TypeError: Expected signal to be an instanceof AbortSignal when using SDK and CustomVision libraries

We are trying to create a project on CustomVision using SDK. We are using the following libraries: import { TrainingAPIClient } from "@azure/cognitiveservices-customvision-training"; import { PredictionAPIClient } from "@azure/cognitiveservices-customvision-prediction"; import { ApiKeyCredentials } from "@azure/ms-rest-js";

here are the versions: "@azure/cognitiveservices-customvision-prediction": "^5.1.2", "@azure/cognitiveservices-customvision-training": "^5.2.0", "@azure/ms-rest-azure-js": "^2.1.0", Unfortunately, no matter what i do, i keep getting this error: TypeError: Expected signal to be an instanceof AbortSignal

We tried switching to "@azure/ms-rest-azure-js"; which did not help.

We tried passing AbortSignal but this did not work as well.

Current versions of these libraries do not support modern libs like core-auth so we can't use that.

Any other ideas?

The error Expected signal to be an instance of AbortSignal typically occurs when using abort-controller in the following format:

const controller = new AbortController();
doAsyncWork({ abortSignal: controller.signal });
controller.abort();

Note :
Refer to these git and MSDOC On the same issue .

To avoid the conflict, install the AbortSignal library (or remove the part of the code that contains it).AbortSignal .

npm install @azure/abort-controller

npm install abort-controller

By referring to the doc, I have executed the following code for creating a project, adding tags, uploading images, training the model, and publishing the iteration using Azure Custom Vision.

const  util = require('util');
const  fs = require('fs');
const TrainingApi = require("@azure/cognitiveservices-customvision-training");
const PredictionApi = require("@azure/cognitiveservices-customvision-prediction");
const msRest = require("@azure/ms-rest-js");

const trainingKey = "VISION_TRAINING_KEY";
const trainingEndpoint = "VISION_TRAINING_ENDPOINT";

const VISIONKey = "VISION_PREDICTION_KEY";
const predictionResourceId ="VISION_PREDICTION_RESOURCE_ID";
const predictionEndpoint = "VISION_PREDICTION_ENDPOINT";

const publishIterationName = "classifyModel";
const setTimeoutPromise = util.promisify(setTimeout);

(async () => {
    console.log("Creating project...");
    const credentials = new msRest.ApiKeyCredentials({ inHeader: { "Training-key": trainingKey } });
    const trainer = new TrainingApi.TrainingAPIClient(credentials, trainingEndpoint);
    const sampleProject = await trainer.createProject("Sample Project");

    console.log("Adding tags...");
    const hemlockTag = await trainer.createTag(sampleProject.id, "Hemlock");
    const cherryTag = await trainer.createTag(sampleProject.id, "Japanese Cherry");

    console.log("Adding images...");
    const sampleDataRoot = "Images";
    let fileUploadPromises = [];

    const hemlockDir = `${sampleDataRoot}/Hemlock`;
    const hemlockFiles = fs.readdirSync(hemlockDir);
    hemlockFiles.forEach(file => {
        fileUploadPromises.push(trainer.createImagesFromData(sampleProject.id, fs.readFileSync(`${hemlockDir}/${file}`), { tagIds: [hemlockTag.id] }));
    });

    const cherryDir = `${sampleDataRoot}/Japanese_Cherry`;
    const japaneseCherryFiles = fs.readdirSync(cherryDir);
    japaneseCherryFiles.forEach(file => {
        fileUploadPromises.push(trainer.createImagesFromData(sampleProject.id, fs.readFileSync(`${cherryDir}/${file}`), { tagIds: [cherryTag.id] }));
    });

    await Promise.all(fileUploadPromises);

    console.log("Training...");
    let trainingIteration = await trainer.trainProject(sampleProject.id);

    console.log("Training started...");
    while (trainingIteration.status == "Training") {
        console.log("Training status: " + trainingIteration.status);
        await setTimeoutPromise(1000, null);
        trainingIteration = await trainer.getIteration(sampleProject.id, trainingIteration.id);
    }
    console.log("Training status: " + trainingIteration.status);

    await trainer.publishIteration(sampleProject.id, trainingIteration.id, publishIterationName, predictionResourceId);

    const predictor_credentials = new msRest.ApiKeyCredentials({ inHeader: { "Prediction-key": VISIONKey } });
    const predictor = new PredictionApi.PredictionAPIClient(predictor_credentials, predictionEndpoint);
    const testFile = fs.readFileSync(`${sampleDataRoot}/Test/test_image.jpg`);

    const results = await predictor.classifyImage(sampleProject.id, publishIterationName, testFile);
    console.log("Results:");
    results.predictions.forEach(predictedResult => {
        console.log(`\t ${predictedResult.tagName}: ${(predictedResult.probability * 100.0).toFixed(2)}%`);
    });
})();