无法在 RStudio 中将大型 ish .RDS 文件部署到 Shinyapp.io [英] Can't deploy large ish .RDS files to shinyapp.io in RStudio
问题描述
我正在尝试从 RStudio 部署一个闪亮的应用程序.除了ui.r"和server.r"文件外，我还有 3 个 .RDS 文件(50kb、15mb 和 18mb)，我在 server.r 文件中读取这些文件，它们在本地运行良好.当我开始上传时，它无法给出以下错误:
I'm trying to deploy a shiny app from RStudio. Along with the "ui.r" and "server.r" files, I have 3 .RDS files (50kb, 15mb and 18mb), which I read in server.r file, and they work fine locally. When I commence upload, it fails giving the following error:
Preparing to deploy application...DONE
Uploading bundle for application: 794264...
Error in force(code) : Could not upload file.
Calls: <Anonymous> -> withStatus -> force
Execution halted
我尝试选择 50kb 和 15mb 文件，这很有效.当我单独选择 18 mb 时它可以工作，但是当我选择所有三个时，它似乎不起作用.我将文件大小增加到大约.500 mb(最有可能缩小包大小)，使用以下内容:
I tried selecting on the 50kb and 15mb file, that worked. When I select the 18 mb on its own it works, but when I select all the three of them, it doesn't seem to work. I increased the file size to approx. 500 mb (which was most likely downsizing the bundle size) using the following:
options(rsconnect.max.bundle.files = 500000000)
当我读到它时，它以字节为单位.
As I read it takes the value in bytes.
那是徒劳的.
我使用控制台进行部署:
I deployed using the console:
deployApp(appDir="Documents/GitHub/DataScience_JH_Coursera_Assignments/10.Capstone/shinyapp/PredictionText/",logLevel="verbose",lint=TRUE)
无济于事，但我设法将问题归零并得到以下日志:
To no avail, but I'm managed to zero in to the problematic line and I got the following log:
----- Deployment error -----
Error in force(code) : Could not upload file.
----- Error stack trace -----
4: stop("Could not upload file.")
3: force(code)
2: withStatus(paste0("Uploading bundle for ", assetTypeName, ": ",
application$id), {
bundlePath <- bundleApp(target$appName, appDir, appFiles,
appPrimaryDoc, assetTypeName, contentCategory, verbose)
if (isShinyapps(accountDetails)) {
bundleSize <- file.info(bundlePath)$size
checkSum <- md5sum(bundlePath)
bundle <- client$createBundle(application$id, "application/x-tar",
bundleSize, checkSum)
if (verbose)
timestampedLog("Starting upload now")
if (!uploadBundle(bundle, bundleSize, bundlePath)) {
stop("Could not upload file.")
}
if (verbose)
timestampedLog("Upload complete")
response <- client$updateBundleStatus(bundle$id, status = "ready")
bundle <- client$getBundle(bundle$id)
}
else {
bundle <- client$uploadApplication(application$id, bundlePath)
}
})
1: deployApp(appDir = "Documents/GitHub/DataScience_JH_Coursera_Assignments/10.Capstone/shinyapp/PredictionText/",
logLevel = "verbose", lint = TRUE)
所以它在这里失败了，我假设:
So it's failing here, I'm assuming:
if (!uploadBundle(bundle, bundleSize, bundlePath)) {
仅供参考，我的包大小设置为大约 3GB，如果我在得到以下内容时正确进行了转换:
FYI my bundle size is set to approx 3gb, if I did the conversion right when I got the following:
getOption("rsconnect.max.bundle.size")
[1] 3145728000
我已经坚持了 5 天了.我阅读了我可以阅读的所有论坛，尝试单独上传文件，却发现 Shiny/rsconnect 似乎在此之前删除了实例，因此无法逐个上传文件.我什至试图弄乱 .dcf rsconnect 文件，但那里也没有.任何帮助/见解将不胜感激.
I've been stuck on this for 5 days now. I read all the forums I can read, tried uploading the files individually only to find that shiny/rsconnect seems to delete the instance before that, hence not being able to upload files one-by-one. I even try to mess with .dcf rsconnect file and nothing there too. Any help/insight would be much appreciated.
推荐答案
我已经设法解决了这个问题，因为我将 *.RDS 文件与 server.r 和 ui.r 放在同一个文件夹中文件.我只需要在与 *.r 文件相同的文件夹中创建一个名为data"的文件夹，然后将 *.RDS 文件转储到那里并上传工作.
I've managed to solve the problem, as it turns out I had the *.RDS files in the same folder as the server.r and ui.r files. I just needed to create a folder called "data" in the same folder as *.r files and dump the *.RDS files there and the upload work.
日志本来可以更清楚一点.
The logs could've been a bit more clear.
另外，不相关，但它确实有帮助，刷新缓存"，导航到 rsconnect 和所有后续文件夹，直到到达您的 [appname].dcf 文件并删除该文件.
Also, not related, but it did help, to "flush the cache", navigate into the rsconnect and all subsequent folders till you reach your [appname].dcf file and delete that file.
这篇关于无法在 RStudio 中将大型 ish .RDS 文件部署到 Shinyapp.io的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
