Skip to content

armgit / hypit

Public Imported from GitHub
0

Clone any viral video with AI agents. Not just a script, the whole workflow: swap the face, the words, the B-roll, ship 100 variants in one command, and get your 100M views.

feat(whisperx): support Spanish alignment

kashorin2350bebcdcb9
diff --git a/packages/provider-whisperx-local/src/provider.ts b/packages/provider-whisperx-local/src/provider.ts
index 98acbd28..1be0a65a 100644
--- a/packages/provider-whisperx-local/src/provider.ts
+++ b/packages/provider-whisperx-local/src/provider.ts
@@ -72,7 +72,7 @@ function alignmentRequest(value: CanonicalValue): WhisperXAlignmentRequest {
&& item.audio.mediaType === "audio/wav"
&& Number.isSafeInteger(item.sampleFrames)
&& item.sampleFrames > 0
- && (item.language === "en" || item.language === "zh"),
+ && (item.language === "en" || item.language === "zh" || item.language === "es"),
"WhisperX alignment request is invalid");
return item;
}
diff --git a/packages/whisperx/src/component.ts b/packages/whisperx/src/component.ts
index d1c0e834..70cc5c3b 100644
--- a/packages/whisperx/src/component.ts
+++ b/packages/whisperx/src/component.ts
@@ -19,7 +19,9 @@ export const whisperXComponent = {
handler: ({ inputs }) => {
const evidence = inline(inputs.evidence!.value, "SpeechEvidenceAudio") as unknown as SpeechEvidenceAudio;
const language = inline(inputs.language!.value, "WhisperXLanguage") as unknown as WhisperXLanguage;
- if (language !== "en" && language !== "zh") throw new Error("WhisperXLanguage must be en or zh");
+ if (language !== "en" && language !== "zh" && language !== "es") {
+ throw new Error("WhisperXLanguage must be en, zh, or es");
+ }
return {
outputs: {},
needs: { alignment: canonicalize(whisperXRequestForEvidenceAudio(evidence, { language })) },
diff --git a/packages/whisperx/src/manifest.ts b/packages/whisperx/src/manifest.ts
index 7be30f75..24defab0 100644
--- a/packages/whisperx/src/manifest.ts
+++ b/packages/whisperx/src/manifest.ts
@@ -38,8 +38,8 @@ export const whisperXMarkupSurfaces = [{
accepts: [mediaTypes.synchronized],
summary: "Selects the already normalized SynchronizedMedia measured by WhisperX." },
{ name: "language", kind: "literal", required: true,
- values: ["en", "zh"],
- summary: "Explicitly selects the English or Chinese WhisperX transcription and alignment models." },
+ values: ["en", "zh", "es"],
+ summary: "Explicitly selects the English, Chinese or Spanish WhisperX transcription and alignment models." },
],
ports: [
{ name: "take", type: speechTypes.semanticTake,
@@ -49,7 +49,7 @@ export const whisperXMarkupSurfaces = [{
segment={story.segment.opening} media={opening-media.media} language="en"/>`,
notes: [
"All five attributes are required; the element accepts no children and no text content.",
- "Language is never detected from Script text or audio; each alignment call states en or zh explicitly.",
+ "Language is never detected from Script text or audio; each alignment call states en, zh or es explicitly.",
"Importing this package is what selects the WhisperX model family; the Runtime separately binds the alignment Need to an Endpoint.",
],
},
diff --git a/packages/whisperx/src/surface.ts b/packages/whisperx/src/surface.ts
index 0a4ae08f..957a35e8 100644
--- a/packages/whisperx/src/surface.ts
+++ b/packages/whisperx/src/surface.ts
@@ -38,8 +38,8 @@ export const decodeWhisperXSemanticTakeSurface: StructuredSurfaceHandler = ({ el
const segment = reference(element, "segment", narrativeTypes.excerpt, resolveReference);
const media = reference(element, "media", mediaTypes.synchronized, resolveReference);
const language = stringAttribute(element, "language");
- if (language !== "en" && language !== "zh") {
- throw new Error(`${element.name}.language must be en or zh`);
+ if (language !== "en" && language !== "zh" && language !== "es") {
+ throw new Error(`${element.name}.language must be en, zh, or es`);
}
const languageId = `${id}.language`;
return {
diff --git a/packages/whisperx/src/types.ts b/packages/whisperx/src/types.ts
index 1b0482d0..1d9da5c8 100644
--- a/packages/whisperx/src/types.ts
+++ b/packages/whisperx/src/types.ts
@@ -1,6 +1,6 @@
import type { SpeechEvidenceAudio } from "@hypit/speech";
-export type WhisperXLanguage = "en" | "zh";
+export type WhisperXLanguage = "en" | "zh" | "es";
export type WhisperXAlignmentRequest = {
readonly audio: SpeechEvidenceAudio["artifact"];
diff --git a/packages/whisperx/test/whisperx.test.ts b/packages/whisperx/test/whisperx.test.ts
index 098eb402..61be9aee 100644
--- a/packages/whisperx/test/whisperx.test.ts
+++ b/packages/whisperx/test/whisperx.test.ts
@@ -29,11 +29,11 @@ function evidenceAudio(): SpeechEvidenceAudio {
test("WhisperX receives normalized bytes without authored Segment truth", () => {
const valid = evidenceAudio();
- const request = whisperXRequestForEvidenceAudio(valid, { language: "zh" });
+ const request = whisperXRequestForEvidenceAudio(valid, { language: "es" });
assert.equal(request.audio.digest, valid.artifact.digest);
assert.equal("segments" in request, false);
assert.equal(request.sampleFrames, 16_000);
- assert.equal(request.language, "zh");
+ assert.equal(request.language, "es");
});
test("WhisperX installs into the host-neutral compute port without a Node Driver", () => {