads the data file for analysis using the secure URL.

  # Upload input file
  cat("Uploading input file ... \n")
  upload_url = upload_url_result$presigned$url
  headers <- c()
  body <- list(
    key = upload_url_result$presigned$fields$key,
    "x-amz-algorithm" = upload_url_result$presigned$fields$`x-amz-algorithm`,
    "x-amz-credential" = upload_url_result$presigned$fields$`x-amz-credential`,
    "x-amz-date" = upload_url_result$presigned$fields$`x-amz-date`,
    "x-amz-security-token" = upload_url_result$presigned$fields$`x-amz-security-token`,
    policy = upload_url_result$presigned$fields$policy,
    "x-amz-signature" = upload_url_result$presigned$fields$`x-amz-signature`,
    file = upload_file(normalizePath(input))
  )
  
  upload_response <- POST(url = upload_url, body = body)

Initiate Analysis

This code initiates the analysis by Aplos NCA.

  cat("Loading analysis configurations \n")
  config.list <- fromJSON(txt = config)
  
  cat("Loading analysis meta data \n")
  meta.list <- fromJSON(txt = meta)
  
  cat("Initiating analysis ... \n")
  tenant_id <- decode_jwt(token)$payload$`custom:aplos_user_tenant_id`
  user_id <- decode_jwt(token)$payload$`custom:aplos_user_id`
  headers <- c("Content-Type" = "application/json",
               "Authorization" = paste0("Bearer ",token))
  body <- list(
    file = list(
      id = upload_url_result$file$id
    ),
    meta_data = meta.list,
    configuration = config.list
  )
  
  analysis_response <- POST(url = paste0(url,"/tenants/",tenant_id,"/users/",user_id,"/nca/executions"),
                            add_headers(.headers = headers),
                            body = toJSON(body, auto_unbox = TRUE))
  if (http_status(analysis_response)$category == "Success") {
    cat("Execution initiated. \n")
  } else {
    cat("Error in the request. \n")
  }
  analysis_response2 <- fromJSON(content(analysis_response, "text", encoding = "UTF-8"))
  execution_id <- analysis_response2$execution_id

Request Execution Status

This code checks the status of the analysis, prints that status, and then returns the download URL when the analysis is complete.

  cat("Checking status ... \n")
  tenant_id <- decode_jwt(token)$payload$`custom:aplos_user_tenant_id`
  user_id <- decode_jwt(token)$payload$`custom:aplos_user_id`
  headers <- c("Content-Type" = "application/json",
               "Authorization" = paste0("Bearer ",token))
  complete <- FALSE
  while (!complete) {
    response <- GET(paste0(url,"/tenants/",tenant_id,"/users/",user_id,"/nca-engine/executions/",execution_id),
                    add_headers(.headers = headers))
    result <- fromJSON(content(response, "text", encoding = "UTF-8"))
    if(result$status == "failed") {break}
    complete <- result$status == "complete"
    if(!complete) {
      cat("Not yet complete ... ",result$status," \n")
      Sys.sleep(4)}
  }
  if(result$status == "complete") {
    cat("Execution complete. Execution ID = ",execution_id,"\n")
  } else {
    cat(paste0("Execution failed. Execution ID = ",execution_id,"\n"))
  }
  cat("Downloading results. \n")
  if(result$status == "complete") {
    download_url <- result$presigned$url
    output_file <- paste0(output,"results-",substr(execution_id,1,10),".zip")
    # create output directory if doesn't already exist
    download(download_url, dest=output_file, mode = "wb", quiet = TRUE)
    if(unzip) {
      unzip_folder = paste0(output,"results-",substr(execution_id,1,10))
      unzip(output_file, exdir = unzip_folder)
      cat("Results file downloaded and unziped. \n")
      cat(paste0("Location is ",unzip_folder,"/ \n"))
    } else {
      cat("Results file downloaded. \n")
    }
  }