由于反应性错误，闪亮的 RStudio 应用程序无法正常工作

Posted 2023-03-24

【中文标题】由于反应性错误，闪亮的 RStudio 应用程序无法正常工作

【英文标题】：Shiny RStudio app not working due to reactive error

【发布时间】：2020-07-09 09:29:31

【问题描述】：

运行应用程序时出现以下错误：

.getReactiveEnvironment()$currentContext() 中的错误： 如果没有活动的反应上下文，则不允许操作。 （你试图做一些只能在反应式表达式或观察者内部完成的事情。）

我第一次尝试了闪亮的应用程序，所以我可能会错过一些我无法弄清楚的关键方面。如果有人可以提供帮助，将不胜感激。

这是我制作的代码：

用户界面：

library(shiny)
shinyUI(pageWithSidebar(
  headerPanel(h1('Sentiment Analyzer for .txt Files'), 
              windowTitle = "Sentiment Analysis"),

  # Getting User Inputs
  sidebarPanel(
    fileInput("fileInput","Please Upload the File"),
    textInput("findPhr5ase", "Enter data to be searched"),
    submitButton('Analyze')
  ),

  mainPanel(
    sidebarPanel(
      verbatimTextOutput("file_Health1")
    ),
    sidebarPanel(
      textOutput("file_Health2")
    ),
    sidebarPanel(
      textOutput("search_Result1")
    ),
    sidebarPanel(
      verbatimTextOutput("search_Result2")
    ),
    sidebarPanel(
      verbatimTextOutput("search_Result3")
    ),
    sidebarPanel(
      wordcloud2Output("sentimentCloud", width = "100%", height = "400px")
    ),
    sidebarPanel(
      plotOutput("sentiment_Analysis", width = "40%", height = "200px")
    ),
    sidebarPanel(
      plotOutput("topSentiments", width = "40%", height = "200px")
    ),
    sidebarPanel(
      dataTableOutput("sentimentScores")
    )
  )
)
)

服务器：

shinyServer <- function(input, output) 

  observeEvent(input$analysis, 
    #Function to read file
    searchText <- reactive(renderText(input$findPhrase)) 
    #fileType <- renderText(input$selectType)
    read_data_from_file <- function ()
      text <- reactive(readtext(input$fileInput))
      text <- toString(text[,2])
      text<- gsub("\n", "", text)
      text <- gsub("[\r\n]", "", text)
      return(text)
      

    check_data <- function ()
      if (is.null(read_data_from_file()) || is.null(searchText)) 
      return("File or Search Text is empty")
      
      else 
        return("Starting to analyze file")
      
    
    output$file_Health1 <- renderPrint(check_data())

    #WORD COUNT
    output$file_Health2 <- renderText(sapply(read_data_from_file(),stri_stats_latex)['Words',])

    #Function to clean the text file
    clean_data <- function(text)
      #Remove NA values
      NAindices = which(is.na(text))
      if (is.numeric(NAindices) & length(NAindices) > 0) 
        text <- text[-c(NAindices)]
      
      #Remove single letters and hyphen
      text <- gsub(pattern = "\\b[A-z]\\b1", replace = " ", text)
      text <- gsub(pattern = "[-]", replace = " ", text)
      #Conversion to corpus
      docs <- Corpus(VectorSource(text))
      #Cleaning the corpus
      docs <- tm_map(docs, content_transformer(tolower))
      docs <- tm_map(docs, removeNumbers)
      #Remove Stopwords
      docs <- tm_map(docs, removeWords, stopwords("english"))
      docs <- tm_map(docs, removeWords, c("don", "s", "t"))
      docs <- tm_map(docs, removePunctuation)
      docs <- tm_map(docs, stripWhitespace)
      #Rewriting as text file
      writeCorpus(docs, filenames = "clean.txt")
      #Reading cleaned text file
      text1 <- readtext("clean.txt")
      text1 <- toString(text1[,2])
      return(text1)
    


    #Function to tokenize the words
    tokenize <- function(text)
      words = tokenize_words(text)
      return (words)
    

    #Hashmap function to get frequency of all words
    get_freq_map <- function(words) 
      hashmap <- hashmap("character")
      for(word in words)
        hashmap[word] <-  0
      
      for(word in words)
        hashmap[word] <- as.numeric(hashmap[word]) +  1
      
      return (hashmap)
    

    #Calling the functions
    text2 = tokenize(clean_data(read_data_from_file()))
    freq_map = get_freq_map(text2[[1]])

    #Function to find frequency of the word given as input
    search_keyword_freq <- function(X) 
      return (freq_map[X])
    

    #Word count Frequency function
    output$search_Result1 <-renderText(search_keyword_freq(searchText))


    #Reading text file
    text = read_data_from_file()
    #Tokeninzing the sentences
    sentences <- tokenize_sentences(text)

    #Function to create a vector of line numbers where the searched word occurs
    line_nums <- vector()
    for (i in 1:length(sentences[[1]])) 
      filtered_sent_words = tokenize(clean_data(sentences[[1]][i]))
      sentence_freq_map = get_freq_map(filtered_sent_words[[1]])
      freq <- tryCatch(sentence_freq_map[searchText], error = function(e) NULL)
      if (!(is.null(freq))) 
        if (freq > 0) 
          line_nums <- c(line_nums, i)
        
      
    
    #Printing the vector and sentences where the word occurs
    output$search_Result2 <- renderPrint(line_nums)
    output$search_Result3 <- renderPrint(for (i in 1:length(line_nums)) 
                                print(sentences[[1]][line_nums[i]]))

    ##################################################################
    # Word Cloud
    ##################################################################
    # take all the phrases
    df_data <- data.frame(text = sapply(docs, as.character), stringsAsFactors = FALSE)
    # add an id, from 1 to n
    df_data$doc_id <- row.names(df_data)

    #Tokenize
    tidy_docs <- df_data %>% unnest_tokens(word, text)
    par(mar = c(1,1,1,1))
     # split all the words and Word Cloud
    output$sentimentCloud <- renderWordcloud2(tidy_docs %>%inner_join(get_sentiments("bing")) %>%
      count(word, sentiment, sort = TRUE) %>%acast(word ~ sentiment, value.var = "n", fill = 0) %>%
      comparison.cloud(scale=c(4,.3), colors = c("#00B2FF", "red", "#FF0099", "#6600CC"), title.size = 1,
                       title.colors=c("red","blue"),title.bg.colors=c("grey40","grey70"),max.words = 100), 
      env = parent.frame(), quoted = FALSE)

    ####################################################################
    # polarity
    ####################################################################
    #calculate polarity on text column by grouping doc_id
    df_polarity <- polarity(text.var = df_data$text, 
                            grouping.var = df_data$doc_id, 
                            constrain = TRUE,
                            polarity.frame = key.pol,
                            negators = qdapDictionaries::negation.words,
                            amplifiers = qdapDictionaries::amplification.words, 
                            deamplifiers = qdapDictionaries::deamplification.words)
    polarity_score <- scores(df_polarity)
    polarity_res <- as.data.frame((counts(df_polarity)))

    ####################################################################
    # sentiment analysis
    ####################################################################
    df_data2 <- df_data$text
    data_vec <- as.vector(df_data2)
    df_emotion <- get_nrc_sentiment(data_vec)
    df_emotion_sub <- df_emotion[c(1:8)]

    #get the top emotions for each sentence
    df_emotion_sub$top_emotions <- apply(df_emotion_sub[-1], 1, FUN = function(x) 
      paste(names(head(sort(x[x!=0], decreasing = TRUE), 3)), collapse=", "))

    #build the final dataframe containing the results of polarity and sentiment analysis
    pos_vector <- as.character(polarity_res$pos.words)
    neg_vector <- as.character(polarity_res$neg.words)
    neg_words <- polarity_res$neg.words
    #creating the final data frame
    df_final <- data.frame(
      text = df_data$text,
      words = polarity_res$wc,
      positive_words = pos_vector,
      negative_words = neg_vector,
      polarity = polarity_score$ave.polarity,
      emotions = df_emotion_sub$top_emotions
    )

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

    df_final$Pos_Neg <- ifelse(df_final$polarity > 0,"Positive", 
                               ifelse(df_final$polarity < 0,"Negative","Neutral"))

    output$sentiment_Analysis <- renderPlot(ggplot(df_final, aes(x = factor(1), y = polarity)) +
      geom_boxplot(width = 0.4, fill = "white") +
      geom_jitter(aes(color = Pos_Neg, shape = Pos_Neg), 
                  width = 0.1, size = 1) +
      scale_color_manual(values = c("#00AFBB", "#E7B800", "tomato1")) + 
      labs(x = NULL))   # Remove x axis label

    #counting frequencies of emotions
    f<-data.frame(table(unlist(strsplit(tolower(df_final$emotions), ", "," ,")))) %>% mutate(Freq = sort(Freq))
    #######
    bar <- ggplot(data = f) +
      geom_bar(
        mapping = aes(x = Var1, fill = Freq),
        show.legend = TRUE,
        width = 1
      ) +
      theme(aspect.ratio = 1) +
      labs(x = NULL, y = NULL)

    output$topSentiments <- renderPlot(bar + coord_flip())

    output$sentimentScores <- renderDataTable(df_final[["text","positive_words","negative_words","polarity"]])
  )

感谢您帮我解决这个问题。

【问题讨论】：

请尽量使您的示例最小。尽可能多地删除与问题没有直接关系的代码。最有可能的是，当您将问题缩小时，更容易找到确切的原因。还要确保列出所有依赖项。例如wordcloud2Output 不需要来自shiny。当我把它拿出来时，它似乎运行没有错误。你传入的是什么文件？是不是上传后才会报错？

我想回应@MrFlick 的评论。这是非常复杂的代码，具有许多功能，因此难以评估。如果这对您来说是新的，我会从 shiny 开始。阅读reactivity。您会发现在observeEvent 中有许多嵌入式函数和output 语句可能不是要走的路，它将决定您将包含哪些reactive 函数。让我们知道这是否有帮助。

【参考方案1】：

你想要做的是用反应调用的服务器函数替换你的函数。像这样的东西。 而不是：

read_data_from_file <- function ()
      text <- reactive(readtext(input$fileInput))
      text <- toString(text[,2])
      text<- gsub("\n", "", text)
      text <- gsub("[\r\n]", "", text)
      return(text)
      

    check_data <- function ()
      if (is.null(read_data_from_file()) || is.null(searchText)) 
        return("File or Search Text is empty")
      
      else 
        return("Starting to analyze file")
      
    

你最好写：

read_data_from_file <- reactve(
      text <- reactive(readtext(input$fileInput))
      text <- toString(text[,2])
      text<- gsub("\n", "", text)
      text <- gsub("[\r\n]", "", text)
      return(text)
    )  

    check_data <- reactive(
      if (is.null(read_data_from_file()) || is.null(searchText() )) 
        return("File or Search Text is empty")
      
      else 
        return("Starting to analyze file")
      
    )

请不要在searchText 之后添加() 这是引用反应对象的方式。它们就像没有任何参数的函数。

希望这会有所帮助！