JupyterLab, TypeError: getOwnMetadataKeys is not a function

[manoj30075]

I’m currently developing an extension for JupyterLab. As part of this process, I’m utilizing the InversifyJS library from npm. To use it, I have to import ‘reflect-metadata’ in my index.ts file, which I have done. However, when I create an injectable class and use it in my extension, I encounter the following error:

TypeError: getOwnMetadataKeys is not a function
    at Object.isProviderFor (Reflect.js:1135:1)
    at getProviderNoCache (Reflect.js:941:1)
    at Object.getProvider (Reflect.js:955:1)
    at GetMetadataProvider (Reflect.js:1158:1)
    at OrdinaryDefineOwnMetadata (Reflect.js:611:1)
    at decorator (Reflect.js:195:1)
    at DecorateConstructor (Reflect.js:549:1)
    at Reflect.decorate (Reflect.js:143:1)
    at push.../shared/components/localTelemetry.ts.__decorate (icons.ts:4:72)
    at ../shared/components/localTelemetry.ts (localTelemetry.ts:7:28)

This is part of my code

@injectable()
export class LocalTelemetry implements TelemetryService {
  constructor(
    @inject(NAMES.customerId)
    private customerId: string = '',
    @inject(NAMES.outputChannelTelemetry)
    private outputChannel: Logger
  ) {}

And I have imported the reflect-metadata at the index.ts 1st line.

Upon researching a solution for this error, the majority of sources suggest importing reflect-metadata at the root level file, which I have already done. I wrote Jest test cases for the Injectable class and the test cases are passing when I run them in the terminal, but when I run the extension in the browser, I get the above error. I’m uncertain about what I might be doing incorrectly.

[PodaruDragos]

there could be more than 1 reasons for this one. I assume you have the tsconfig emitMetadata and you are using at least 4.5 version, right ?

[manoj30075]

Yes, I have those configurations set in tsconfig and using version 5.2.2. The strange thing that I found out while debugging is that I created another Jupyter extension just to test InversifyJS. I added a simple injectable class. When I installed that sample extension, my main extension started working fine, but the sample extension threw this error in the browser console: Error: Missing required @Injectable annotation in: DependencyInjector. Then, when I uninstalled the sample extension, the old error returned for the main extension. Now I'm not sure what's the issue here.

[femisystems]

@manoj30075 were you eventually able to fix this problem? I also have a similar problem right now with Jupyterlab 4. With Jupyterlab3, it works well.

[notaphplover]

Hey, just open a new issue with a minimum recreation code and I'll have a look at it