r - 闪亮的帮助：发布和“运行应用程序”将“变灰”，但使用 shinyApp() 作品
问题描述
我正在开发我的第一个 Shiny 应用程序，当我在本地运行它时它看起来很棒，但如果使用 Rstudio“运行应用程序”按钮或通过 shinyapp.io 发布时会失败。
我确保在通过 shinyapp.io 发布时该目录不是本地目录，并且 csv 文件位于应用程序文件夹中。我使用 profvis 检查了应用程序的性能，该应用程序使用 352.9 MB 内存和 1060 毫秒。
这是ui文档：
library(shiny)
library(ggplot2)
library(readr)
library(tidyr)
library(data.table)
library(reshape2)
library(dplyr)
library(DT)
library(shinyWidgets)
library(ggrepel)
library(rsconnect)
library(devtools)
#master document. replaced with github data for sharing
df1 <- read_csv("https://raw.githubusercontent.com/KeithTStory/EnvironmentalAnalytics/master/Shiny/CABIN_explorer/20190522(2)_cabin_habitat_data_mda05_1987-present.csv",
col_types = cols_only(Site = "c", Day = "i", Year = "i", Status = "c", Type = "c", Variable = "c", Unit = "c", Value = "n"),
locale = locale(encoding = "latin1")
)
attach(df1) #odd that I am having to attach names. Seems to work with names attached though.
#Melt data to give each variable a column for selecting
df2 <- df1 %>%
mutate(id = row_number()) %>%
replace(., is.na(.), 0) %>%
filter(Site != 0, Value > 0) %>%
spread(key = Variable, value = Value) %>%
group_by(Year)
attach(df2)
#remove symbols that are causing issues. Tried to convert encoding but had no success.
colnames(df2) <- gsub(x = colnames(df2), pattern = "[+%-]", replacement = "")
colnames(df2) <- gsub(x = colnames(df2), pattern = " ", replacement = ".")
df1$Variable <-gsub(x = df1$Variable, pattern = "[+%-]", replacement = "")
df1$Variable <-gsub(x = df1$Variable, pattern = " ", replacement = ".")
#This function is used to in the ui input choices for the y axis
FilterFunc <- function(condition, df) {
subset.data <- df %>%
filter(Type == condition)
data_list <- unique(subset.data$Variable)
return(as.character(data_list))
}
#useful variables for ui choices
types <- unique(as.factor(df1$Type))
df_sites <- unique(df2$Site)
### ui code_____________________________________________
ui <- fluidPage(
titlePanel("CABIN Data Explorer (Nelson River)"),
sidebarPanel(
selectInput(inputId = "site",
label = "Select Sampling Site:",
choices = df_sites,
selectize = TRUE,
multiple = TRUE,
selected = "521"
),
selectInput(inputId = "y",
label = "Y-axis:",
choices = list(
"Physical Data" = FilterFunc(types[1], df1), #pulling from df1 since it is pre-spread and simpler to work with
"Sediment Chemistry" = FilterFunc(types[2], df1),
"Water Chemistry" = FilterFunc(types[3], df1),
"Channel" = FilterFunc(types[4], df1),
"Climate" = FilterFunc(types[5], df1),
"Substrate Data" = FilterFunc(types[6], df1),
"Hydrology" = FilterFunc(types[7], df1),
"Topography" = FilterFunc(types[8], df1),
"Landcover" = FilterFunc(types[9], df1),
"Bedrock Geology " = FilterFunc(types[10], df1)
),
selectize = TRUE,
selected = "Ca"
)
),
mainPanel(
plotOutput('histplot'),
DT::dataTableOutput(outputId = "site_table")
)
)
和服务器文档：
#
server <- function(input, output) {
DF <- reactive({
df2 %>%
select(Site, Day, Year, Unit, Value = input$y) %>%
filter(Site %in% input$site) %>%
na.omit()
})
output$histplot <- renderPlot({
req(input$site)
req(input$y)
p <- ggplot(data = DF(), aes(y = Value, x = Year, label = Site, color = Site)) +
geom_point(stat = "identity",
size = 3) +
geom_label_repel(fill = "white",
size = 6) +
scale_x_continuous(limits = c(2005, 2018)) +
# scale_y_continuous(limits = c(0, max(Var))) +
theme_bw() +
theme(panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
legend.position = "none",
panel.background = element_rect(fill = "floralwhite")
) +
facet_wrap(~.Unit) +
labs(ylab(input$y))
print(p)
}, height=300)
output$site_table <- DT::renderDataTable({
req(input$site)
req(input$y)
DT::datatable(data = DF(),
options = list(pageLength = 10),
rownames = FALSE)
})
}
shinyApp(ui, server)
当我在本地运行 shinyApp(ui, server) 时，一切看起来都很棒。
当我在 RStudio 中使用“运行应用程序”按钮时，我收到以下消息：
Warning: Error in serverFuncSource: server.R returned an object of unexpected type: list
[No stack trace available]
Error in serverFuncSource() :
server.R returned an object of unexpected type: list
当我使用 RStudio 中的发布按钮发布应用程序时，Web 文档开始加载，但随后“灰显”并显示“与服务器断开连接”，并可选择重新加载。
检查控制台输出会产生对我不知道如何解释的 js 错误的引用（也许您可以重现并指导我？）。
感谢您的时间和任何建议！（欢迎提供有关如何解释这些错误的信息）。