feat: convert to TypeScript

[alexander-fenster]

This one is complicated. It has a very weird dark magic with a helper replacing one of the auto-generated method:

nodejs-speech/src/index.js

Lines 53 to 57 in cace910

	// Augment the SpeechClient objects with the helpers.
	for (const gapicVersion of Object.keys(gapic)) {
	const clientProto = gapic[gapicVersion].SpeechClient.prototype;
	Object.assign(clientProto, helpers());
	}

I'm replacing it with an equal (but probably not so dark) magic in TypeScript, making kind of a mixin.

nodejs-speech/src/index.ts

Lines 19 to 36 in b4c87e7

	// The following code is adapted from http://www.typescriptlang.org/docs/handbook/mixins.html
	// tslint:disable-next-line no-any
	Object.defineProperty(
	v1.SpeechClient.prototype,
	'streamingRecognize',
	Object.getOwnPropertyDescriptor(
	ImprovedStreamingClient.prototype,
	'streamingRecognize'
	)!
	);
	Object.defineProperty(
	v1p1beta1.SpeechClient.prototype,
	'streamingRecognize',
	Object.getOwnPropertyDescriptor(
	ImprovedStreamingClient.prototype,
	'streamingRecognize'
	)!
	);

nodejs-speech/src/v1/speech_client.ts

Lines 477 to 478 in b4c87e7

	import {ImprovedStreamingClient} from '../helpers';
	export interface SpeechClient extends ImprovedStreamingClient {}

Other than that, looks good to me, and all the system tests just work.

[alexander-fenster]

(figuring out how to make samples linted correctly)

[codecov]

Codecov Report

Merging #472 into master will increase coverage by 21.68%.
The diff coverage is 91.49%.

@@             Coverage Diff             @@
##           master     #472       +/-   ##
===========================================
+ Coverage   69.37%   91.06%   +21.68%     
===========================================
  Files          17       11        -6     
  Lines         209     2932     +2723     
  Branches        0       74       +74     
===========================================
+ Hits          145     2670     +2525     
- Misses         64      260      +196     
- Partials        0        2        +2

Impacted Files	Coverage Δ
src/v1/index.ts	100% <100%> (ø)
src/index.ts	100% <100%> (ø)
src/helpers.ts	97.67% <100%> (ø)
src/v1p1beta1/index.ts	100% <100%> (ø)
src/v1/speech_client.ts	93.09% <90.03%> (ø)
src/v1p1beta1/speech_client.ts	93.11% <90.07%> (ø)
samples/betaFeatures.js	82.13% <0%> (ø)
samples/recognize.js	93.96% <0%> (ø)
... and 7 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update cace910...25c718b. Read the comment docs.

[bcoe]

can we just replace this with c8 mocha build/test/*.js, it seems like this has weird bespoke coverage logic (perhaps this was the first module we added coverage to).

[bcoe]

What are the main benefits of the ImprovedStreamingClient mix-in? better types in the response?

I wonder if we should consider upstreaming some of this to the generator itself, worth making a tracking ticket?

[alexander-fenster]

This looks like a speech-specific change. They want streamingRecognize, which is a bi-di streaming method, to accept a separate config object:

  streamingRecognize(
    streamingConfig?:
      | protosTypes.google.cloud.speech.v1.IStreamingRecognitionConfig
      | protosTypes.google.cloud.speech.v1p1beta1.IStreamingRecognitionConfig,
    options?: gax.CallOptions
  )

By our design, bi-directional streaming calls don't accept any request as a parameter (only as a stream). This one is unique.