In fmichelsendis/sendis: A Systematic Evaluated Nuclear Data Integral Survey

Screening evaluated nuclear data integral performance

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>", 
  echo=FALSE, 
  message=FALSE, 
  warning=FALSE, 
  eval=TRUE,
  fig.align='center'
  #fig.width = 100pct
  ) 

htmltools::tagList(rmarkdown::html_dependency_font_awesome())

library(dplyr)
library(plotly) 

The sendis R package provides datasets and simple functions to expedite common tasks in the comparison of integral benchmarking performance of nuclear data libraries.

Background

The ability of a nuclear data library to accurately reproduce the observables of integral benchmark experiments in transport simulations plays a key role in the data validation process.

• How to consistently compare different nuclear data libraries in terms of integral performance?
• How do different benchmarking suites compare to each other in terms of how they rank the different libraries ?
• In general, how to easily spot general trends and outliers?

Some helpful tools that help with the questions above have been developed in this package.

Examples

Interactive graphs are easily rendered using plotly :

data(sendis) 
df<-filter(sendis, INST=="NEA", VER!="2.2")
p1<-plot_cumulchi(df)%>%
  layout(xaxis = list(title = "Benchmark suite"),
         yaxis = list(title = TeX("\\chi^2")))%>%
  config(mathjax = "cdn")
# to display later using iframe, in order to render Mathjax : 
# htmlwidgets::saveWidget(p1, "myplot.html")

```{css, echo=FALSE}

div.blue { background-color:#e6f0ff; border-radius: 5px; padding: 20px;}

.holds-the-iframe { background:url(files/pacman_loader.gif) center center no-repeat; }

.col-left{ float: left; width: 50%; text-align: left; }

.col-right{ float: right; width: 50%; text-align: right; }

```r
# Load library
library(sendis)

# Filter through data as needed 
df<-filter(sendis, INST=="NEA")

# Plot
sendis::plot_cumulchi(df)

The sendis app

For seeing trends and filtering through the data, an interactive Shiny application has been developed.

• For Mac users : download a launcher for the sendis app and double-click the sendis icon.

• Alternatively, use the command line in an R session :

library(sendis)
sendis::runApp()

Reproducible, automated reporting

Integrate sendis into an Rmarkdown workflow to generate automated diagnosis pdf reports.

• Automatically compile plots and comparison tables;
• Transparent, reproducible workflow;
• Easily updatable with new or corrected data;

fmichelsendis/sendis documentation built on June 16, 2019, 7:30 a.m.