Endpoint to transcribe videos #554

ekrojo77 · 2024-12-05T11:13:17Z

Testing
Endepunkt for å polle om filen er opprettet
Bedre return fra GET

ekrojo77 · 2024-12-12T13:23:07Z

audio-api/src/main/scala/no/ndla/audioapi/controller/TranscriptionController.scala

+            case Success(_)  => Right(())
+            case Failure(ex) => returnLeftError(ex)
+          }
+        }


Beholder disse i første omgang før vi vet hva som trengs av endepunkt. Men tror det rene transcribe-endepunktet er det som vil bli mest brukt

gunnarvelle

Lisenspirk

audio-api/src/main/scala/no/ndla/audioapi/controller/TranscriptionController.scala

audio-api/src/main/scala/no/ndla/audioapi/model/api/TranscriptionResultDTO.scala

audio-api/src/main/scala/no/ndla/audioapi/service/TranscriptionService.scala

audio-api/src/test/scala/no/ndla/audioapi/service/TranscriptionServiceTest.scala

common/src/main/scala/no/ndla/common/brightcove/NdlaBrightcoveClient.scala

audio-api/src/main/scala/no/ndla/audioapi/controller/TranscriptionController.scala

common/src/main/scala/no/ndla/common/aws/NdlaS3Client.scala

common/src/main/scala/no/ndla/common/brightcove/NdlaBrightcoveClient.scala

audio-api/src/main/scala/no/ndla/audioapi/AudioApiProperties.scala

audio-api/src/main/scala/no/ndla/audioapi/controller/TranscriptionController.scala

audio-api/src/main/scala/no/ndla/audioapi/service/TranscriptionService.scala

common/src/main/scala/no/ndla/common/brightcove/NdlaBrightcoveClient.scala

jnatten · 2025-01-08T12:51:34Z

audio-api/src/main/scala/no/ndla/audioapi/service/TranscriptionService.scala

+    def getVideoTranscription(
+        videoId: String,
+        language: String
+    ): Try[Either[String, String]] = {


Jeg synes fortsatt det er vanskelig å vite hva denne typen representerer uten å dykke ned i hva funksjonen gjør og returnerer.

Kanskje vi kunne brukt en trait som enten kunne være success eller ikke?
Noe sånt?

sealed trait TranscriptionResult case class TranscriptionComplete(transcription: String) extends TranscriptionResult case class TranscriptionNonComplete(transcriptionStatus: String) extends TranscriptionResult

Da kan vi gjøre feks sånn:

getObjectFromS3(transcribeUri).map(TranscriptionComplete) } else { Success(TranscriptionNonComplete(transcriptionJobStatus))

jnatten · 2025-01-08T12:55:11Z

audio-api/src/main/scala/no/ndla/audioapi/service/TranscriptionService.scala

+
+      transcribeClient.getTranscriptionJob(jobName).flatMap { transcriptionJobResponse =>
+        val transcriptionJob       = transcriptionJobResponse.transcriptionJob()
+        val transcriptionJobStatus = transcriptionJob.transcriptionJobStatus().toString


Denne trenger vel egentlig ikke konverteres til en streng?

jnatten · 2025-01-08T12:58:58Z

common/src/main/scala/no/ndla/common/brightcove/NdlaBrightcoveClient.scala

+      val authResponse = request.send(backend)
+      Try {
+        authResponse.body match {
+          case Right(jsonString) =>
+            decode[TokenResponse](jsonString) match {
+              case Right(tokenResponse) => tokenResponse.access_token
+              case Left(error)          => throw new Exception(s"Failed to decode token response: ${error.getMessage}")
+            }
+          case Left(error) => throw new Exception(s"Failed to get token: ${error}")
+        }


request.send(backend) kan potensielt kaste exception så den burde nok være med i Try'en, kanskje noe sånt?

Suggested change

val authResponse = request.send(backend)

Try {

authResponse.body match {

case Right(jsonString) =>

decode[TokenResponse](jsonString) match {

case Right(tokenResponse) => tokenResponse.access_token

case Left(error) => throw new Exception(s"Failed to decode token response: ${error.getMessage}")

}

case Left(error) => throw new Exception(s"Failed to get token: ${error}")

}

Try(request.send(backend).body) match {

case Success(Right(jsonString)) =>

decode[TokenResponse](jsonString) match {

case Right(tokenResponse) => Success(tokenResponse.access_token)

case Left(error) => Failure(new Exception(s"Failed to decode token response: ${error.getMessage}"))

}

case Success(Left(error)) => Failure(new Exception(s"Failed to get token: ${error}"))

case Failure(ex) => Failure(ex)

}

Ideelt sett så burde vi sikkert lage noen spesifikke Exception typer og (eller gjenbruke om vi har noen som finnes allerede)

jnatten · 2025-01-08T13:02:38Z

common/src/main/scala/no/ndla/common/brightcove/NdlaBrightcoveClient.scala

+      val response = request.send(backend)
+      Try {
+        response.body match {
+          case Right(jsonString) =>
+            parse(jsonString) match {
+              case Right(json) =>
+                json.asArray match {
+                  case Some(videoSources) => videoSources
+                  case None               => throw new Exception("Failed to parse video source")
+                }
+              case Left(error) => throw new Exception(s"Failed to parse video source: ${error.getMessage}")
+            }
+          case Left(error) => throw new Exception(s"Failed to get video source: ${error}")
+        }
+      }


Samme her:

Suggested change

val response = request.send(backend)

Try {

response.body match {

case Right(jsonString) =>

parse(jsonString) match {

case Right(json) =>

json.asArray match {

case Some(videoSources) => videoSources

case None => throw new Exception("Failed to parse video source")

}

case Left(error) => throw new Exception(s"Failed to parse video source: ${error.getMessage}")

}

case Left(error) => throw new Exception(s"Failed to get video source: ${error}")

}

}

Try(request.send(backend).body) match {

case Success(Right(jsonString)) =>

parse(jsonString) match {

case Right(json) =>

json.asArray match {

case Some(videoSources) => Success(videoSources)

case None => Failure(new Exception("Failed to parse video source"))

}

case Left(error) => Failure(new Exception(s"Failed to parse video source: ${error.getMessage}"))

}

case Success(Left(error)) => Failure(new Exception(s"Failed to get video source: ${error}"))

case Failure(ex) => Failure(ex)

}

ekrojo77 added 9 commits December 5, 2024 12:11

Endpoint to extract audio from brightcove files

94d8ea4

Add get endpoint

f46fd63

Update NotFound Error handling in audio-api

8730288

Add transcription client

37edf26

Add Transcription endpoints

87f3560

Add maxSpeakers and subtitles

68690a5

Fix test coverage for transcription service

1178838

remove unused import

461940c

Merge branch 'master' into transcribe-audio

e7dedf9

ekrojo77 commented Dec 12, 2024

View reviewed changes

ekrojo77 added 3 commits December 13, 2024 08:14

Add 400 if job already exist

2337237

Clean up folder names and logs

5cdb331

Update descriptions of transcription endpoints

d6e64b2

ekrojo77 marked this pull request as ready for review December 13, 2024 10:37

ekrojo77 requested a review from a team December 13, 2024 10:37

ekrojo77 marked this pull request as draft December 13, 2024 13:53

ekrojo77 changed the title ~~Endpoint to extract audio from brightcove files~~ Endpoint to transcribe videos Dec 13, 2024

ekrojo77 added 5 commits December 16, 2024 09:41

Update function to send subtitles as response if Get returns 200

47834f4

Remove unused imports

493e7f1

Add json return for transcription results

0342b7f

Fix syntax error

7d7e95e

Add mock for getting transcription results

e0109cd

ekrojo77 marked this pull request as ready for review December 16, 2024 10:43

ekrojo77 added 2 commits December 17, 2024 10:23

add a temp audio endpoint

850dbd9

fix audioprop

9fb082a

ekrojo77 marked this pull request as draft December 17, 2024 13:19

fix audio endpoints

60efcd7

gunnarvelle reviewed Dec 18, 2024

View reviewed changes

ekrojo77 added 2 commits December 18, 2024 13:07

Fix licensing and unused audioName in get call

8a24b8d

Make creation of subtitle file optional

e54886d

fix test

31b918e

gunnarvelle reviewed Dec 20, 2024

View reviewed changes

ekrojo77 added 3 commits December 20, 2024 13:01

Update permissions and descriptions, refactor to existing s3 putobject

5163c80

Put urls into props

5759c2d

Fix naming and test

46d4e79

ekrojo77 marked this pull request as ready for review January 2, 2025 10:38

gunnarvelle reviewed Jan 3, 2025

View reviewed changes

audio-api/src/main/scala/no/ndla/audioapi/AudioApiProperties.scala Outdated Show resolved Hide resolved

audio-api/src/main/scala/no/ndla/audioapi/AudioApiProperties.scala Outdated Show resolved Hide resolved

jnatten reviewed Jan 6, 2025

View reviewed changes

ekrojo77 added 6 commits January 6, 2025 08:30

Fix naming of props

a2a17f5

Fix error handling in api

0bd5339

Improve error handling

628be06

Fix unused import and try handling of brightcove client

6c355db

Remove unused import

ea32cc9

update getVideo after brightcove changes

574bc37

ekrojo77 requested a review from jnatten January 8, 2025 06:40

jnatten reviewed Jan 8, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Endpoint to transcribe videos #554

Endpoint to transcribe videos #554

ekrojo77 commented Dec 5, 2024 •

edited

Loading

ekrojo77 Dec 12, 2024

gunnarvelle left a comment

jnatten Jan 8, 2025

jnatten Jan 8, 2025

jnatten Jan 8, 2025

jnatten Jan 8, 2025

jnatten Jan 8, 2025

jnatten Jan 8, 2025

Endpoint to transcribe videos #554

Are you sure you want to change the base?

Endpoint to transcribe videos #554

Conversation

ekrojo77 commented Dec 5, 2024 • edited Loading

ekrojo77 Dec 12, 2024

Choose a reason for hiding this comment

gunnarvelle left a comment

Choose a reason for hiding this comment

jnatten Jan 8, 2025

Choose a reason for hiding this comment

jnatten Jan 8, 2025

Choose a reason for hiding this comment

jnatten Jan 8, 2025

Choose a reason for hiding this comment

jnatten Jan 8, 2025

Choose a reason for hiding this comment

jnatten Jan 8, 2025

Choose a reason for hiding this comment

jnatten Jan 8, 2025

Choose a reason for hiding this comment

ekrojo77 commented Dec 5, 2024 •

edited

Loading