fix(api): preserve plaintext and rich audio response types

.castiron.stats.yml

@@ -1,8 +1,8 @@
 schema_version: 1
-generation_id: cf94c82f-0604-4dfc-aff0-cec482e84953
-openapi_spec_hash: ce912f79d7f8cb19a7a515513c214b8a
-openapi_transformed_spec_hash: 352b5a7a028f990b6475b6593e381eee
-config_hash: dde98d9c41cd6f4163320f92c70f83e6
-codegen_sha: 326831af0c0fdf859335a189df2b6716e4b605d4
-codegen_hash: 00e7f6a53e1a6a1adf02385bc368fce999fdd775d1f1c28615f790c7a4109ad0
-public_codegen_sha: c2abe7c282c83c8fb9a7330254235c20b67a2b10
+generation_id: 8e5ebfdb-d6fd-41d4-9173-a575e6ba13fc
+openapi_spec_hash: 8c73dea75f8570b7ea967542dafa0b0f
+openapi_transformed_spec_hash: 19d647b4e088bfa2499083e65eacd54e
+config_hash: ad0fbf2fbcf7981fe7a9910664285991
+codegen_sha: 9244e44f047f4d2a30e58bf5423011673623ce81
+codegen_hash: 391329cd5b43dc65c96676993a33fa22103908dd8b8e2f3cf461a837b31df84b
+public_codegen_sha: 1252578bddc8b8637004578d7a16f3c8ee7ee22e

api_reference/openapi.transformed.yml

@@ -601,11 +601,64 @@ paths:
               schema:
                 type: string
                 format: binary
+            audio/mpeg:
+              schema:
+                type: string
+                format: binary
+            audio/aac:
+              schema:
+                type: string
+                format: binary
+            audio/opus:
+              schema:
+                type: string
+                format: binary
+            audio/flac:
+              schema:
+                type: string
+                format: binary
+            audio/pcm:
+              schema:
+                type: string
+                format: binary
+            audio/wav:
+              schema:
+                type: string
+                format: binary
             text/event-stream:
               schema:
                 $ref: '#/components/schemas/CreateSpeechResponseStreamEvent'
+        '400':
+          description: Invalid speech request, input, output format, or voice.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
+        '401':
+          description: Authentication failed because the API key is missing or revoked, or the client IP is not authorized.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
+            text/plain:
+              schema:
+                type: string
+                description: JSON-encoded error text containing an error object and the HTTP status.
+        '403':
+          description: Access to a personal API organization is blocked by the organization policy.
+          content:
+            text/plain:
+              schema:
+                type: string
+                description: JSON-encoded error text containing an error object and the HTTP status.
         '429':
           $ref: '#/components/responses/InferenceRateLimited'
+        '500':
+          description: The voice or speech audio could not be processed.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
         '503':
           $ref: '#/components/responses/InferenceServiceUnavailable'
       x-oaiMeta:
@@ -694,7 +747,8 @@ paths:
         Transcribes audio into the input language.
 
         Returns a transcription object in `json`, `diarized_json`, or `verbose_json`
-        format, or a stream of transcript events.
+        format, plain text in `text`, `srt`, or `vtt` format, or a stream of transcript
+        events. Supported formats depend on the model.
       requestBody:
         required: true
         content:
@@ -710,16 +764,58 @@ paths:
                 anyOf:
                 - $ref: '#/components/schemas/CreateTranscriptionResponseJson'
                 - $ref: '#/components/schemas/CreateTranscriptionResponseDiarizedJson'
-                  x-stainless-skip:
-                  - go
                 - $ref: '#/components/schemas/CreateTranscriptionResponseVerboseJson'
                 discriminator:
                   propertyName: task
+            text/plain:
+              schema:
+                type: string
             text/event-stream:
               schema:
                 $ref: '#/components/schemas/CreateTranscriptionResponseStreamEvent'
+        '400':
+          description: Invalid audio input or request parameters.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
+        '401':
+          description: Authentication failed because the API key is missing or revoked, or the client IP is not authorized.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
+            text/plain:
+              schema:
+                type: string
+                description: JSON-encoded error text containing an error object and the HTTP status.
+        '403':
+          description: Access to a personal API organization is blocked by the organization policy.
+          content:
+            text/plain:
+              schema:
+                type: string
+                description: JSON-encoded error text containing an error object and the HTTP status.
+        '413':
+          description: The audio request exceeds the supported size limit.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
         '429':
           $ref: '#/components/responses/InferenceRateLimited'
+        '500':
+          description: The audio could not be processed.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
+        '502':
+          description: The upstream audio service connection failed.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
         '503':
           $ref: '#/components/responses/InferenceServiceUnavailable'
       x-oaiMeta:
@@ -1271,8 +1367,52 @@ paths:
                 - $ref: '#/components/schemas/CreateTranslationResponseVerboseJson'
                   x-stainless-skip:
                   - go
+            text/plain:
+              schema:
+                type: string
+        '400':
+          description: Invalid audio input or request parameters.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
+        '401':
+          description: Authentication failed because the API key is missing or revoked, or the client IP is not authorized.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
+            text/plain:
+              schema:
+                type: string
+                description: JSON-encoded error text containing an error object and the HTTP status.
+        '403':
+          description: Access to a personal API organization is blocked by the organization policy.
+          content:
+            text/plain:
+              schema:
+                type: string
+                description: JSON-encoded error text containing an error object and the HTTP status.
+        '413':
+          description: The audio request exceeds the supported size limit.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
         '429':
           $ref: '#/components/responses/InferenceRateLimited'
+        '500':
+          description: The audio could not be processed.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
+        '502':
+          description: The upstream audio service connection failed.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/ErrorResponse'
         '503':
           $ref: '#/components/responses/InferenceServiceUnavailable'
       x-oaiMeta:
@@ -66730,6 +66870,36 @@ components:
             "type": "safety.warning_issued",
             "data": {"id": "C-abc123"}
           }
+    _MisalignmentErrorType:
+      anyOf:
+      - type: string
+      - type: string
+        enum:
+        - potentially_unintended_data_transfer
+        - potentially_unintended_data_access
+        - potentially_unintended_destructive_activity
+        - other
+    _MisalignmentSteer:
+      properties:
+        message:
+          type: string
+          description: The public continuation instruction.
+      type: object
+      required:
+      - message
+    MisalignmentErrorDetailsResource:
+      properties:
+        error_type:
+          $ref: '#/components/schemas/_MisalignmentErrorType'
+          description: An optional classification; clients must accept additional values.
+        detailed_explanation:
+          type: string
+          description: The public explanation for this block.
+        steer:
+          $ref: '#/components/schemas/_MisalignmentSteer'
+          description: An optional public continuation instruction.
+      type: object
+      required: []
     ModerationInputType:
       type: string
       enum:
@@ -70206,36 +70376,6 @@ components:
       - object
       - metadata
       - created_at
-    _MisalignmentErrorType:
-      anyOf:
-      - type: string
-      - type: string
-        enum:
-        - potentially_unintended_data_transfer
-        - potentially_unintended_data_access
-        - potentially_unintended_destructive_activity
-        - other
-    _MisalignmentSteer:
-      properties:
-        message:
-          type: string
-          description: The public continuation instruction.
-      type: object
-      required:
-      - message
-    MisalignmentErrorDetailsResource:
-      properties:
-        error_type:
-          $ref: '#/components/schemas/_MisalignmentErrorType'
-          description: An optional classification; clients must accept additional values.
-        detailed_explanation:
-          type: string
-          description: The public explanation for this block.
-        steer:
-          $ref: '#/components/schemas/_MisalignmentSteer'
-          description: An optional public continuation instruction.
-      type: object
-      required: []
     ImageGenOutputTokensDetails:
       properties:
         image_tokens:

examples/audio/record_and_play.rb

@@ -14,6 +14,6 @@
 client = OpenAI::Client.new
 clip = OpenAI::LocalAudio.record(duration: 5, timeout: 10)
 transcription = client.audio.transcriptions.create(model: TRANSCRIPTION_MODEL, file: clip)
-puts(transcription.text)
+puts(transcription.is_a?(StringIO) ? transcription.read : transcription.text)
 speech = client.audio.speech.create(model: SPEECH_MODEL, voice: VOICE, input: "Recording complete.")
 OpenAI::LocalAudio.play(speech, timeout: 30)

lib/openai/resources/admin/organization.rb

@@ -4,9 +4,6 @@ module OpenAI
   module Resources
     class Admin
       class Organization
-        # @return [OpenAI::Resources::Admin::Organization::ExternalStorage]
-        attr_reader :external_storage
-
         # List user actions and configuration changes within this organization.
         # @return [OpenAI::Resources::Admin::Organization::AuditLogs]
         attr_reader :audit_logs
@@ -32,6 +29,9 @@ class Organization
         # @return [OpenAI::Resources::Admin::Organization::DataRetention]
         attr_reader :data_retention
 
+        # @return [OpenAI::Resources::Admin::Organization::ExternalStorage]
+        attr_reader :external_storage
+
         # @return [OpenAI::Resources::Admin::Organization::SpendLimit]
         attr_reader :spend_limit
 
@@ -49,7 +49,6 @@ class Organization
         # @param client [OpenAI::Client]
         def initialize(client:)
           @client = client
-          @external_storage = OpenAI::Resources::Admin::Organization::ExternalStorage.new(client: client)
           @audit_logs = OpenAI::Resources::Admin::Organization::AuditLogs.new(client: client)
           @admin_api_keys = OpenAI::Resources::Admin::Organization::AdminAPIKeys.new(client: client)
           @usage = OpenAI::Resources::Admin::Organization::Usage.new(client: client)
@@ -58,6 +57,7 @@ def initialize(client:)
           @groups = OpenAI::Resources::Admin::Organization::Groups.new(client: client)
           @roles = OpenAI::Resources::Admin::Organization::Roles.new(client: client)
           @data_retention = OpenAI::Resources::Admin::Organization::DataRetention.new(client: client)
+          @external_storage = OpenAI::Resources::Admin::Organization::ExternalStorage.new(client: client)
           @spend_limit = OpenAI::Resources::Admin::Organization::SpendLimit.new(client: client)
           @spend_alerts = OpenAI::Resources::Admin::Organization::SpendAlerts.new(client: client)
           @certificates = OpenAI::Resources::Admin::Organization::Certificates.new(client: client)

lib/openai/resources/audio/transcriptions.rb

@@ -11,7 +11,8 @@ class Transcriptions
         # Transcribes audio into the input language.
         #
         # Returns a transcription object in `json`, `diarized_json`, or `verbose_json`
-        # format, or a stream of transcript events.
+        # format, plain text in `text`, `srt`, or `vtt` format, or a stream of transcript
+        # events. Supported formats depend on the model.
         #
         # `String`, `StringIO`, and pathless `IO` inputs are sent with generic upload
         # metadata. Use `OpenAI::FilePart` when you need to override the filename or
@@ -108,7 +109,7 @@ class Transcriptions
         #
         # @param request_options [OpenAI::RequestOptions, Hash{Symbol=>Object}, nil]
         #
-        # @return [OpenAI::Models::Audio::Transcription, OpenAI::Models::Audio::TranscriptionDiarized, OpenAI::Models::Audio::TranscriptionVerbose]
+        # @return [OpenAI::Models::Audio::Transcription, OpenAI::Models::Audio::TranscriptionDiarized, OpenAI::Models::Audio::TranscriptionVerbose, StringIO]
         #
         # @see OpenAI::Models::Audio::TranscriptionCreateParams
         def create(params)
@@ -118,12 +119,22 @@ def create(params)
             raise ArgumentError.new(message)
           end
 
+          body = OpenAI::Internal::Transport::RequestBodyMerge.merge(parsed, options[:extra_body])
+          model = case body.fetch(:response_format) { body["response_format"] }
+          in :verbose_json | "verbose_json"
+            OpenAI::UnionOf[OpenAI::Audio::TranscriptionVerbose, OpenAI::Audio::TranscriptionCreateResponse]
+          in :diarized_json | "diarized_json"
+            OpenAI::UnionOf[OpenAI::Audio::TranscriptionDiarized, OpenAI::Audio::TranscriptionCreateResponse]
+          else
+            OpenAI::Audio::TranscriptionCreateResponse
+          end
+
           @client.request(
             method: :post,
             path: "audio/transcriptions",
             headers: {"content-type" => "multipart/form-data"},
             body: parsed,
-            model: OpenAI::Models::Audio::TranscriptionCreateResponse,
+            model: model,
             security: {bearer_auth: true},
             options: options
           )
@@ -135,7 +146,8 @@ def create(params)
         # Transcribes audio into the input language.
         #
         # Returns a transcription object in `json`, `diarized_json`, or `verbose_json`
-        # format, or a stream of transcript events.
+        # format, plain text in `text`, `srt`, or `vtt` format, or a stream of transcript
+        # events. Supported formats depend on the model.
         #
         # `String`, `StringIO`, and pathless `IO` inputs are sent with generic upload
         # metadata. Use `OpenAI::FilePart` when you need to override the filename or