Question

我正在开发一个R包，其中包含以下功能：

#' @export
makeEnvironment = function(name) {
  envir = R6::R6Class(
    public = list(
      initialize = function(name) {
        remote.base = "http://127.0.0.1:5000"
        private$client = gym::create_GymClient(remote.base)
        private$instance.id = gym::env_create(private$client, name)
        outdir = "/tmp/random-agent-results"
        gym::env_monitor_start(private$client, private$instance.id, outdir, 
          force = TRUE, resume = FALSE)
        print(gym::env_action_space_info(private$client, private$instance.id))
      }
    ),
    private = list(
      client = NULL,
      instance.id = NULL
    )
  )
  package.path = system.file(package = "reinforcelearn")
  path2pythonfile = paste0(package.path, "/gym_http_server.py")
  system2("python", args = path2pythonfile, stdout = NULL, wait = FALSE)
  envir$new(name)
}

我可以在此处找到我调用的python脚本：https://github.com/openai/gym-http-api/blob/master/gym_http_server.py

当我启动RStudio时，加载包并运行makeEnvironment("MountainCar-v0")我收到错误

curl :: curl_fetch_memory（url，handle = handle）出错：无法执行连接到127.0.0.1端口5000：拒绝连接

由gym::env_create(private$client, name)引起。如果我再次重复makeEnvironment("MountainCar-v0")，它就可以了。

有人可以帮忙吗？为什么我会收到此错误以及如何解决此问题？

会话信息

R版本3.4.1（2017-06-30）平台：x86_64-w64-mingw32 / x64（64位）   运行于：Windows＆gt; = 8 x64（build 9200）

Matrix产品：默认

locale：[1] LC_COLLATE = German_Germany.1252   LC_CTYPE = German_Germany.1252 LC_MONETARY = German_Germany.1252 [4]   LC_NUMERIC = C LC_TIME = German_Germany.1252

附加基础包：[1] stats graphics grDevices utils
  数据集方法基础

其他附件包：[1] reinforcelearn_0.0.0.9070

通过命名空间加载（而不是附加）：[1] httr_1.2.1
  compiler_3.4.1 backports_1.1.0 R6_2.2.2 Matrix_1.2-10
  tools_3.4.1 [7] nnet_7.3-12 curl_2.8.1
  MDPtoolbox_4.0.3 grid_3.4.1 checkmate_1.8.3 jsonlite_1.5
  [13] gym_0.1.0 linprog_0.9-2 lattice_0.20-35

curl :: curl_fetch_memory（url，handle = handle）出错：无法连接到127.0.0.1端口5000：拒绝连接

0 个答案: