feat(server): support sliced audio (#11562)

fix AF-2479

packages/backend/server/src/__tests__/__snapshots__/copilot.e2e.ts.md

@@ -88,3 +88,47 @@ Generated by [AVA](https://avajs.dev).
         ],
       },
     ]
+
+> should submit audio transcription job
+
+    [
+      {
+        status: 'running',
+      },
+    ]
+
+> should claim audio transcription job
+
+    [
+      {
+        status: 'claimed',
+        summary: '[{"a":"A","s":30,"e":45,"t":"Hello, everyone."},{"a":"B","s":46,"e":70,"t":"Hi, thank you for joining the meeting today."}]',
+        title: '[{"a":"A","s":30,"e":45,"t":"Hello, everyone."},{"a":"B","s":46,"e":70,"t":"Hi, thank you for joining the meeting today."}]',
+        transcription: [
+          {
+            end: '00:00:45',
+            speaker: 'A',
+            start: '00:00:30',
+            transcription: 'Hello, everyone.',
+          },
+          {
+            end: '00:01:10',
+            speaker: 'B',
+            start: '00:00:46',
+            transcription: 'Hi, thank you for joining the meeting today.',
+          },
+          {
+            end: '00:10:45',
+            speaker: 'A',
+            start: '00:10:30',
+            transcription: 'Hello, everyone.',
+          },
+          {
+            end: '00:11:10',
+            speaker: 'B',
+            start: '00:10:46',
+            transcription: 'Hi, thank you for joining the meeting today.',
+          },
+        ],
+      },
+    ]

packages/backend/server/src/__tests__/copilot.e2e.ts

@@ -883,34 +883,67 @@ test('should be able to transcript', async t => {
     '[{"a":"A","s":30,"e":45,"t":"Hello, everyone."},{"a":"B","s":46,"e":70,"t":"Hi, thank you for joining the meeting today."}]'
   );
 
-  const job = await submitAudioTranscription(
-    app,
-    workspaceId,
-    'blobId',
-    'test.mp3',
-    Buffer.from([1, 1])
-  );
-  t.snapshot(
-    cleanObject([job], ['id']),
-    'should submit audio transcription job'
-  );
-  t.truthy(job.id, 'should have job id');
-
-  // wait for processing
   {
-    let { status } = (await audioTranscription(app, workspaceId, job.id)) || {};
+    const job = await submitAudioTranscription(app, workspaceId, '1', '1.mp3', [
+      Buffer.from([1, 1]),
+    ]);
+    t.snapshot(
+      cleanObject([job], ['id']),
+      'should submit audio transcription job'
+    );
+    t.truthy(job.id, 'should have job id');
 
-    while (status !== 'finished') {
-      await new Promise(resolve => setTimeout(resolve, 1000));
-      ({ status } = (await audioTranscription(app, workspaceId, job.id)) || {});
+    // wait for processing
+    {
+      let { status } =
+        (await audioTranscription(app, workspaceId, job.id)) || {};
+
+      while (status !== 'finished') {
+        await new Promise(resolve => setTimeout(resolve, 1000));
+        ({ status } =
+          (await audioTranscription(app, workspaceId, job.id)) || {});
+      }
+    }
+
+    {
+      const result = await claimAudioTranscription(app, job.id);
+      t.snapshot(
+        cleanObject([result], ['id']),
+        'should claim audio transcription job'
+      );
     }
   }
 
   {
-    const result = await claimAudioTranscription(app, job.id);
+    // sliced audio
+    const job = await submitAudioTranscription(app, workspaceId, '2', '2.mp3', [
+      Buffer.from([1, 1]),
+      Buffer.from([1, 2]),
+    ]);
     t.snapshot(
-      cleanObject([result], ['id']),
-      'should claim audio transcription job'
+      cleanObject([job], ['id']),
+      'should submit audio transcription job'
     );
+    t.truthy(job.id, 'should have job id');
+
+    // wait for processing
+    {
+      let { status } =
+        (await audioTranscription(app, workspaceId, job.id)) || {};
+
+      while (status !== 'finished') {
+        await new Promise(resolve => setTimeout(resolve, 1000));
+        ({ status } =
+          (await audioTranscription(app, workspaceId, job.id)) || {});
+      }
+    }
+
+    {
+      const result = await claimAudioTranscription(app, job.id);
+      t.snapshot(
+        cleanObject([result], ['id']),
+        'should claim audio transcription job'
+      );
+    }
   }
 });

packages/backend/server/src/__tests__/utils/copilot.ts

@@ -335,17 +335,17 @@ export async function submitAudioTranscription(
   workspaceId: string,
   blobId: string,
   fileName: string,
-  content: Buffer
+  content: Buffer[]
 ): Promise<{ id: string; status: string }> {
-  const res = await app
+  let resp = app
     .POST('/graphql')
     .set({ 'x-request-id': 'test', 'x-operation-name': 'test' })
     .field(
       'operations',
       JSON.stringify({
         query: `
-          mutation submitAudioTranscription($blob: Upload!, $blobId: String!, $workspaceId: String!) {
-            submitAudioTranscription(blob: $blob, blobId: $blobId, workspaceId: $workspaceId) {
+          mutation submitAudioTranscription($blob: Upload, $blobs: [Upload!], $blobId: String!, $workspaceId: String!) {
+            submitAudioTranscription(blob: $blob, blobs: $blobs, blobId: $blobId, workspaceId: $workspaceId) {
               id
               status
             }
@@ -353,17 +353,29 @@ export async function submitAudioTranscription(
         `,
         variables: {
           blob: null,
+          blobs: [],
           blobId,
           workspaceId,
         },
       })
     )
-    .field('map', JSON.stringify({ '0': ['variables.blob'] }))
-    .attach('0', content, {
+    .field(
+      'map',
+      JSON.stringify(
+        Array.from<any>({ length: content.length }).reduce((acc, _, idx) => {
+          acc[idx.toString()] = [`variables.blobs.${idx}`];
+          return acc;
+        }, {})
+      )
+    );
+  for (const [idx, buffer] of content.entries()) {
+    resp = resp.attach(idx.toString(), buffer, {
       filename: fileName,
       contentType: 'application/octet-stream',
-    })
-    .expect(200);
+    });
+  }
+
+  const res = await resp.expect(200);
 
   return res.body.data.submitAudioTranscription;
 }