feat(server): improve transcript error handle (#11293)

packages/backend/server/src/plugins/copilot/context/job.ts

@@ -18,24 +18,6 @@ import { OpenAIEmbeddingClient } from './embedding';
 import { EmbeddingClient } from './types';
 import { readStream } from './utils';
 
-declare global {
-  interface Jobs {
-    'doc.embedPendingDocs': {
-      workspaceId: string;
-      docId: string;
-    };
-
-    'doc.embedPendingFiles': {
-      contextId: string;
-      userId: string;
-      workspaceId: string;
-      blobId: string;
-      fileId: string;
-      fileName: string;
-    };
-  }
-}
-
 @Injectable()
 export class CopilotContextDocJob {
   private supportEmbedding = false;

packages/backend/server/src/plugins/copilot/context/types.ts

@@ -21,6 +21,21 @@ declare global {
       error: string;
     };
   }
+  interface Jobs {
+    'doc.embedPendingDocs': {
+      workspaceId: string;
+      docId: string;
+    };
+
+    'doc.embedPendingFiles': {
+      contextId: string;
+      userId: string;
+      workspaceId: string;
+      blobId: string;
+      fileId: string;
+      fileName: string;
+    };
+  }
 }
 
 export const MAX_EMBEDDABLE_SIZE = 50 * OneMB;

packages/backend/server/src/plugins/copilot/prompt/prompts.ts

@@ -364,6 +364,7 @@ The output should be a JSON array, with each element containing:
 
 - Ensure the accurate differentiation of speakers even if multiple speakers overlap slightly or switch rapidly.
 - Maintain a consistent speaker labeling system throughout the transcription.
+- If the provided audio or data does not contain valid talk, you should return an empty JSON array.
 `,
       },
     ],

packages/backend/server/src/plugins/copilot/transcript/service.ts

@@ -4,9 +4,11 @@ import { AiJobStatus, AiJobType } from '@prisma/client';
 import {
   CopilotPromptNotFound,
   CopilotTranscriptionJobExists,
+  EventBus,
   type FileUpload,
   JobQueue,
   NoCopilotProviderAvailable,
+  OnEvent,
   OnJob,
 } from '../../../base';
 import { Models } from '../../../models';
@@ -34,6 +36,7 @@ export type TranscriptionJob = {
 @Injectable()
 export class CopilotTranscriptionService {
   constructor(
+    private readonly event: EventBus,
     private readonly models: Models,
     private readonly job: JobQueue,
     private readonly storage: CopilotStorage,
@@ -65,16 +68,11 @@ export class CopilotTranscriptionService {
       status: AiJobStatus.running,
     });
 
-    await this.job.add(
+    await this.job.add('copilot.transcript.submit', {
-      'copilot.transcript.submit',
+      jobId,
-      {
+      url,
-        jobId,
+      mimeType: blob.mimetype,
-        url,
+    });
-        mimeType: blob.mimetype,
-      },
-      // retry 3 times
-      { removeOnFail: 3 }
-    );
 
     return { id: jobId, status };
   }
@@ -114,9 +112,11 @@ export class CopilotTranscriptionService {
 
     const ret: TranscriptionJob = { id: job.id, status: job.status };
 
-    const payload = TranscriptPayloadSchema.safeParse(job.payload);
+    if (job.status === AiJobStatus.claimed) {
-    if (payload.success) {
+      const payload = TranscriptPayloadSchema.safeParse(job.payload);
-      ret.transcription = payload.data;
+      if (payload.success) {
+        ret.transcription = payload.data;
+      }
     }
 
     return ret;
@@ -164,81 +164,124 @@ export class CopilotTranscriptionService {
     url,
     mimeType,
   }: Jobs['copilot.transcript.submit']) {
-    const result = await this.chatWithPrompt('Transcript audio', {
+    try {
-      attachments: [url],
+      const result = await this.chatWithPrompt('Transcript audio', {
-      params: { mimetype: mimeType },
+        attachments: [url],
-    });
+        params: { mimetype: mimeType },
-
-    const transcription = TranscriptionSchema.parse(
-      JSON.parse(this.cleanupResponse(result))
-    );
-    await this.models.copilotJob.update(jobId, {
-      payload: { transcription },
-    });
-
-    await this.job.add(
-      'copilot.transcriptSummary.submit',
-      {
-        jobId,
-      },
-      // retry 3 times
-      { removeOnFail: 3 }
-    );
-  }
-
-  @OnJob('copilot.transcriptSummary.submit')
-  async transcriptSummary({ jobId }: Jobs['copilot.transcriptSummary.submit']) {
-    const payload = await this.models.copilotJob.getPayload(
-      jobId,
-      TranscriptPayloadSchema
-    );
-    if (payload.transcription) {
-      const content = payload.transcription
-        .map(t => t.transcription)
-        .join('\n');
-
-      const result = await this.chatWithPrompt('Summary', { content });
-
-      payload.summary = this.cleanupResponse(result);
-      await this.models.copilotJob.update(jobId, {
-        payload,
       });
 
-      await this.job.add(
+      const transcription = TranscriptionSchema.parse(
-        'copilot.transcriptTitle.submit',
+        JSON.parse(this.cleanupResponse(result))
-        { jobId },
-        // retry 3 times
-        { removeOnFail: 3 }
       );
-    } else {
       await this.models.copilotJob.update(jobId, {
-        status: AiJobStatus.failed,
+        payload: { transcription },
       });
+
+      await this.job.add('copilot.transcript.summary.submit', {
+        jobId,
+      });
+      return;
+    } catch (error: any) {
+      // record failed status and passthrough error
+      this.event.emit('workspace.file.transcript.failed', {
+        jobId,
+      });
+      throw error;
     }
   }
 
-  @OnJob('copilot.transcriptTitle.submit')
+  @OnJob('copilot.transcript.summary.submit')
-  async transcriptTitle({ jobId }: Jobs['copilot.transcriptTitle.submit']) {
+  async transcriptSummary({
-    const payload = await this.models.copilotJob.getPayload(
+    jobId,
-      jobId,
+  }: Jobs['copilot.transcript.summary.submit']) {
-      TranscriptPayloadSchema
+    try {
-    );
+      const payload = await this.models.copilotJob.getPayload(
-    if (payload.transcription && payload.summary) {
+        jobId,
-      const content = payload.transcription
+        TranscriptPayloadSchema
-        .map(t => t.transcription)
+      );
-        .join('\n');
+      if (payload.transcription) {
+        const content = payload.transcription
+          .map(t => t.transcription.trim())
+          .join('\n')
+          .trim();
 
-      const result = await this.chatWithPrompt('Summary as title', { content });
+        if (content.length) {
+          const result = await this.chatWithPrompt('Summary', {
+            content,
+          });
 
-      payload.title = this.cleanupResponse(result);
+          payload.summary = this.cleanupResponse(result);
-      await this.models.copilotJob.update(jobId, {
+          await this.models.copilotJob.update(jobId, {
-        payload,
+            payload,
-        status: AiJobStatus.finished,
+          });
-      });
+
-    } else {
+          await this.job.add('copilot.transcript.title.submit', {
-      await this.models.copilotJob.update(jobId, {
+            jobId,
-        status: AiJobStatus.failed,
+          });
+          return;
+        }
+      }
+    } catch (error: any) {
+      // record failed status and passthrough error
+      this.event.emit('workspace.file.transcript.failed', {
+        jobId,
       });
+      throw error;
     }
   }
+
+  @OnJob('copilot.transcript.title.submit')
+  async transcriptTitle({ jobId }: Jobs['copilot.transcript.title.submit']) {
+    try {
+      const payload = await this.models.copilotJob.getPayload(
+        jobId,
+        TranscriptPayloadSchema
+      );
+      if (payload.transcription && payload.summary) {
+        const content = payload.transcription
+          .map(t => t.transcription.trim())
+          .join('\n')
+          .trim();
+
+        if (content.length) {
+          const result = await this.chatWithPrompt('Summary as title', {
+            content,
+          });
+
+          payload.title = this.cleanupResponse(result);
+          await this.models.copilotJob.update(jobId, {
+            payload,
+          });
+          this.event.emit('workspace.file.transcript.finished', {
+            jobId,
+          });
+          return;
+        }
+      }
+    } catch (error: any) {
+      // record failed status and passthrough error
+      this.event.emit('workspace.file.transcript.failed', {
+        jobId,
+      });
+      throw error;
+    }
+  }
+
+  @OnEvent('workspace.file.transcript.finished')
+  async onFileTranscriptFinish({
+    jobId,
+  }: Events['workspace.file.transcript.finished']) {
+    await this.models.copilotJob.update(jobId, {
+      status: AiJobStatus.finished,
+    });
+  }
+
+  @OnEvent('workspace.file.transcript.failed')
+  async onFileTranscriptFailed({
+    jobId,
+  }: Events['workspace.file.transcript.failed']) {
+    await this.models.copilotJob.update(jobId, {
+      status: AiJobStatus.failed,
+    });
+  }
 }

packages/backend/server/src/plugins/copilot/transcript/types.ts

@@ -22,16 +22,24 @@ export type Transcription = z.infer<typeof TranscriptionSchema>;
 export type TranscriptionPayload = z.infer<typeof TranscriptPayloadSchema>;
 
 declare global {
+  interface Events {
+    'workspace.file.transcript.finished': {
+      jobId: string;
+    };
+    'workspace.file.transcript.failed': {
+      jobId: string;
+    };
+  }
   interface Jobs {
     'copilot.transcript.submit': {
       jobId: string;
       url: string;
       mimeType: string;
     };
-    'copilot.transcriptSummary.submit': {
+    'copilot.transcript.summary.submit': {
       jobId: string;
     };
-    'copilot.transcriptTitle.submit': {
+    'copilot.transcript.title.submit': {
       jobId: string;
     };
   }