Azure (C#) System.IO.FileNotFoundException: Could not find file - visual-studio

Question

I've created a ML model with Visual Studio. I uploaded the web app to Azure with Visual Studio too. However, when I fill the fields for my ML model and click "run" on the website, I get this error which I copied directly from Azure App Service Editor.
I only get this error while trying to run the ML model on Azure website, if I run the web app on my computer I have no errors at all.
Thank you :)
The error:
2020-07-18 01:12:59.138 +00:00 [Error] Microsoft.AspNetCore.Diagnostics.ExceptionHandlerMiddleware: An unhandled exception has occurred while executing the request.
System.IO.FileNotFoundException: Could not find file 'C:\Users\X\X\X\fileML.Model\MLModel.zip'.
File name: 'C:\Users\X\X\X\fileML.Model\MLModel.zip'
____________________
My code:
// This file was auto-generated by ML.NET Model Builder.
using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Text;
using Microsoft.ML;
using fileML.Model;
namespace fileML.Model
{
public class ConsumeModel
{
private static readonly Lazy<PredictionEngine<ModelInput, ModelOutput>> PredictionEngine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);
// For more info on consuming ML.NET models, visit https://aka.ms/mlnet-consume
// Method for consuming model in your app
public static ModelOutput Predict(ModelInput input)
{
ModelOutput result = PredictionEngine.Value.Predict(input);
return result;
}
public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
{
// Create new MLContext
MLContext mlContext = new MLContext();
// Load model & create prediction engine
string modelPath = #"C:\Users\X\X\X\fileML.Model\MLModel.zip";
ITransformer mlModel = mlContext.Model.Load(modelPath, out _);
var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
return predEngine;
}
}
}

Answer 1

Nathan, welcome to the stackoverflow. Here is thing you are missing:
You are trying to access local path from your computer but on Azure there is no local machine so whenever you code tries to access the same path which you have hard coded it's resulting in error.
My recommendation would be to add your zip file to your project, once added right click on that file and mark Copy to Output Directory - Copy always. Please see below
This will help to get the local file path from output directory.
Now it's time to change your code to get file dynamically.
You can use
string directoryPath = Directory.GetCurrentDirectory();
string modelPath= Path.Combine(directoryPath ,"MLModel.zip");
This will get you the file path. Just do a test your code on your local and deploy the app.
The good thing is now your model file will get deployed along with your code. Every time you change your model just replace the file and deploy code again.
Hint to make it more dynamic:- You can also use Azure Blob Storage to keep your zip file, by using this you do not need to deploy your code again and again . Just need to replace the file in side blob.

Answer 2

you are trying to load file MLModel.zip from C:\Users\X\X\X\fileML.Model. Now that's your local computer path. That path not exists into Azure Web App.
There are 2 ways you can do if you really want to store in local directory:
HOMe environment variable in your Azure Web App that resolves to the
equivalent of inetpub for your site. Your app data folder is located
at %HOME%\site\wwwroot\AppData.
TEMP environment both on Azure Web Apps and on your local machine.

Answer 3

public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
{
// Create new MLContext
MLContext mlContext = new MLContext();
// Load model & create prediction engine
string directoryPath = Directory.GetCurrentDirectory();
string modelPath = Path.Combine(#"C:\Users\Admin\source\repos\ShanuASPMLNETML.Model\MLModel.zip","MLModel.zip");
ITransformer mlModel = mlContext.Model.Load(modelPath, out _);
var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
return predEngine;
}
}