[Inference] Remove inference vcr tests (#1375)

same as huggingface/huggingface_hub#3021. 
I’ve just skipped the tests for now and removed VCR related code so that
we can quickly re-enable or adapt them when we introduce proper
inference tests, as mentioned by @Wauplin.

Author: hanouticelina

.github/workflows/test.yml

@@ -38,7 +38,7 @@ jobs:
           pnpm --filter ...[${{ steps.since.outputs.SINCE }}]... build
 
       - name: Test
-        run: VCR_MODE=playback pnpm --filter ...[${{ steps.since.outputs.SINCE }}] test
+        run: pnpm --filter ...[${{ steps.since.outputs.SINCE }}] test
         env:
           HF_TOKEN: ${{ secrets.HF_TOKEN }}
 
@@ -76,7 +76,7 @@ jobs:
           pnpm --filter ...[${{ steps.since.outputs.SINCE }}]... build
 
       - name: Test in browser
-        run: VCR_MODE=playback pnpm --filter ...[${{ steps.since.outputs.SINCE }}] test:browser
+        run: pnpm --filter ...[${{ steps.since.outputs.SINCE }}] test:browser
         env:
           HF_TOKEN: ${{ secrets.HF_TOKEN }}
 

packages/inference/test/InferenceClient.spec.ts

@@ -14,7 +14,6 @@ import {
 import { isUrl } from "../src/lib/isUrl";
 import { HARDCODED_MODEL_INFERENCE_MAPPING } from "../src/providers/consts";
 import { readTestFile } from "./test-files";
-import "./vcr";
 
 const TIMEOUT = 60000 * 3;
 const env = import.meta.env;
@@ -23,7 +22,7 @@ if (!env.HF_TOKEN) {
 	console.warn("Set HF_TOKEN in the env to run the tests for better rate limits");
 }
 
-describe.concurrent("InferenceClient", () => {
+describe.skip("InferenceClient", () => {
 	// Individual tests can be ran without providing an api key, however running all tests without an api key will result in rate limiting error.
 
 	describe("backward compatibility", () => {