Use R to save table into Excel - openmpp/openmpp.github.io GitHub Wiki

Use R to save output table into Excel

It is a convenient to use GNU R to prepare model parameters and analyze output values. There are two different R APIs which we can use for openM++ models:

  • openMpp package: simple and convenient specially for desktop users, upstream and downstream analysis;
  • oms JSON web-service API: preferable choice to run models on computational clusters and in cloud.

There is also an excelent R package created by Matthew T. Warkentin available at: mattwarkentin.github.io/openmpp.

Below is an example how to use oms JSON web-service to read multiple output table values from multiple model runs and save it into XLSX file:

  • using RiskPaths demo model
  • reading model run names and output table names from input Excel file as on screenshots below
  • for each table retriving output values for all model runs
  • retrieving model runs metadata: run name, description, notes, date and time
  • retrieving output tables metadata: name, description and notes
  • saving each table output values as separate Excel workbook sheet
  • saving all model runs metadata and tables metadata as separate sheets

Input Excel workbook:

Model run names in Excel Output table names in Excel

Output Excel workbook:

Model output results and metadata in Excel

R script

#
# Read multiple tables from multiple model runs and save it as XLSX file
# Also save model runs metadata and tables metadata (name, description, notes) into .csv files
# Model run names and table names are coming from another input XLSX file
#
# If any of library below is not installed then do:
#   install.packages("jsonlite")
#   install.packages("httr")
#   install.packages("readxl")
#   install.packages("writexl")
#
library("jsonlite")
library("httr")
library("readxl")
library("writexl")

# Include openM++ helper functions from your $HOME directory
#
source("~/omsCommon.R")

#
# Model digest of RiskPaths version 3.0.0.0: "d90e1e9a49a06d972ecf1d50e684c62b"
# We MUST use model digest if there are multiple versions of the model published.
# We can use model name if only single version of the model is published.
#
md <- "d90e1e9a49a06d972ecf1d50e684c62b"

# oms web-service URL from file: ~/oms_url.txt
#
apiUrl <- getOmsApiUrl()

# model runs can be identified by digest, by run stamp or by run name
# run digest is unique and it preferable way to identify model run
# run names are user friendly may not be unique
#
# read model run names from some XLSX file, 
#   it must have sheet name = "RunNames" with A column "RunNames"
#
rn <- read_xlsx(
  "model-runs-to-read-and-tables-to-read.xlsx", 
  sheet = "RunNames", 
  col_types = "text"
  )

# read table names from some XLSX file, 
#   it must have sheet name = "TableNames" with A column "TableNames"
#
tn <- read_xlsx(
  "model-runs-to-read-and-tables-to-read.xlsx",
  sheet = "TableNames",
  col_types = "text"
  )

# get table information
#
rsp <- GET(paste0(
    apiUrl, "model/", md, "/text"
  ))
if (http_type(rsp) != 'application/json') {
  stop("Failed to get first model info")
}
jr <- content(rsp)
tTxt <- jr$TableTxt

tableInfo <- data.frame()

for (t in tTxt) {
  for (tbl in tn$TableNames)
  {
    if (t$Table$Name == tbl) {
      ti <- data.frame(
          TableName = tbl,
          TableDescription = t$TableDescr,
          TableNotes = t$TableNote
        )
      tableInfo <- rbind(tableInfo, ti)
      break
    }
  }
}

# get run information
#
runInfo <- data.frame()

for (run in rn$RunNames)
{
  rsp <- GET(paste0(
      apiUrl, "model/", md, "/run/", URLencode(run, reserved = TRUE), "/text"
    ))
  if (http_type(rsp) != 'application/json') {
    stop("Failed to get first run info of: ", run)
  }
  jr <- content(rsp)
  ri <- data.frame(
      ModelName = jr$ModelName,
      ModelVersion = jr$ModelVersion,
      RunName = jr$Name,
      SubCount = jr$SubCount,
      RunStarted = jr$CreateDateTime,
      RunCompleted = jr$UpdateDateTime,
      RunDescription = "",
      RunNotes = ""
    )
  if (length(jr$Txt) > 0) {
    ri$RunDescription <- jr$Txt[[1]]$Descr
    ri$RunNotes <- jr$Txt[[1]]$Note
  }
 
  runInfo <- rbind(runInfo, ri)
}

# for each table do:
#   combine all run results and write it into some .xlsx file
#

shts <- list(
    RunInfo = runInfo,
    TableInfo = tableInfo
  )

for (tbl in tn$TableNames)
{

  allCct <- NULL
  isFirst <- TRUE

  for (run in rn$RunNames)
  {
    cct <- read.csv(paste0(
      apiUrl, "model/", md, "/run/", URLencode(run, reserved = TRUE), "/table/", tbl, "/expr/csv"
      ))

    # build a pivot table data frame:
    # use first run results to assign all dimensions and measure(s)
    # from all subsequent model run bind only expr_value column
    if (isFirst) {
      allCct <- rbind(allCct, cct)
      isFirst <- FALSE
    } else {
      cval <- data.frame(expr_value = cct$expr_value)
      allCct <- cbind(allCct, cval)
    }

    # use run name for expression values column name
    names(allCct)[names(allCct) == 'expr_value'] <- run
  }
  shts[[ tbl ]] <- allCct
}

write_xlsx(shts, paste0("output-tables-data.xlsx"))
⚠️ **GitHub.com Fallback** ⚠️