将R标记参数传递给源R脚本

Question

我有一个闪亮的应用程序，用户可以选择传递给参数化R标记报告的选项。然后，Rmd获取一系列R脚本来提取和总结数据，创建报告的情节等。

我正在寻找的提取数据的脚本包括一个参数化的SQL查询，它继承了R标记params的值(而后者又是从闪亮的input继承的)。但是，整个过程在此时停止，我得到一个错误，说明params不存在。

我相当肯定，将输入从闪亮的值传递到R标记参数是很好的--所以问题似乎是将它们传递给源脚本(注意:它只是一个R脚本，而不是一个函数)。我猜想这与脚本在源代码时访问的环境有关(尽管它使用R标记中的前一个块中生成的数据库连接)--但除此之外，对于如何纠正这个问题，我有点不知所措。任何想法都将不胜感激。

这是一个闪亮的应用程序：

##########################################
# SHINY APP - USER INTERFACE:

ui = fluidPage (
    selectInput("pathogen", "Enter pathogen of interest:", c("Campylobacter" = "Campylobacter", "Escherichia" = "Escherichia",
                "Salmonella" = "Salmonella", "Shigella" = "Shigella"), selected = "Salmonella" ),

    radioButtons("pkginstall", "Install required packages?",  c("Yes" = "yes", "No" = "no"),selected = "yes"),

    downloadButton("report", "Generate report")
)


##########################################
# SHINY APP - SERVER LOGIC:

#fileInput("download_location","Select File Location"),
server = function(input, output) {
    # Create the output: 
    output$report = downloadHandler(

      filename = paste0("Pathogen Report ", input$pathogen, "_", format(Sys.time(),"%d-%b-%Y %H.%M"), ".html"),

      content = function(file) {

        # Copy the .Rmd to a temporary directory:
        tempReport <- file.path(tempdir(), "Pathogen_Report.Rmd")
        file.copy("Pathogen_Report.Rmd", tempReport, overwrite = TRUE)


        # Set up parameters to pass to Rmd document:
        params <- list(pathogen = input$pathogen, pkginstall = input$pkginstall)

        # Define name of report:
        outname <- paste0("Pathogen Report ", input$pathogen, "_", format(Sys.time(),"%d-%b-%Y %H.%M"), ".html") 

        # Knit the document:
        created_filename <- rmarkdown::render(input = tempReport, 
                          output_file = outname, 
                          params = params,
                          envir = new.env(parent = globalenv())
                          )
        file.rename(created_filename, file)
      }
    )
  }

##########################################
# SHINY APP - RUN:

# Run app:
shinyApp(ui =ui, server=server)

##################################################################

这是R标记的YAML标题：

---
params:
  pathogen: 
    label: "Enter pathogen of interest:" 
    value: Shigella
    input: select
    choices: [Campylobacter, Escherichia, Salmonella, Shigella]
title: "Pathogen Report"
date: "`r format(Sys.time(), '%d %B %Y')`"
output: 
  phecharts::html_phe:
    includes:
      in_header: phe_logo.html
---

以及相关的块获取R脚本：

{r, GDW Query, echo=FALSE, cache=FALSE, message=FALSE, warning=FALSE, results='hide'}
####################################################################
# QUERY DATABASE AND EXTRACT DATA 

source("Extract_data.R")

Extract_data.R包含一个SQL查询，其中病原体的名称应该取代从R标记参数继承的名称：

# Example SQL to PostgreSQL database:
query <- "SELECT * FROM table1 WHERE table1.organism ~ '^@pathogen'"

# Substituting pathogen for pathogen name from R markdown parameters:
query <- gsub("@pathogen", params$pathogen, query)

# Executing the query:
mydata <- data.table(RPostgres::dbGetQuery(conn = dbcon, statement = query))

请注意，数据库连接已经成功地通过在前面的R标记块中找到另一个脚本来建立。

这是我得到的错误：

Quitting from lines 84-88 (Pathogen_Report.Rmd) 

Warning: Error in gsub: object 'params' not found
  [No stack trace available]

Answer 1

看看local in source (?source)中的参数。

local TRUE、FALSE或环境，以确定解析表达式的计算位置。FALSE (默认值)对应于用户的工作区(全局环境)，并对应于调用源的环境。

当直接呈现Rmd时，params是默认设置的，您是在全局环境中。所以这是可行的：

---
params:
  pathogen: 
    label: "Enter pathogen of interest:" 
    value: Shigella
    input: select
    choices: [Campylobacter, Escherichia, Salmonella, Shigella]
  pkginstall:
    value: no
title: "Pathogen Report"
date: "`r format(Sys.time(), '%d %B %Y')`"
output: html_document
---

```{r GDW Query, echo=TRUE, cache=FALSE, message=FALSE, warning=FALSE, results='hide'}

####################################################################

查询数据库和提取数据

源(Extract_data.R，local=FALSE) #与源(Extract_data.R)相同

However, when running the `Rmd` through a `Shiny App` you want to work in the environment that `Shiny` works in, and you want to source the external script as if it was pasted in line (see [https://shiny.rstudio.com/articles/scoping.html](https://shiny.rstudio.com/articles/scoping.html)). The following should work:

```javascript

params：

病原体：

label: "Enter pathogen of interest:" 

value: Shigella

input: select

choices: [Campylobacter, Escherichia, Salmonella, Shigella]

普金斯托：

value: no

标题：“病原体报告”

日期："r format(Sys.time(), '%d %B %Y')“

输出: html_document

####################################################################
# QUERY DATABASE AND EXTRACT DATA 
source(Extract_data.R, local=TRUE)