IronPython内存释放问题
原始代码:
ScriptEngine eng = IronPython.Hosting.Python.CreateEngine();
优化后的代码:
var options = new Dictionary<string, object> { ["LightweightScopes"] = true }; ScriptEngine eng = IronPython.Hosting.Python.CreateEngine(AppDomain.CurrentDomain, options); var scope = eng.CreateScope(); using (var streamOut = new MemoryStream()) { ICollection<string> Paths = eng.GetSearchPaths(); Paths.Add(Environment.CurrentDirectory + @"\Lib"); Paths.Add(Environment.CurrentDirectory + @"\Lib\site-packages"); eng.SetSearchPaths(Paths); eng.Runtime.IO.SetOutput(streamOut, Encoding.ASCII); var pycode="the python code you need execute";//你需要执行的python代码 eng.Execute(pycode, scope); string printS = Encoding.UTF8.GetString(streamOut.ToArray());//python代码执行的print输出结果 eng.Runtime.Shutdown(); }
注意:
1.GC.Collect()对此处内存回收作用不大。
2.如果需要循环调用python,不建议重复CreateEngine,运行速度会变慢,可使用下面代码生成ScriptEngine:
private ScriptEngine _eng; public ScriptEngine eng { get { if (_eng != null) { return _eng; } else { var options = new Dictionary<string, object> { ["LightweightScopes"] = true }; _eng = Python.CreateEngine(AppDomain.CurrentDomain, options); return _eng; } } set { _eng = value; } }
3.调用eng.Runtime.Shutdown();关闭ScriptEngine的Runtime。
4.如果通过下面代码的方式获取ScriptEngine ,会造成无法获取python的输出结果(也就是上面代码的printS为空)。
ScriptEngine eng
{
get
{
var options = new Dictionary<string, object> { ["LightweightScopes"] = true };
return IronPython.Hosting.Python.CreateEngine(AppDomain.CurrentDomain, options);
}
}