---
title: "Logistic Example"
author: "John Mount"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Logistic Example}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---


We can work an example similar to the [`rquery`](https://winvector.github.io/rquery/) [example](https://winvector.github.io/rquery/index.html) using a [`data.table`](https://rdatatable.gitlab.io/data.table/) 
back-end.

```{r}
library("rqdatatable")
```


```{r}
# data example
dL <- wrapr::build_frame(
   "subjectID", "surveyCategory"     , "assessmentTotal" |
   1          , "withdrawal behavior", 5                 |
   1          , "positive re-framing", 2                 |
   2          , "withdrawal behavior", 3                 |
   2          , "positive re-framing", 4                 )
```


```{r}
scale <- 0.237

# example rquery pipeline
rquery_pipeline <- local_td(dL) %.>%
  extend_nse(.,
             one = 1) %.>%
  extend_nse(.,
             probability =
               exp(assessmentTotal * scale)/
               sum(exp(assessmentTotal * scale)),
             count = sum(one),
             partitionby = 'subjectID') %.>%
  extend_nse(.,
             rank = cumsum(one),
             partitionby = 'subjectID',
             orderby = c('probability', 'surveyCategory')) %.>%
  extend_nse(.,
             isdiagnosis = rank == count,
             diagnosis = surveyCategory) %.>%
  select_rows_nse(., 
                  isdiagnosis == TRUE) %.>%
  select_columns(., 
                 c('subjectID', 'diagnosis', 'probability')) %.>%
  orderby(., 'subjectID')
```

Show expanded form of query tree.

```{r, comment=""}
cat(format(rquery_pipeline))
```

Execute the calculation.

```{r}
ex_data_table(rquery_pipeline)
```