manual fix

Author: ks6088ts

.env.template

@@ -15,6 +15,5 @@ AZURE_BLOB_STORAGE_CONNECTION_STRING="DefaultEndpointsProtocol=https;AccountName
 AZURE_BLOB_STORAGE_CONTAINER_NAME="files"
 
 # Azure AI Speech
-AZURE_SPEECH_KEY="<YOUR_SPEECH_KEY>"
-AZURE_SPEECH_REGION="<YOUR_SPEECH_REGION>"
-AZURE_SPEECH_ENDPOINT="https://<YOUR_SPEECH_REGION>.api.cognitive.microsoft.com/"
+AZURE_AI_SPEECH_API_KEY="<YOUR_AZURE_AI_SPEECH_API_KEY>"
+AZURE_AI_SPEECH_ENDPOINT="https://<speech-api-name>.cognitiveservices.azure.com/"

docs/index.md

@@ -76,32 +76,39 @@ uv run python scripts/files.py delete-multiple-files "file1.txt" "file2.jpg" "fi
 ### Speeches Service
 
 ```shell
+AZURE_BLOB_STORAGE_CONTAINER_SAS_TOKEN="<your_sas_token>"
+AZURE_BLOB_STORAGE_CONTAINER_URL="https://<storage_account_name>.blob.core.windows.net/<container_name>"
+FILE_NAME="path/to/your/audio/file.wav"
+URL="${AZURE_BLOB_STORAGE_CONTAINER_URL}/${FILE_NAME}?${AZURE_BLOB_STORAGE_CONTAINER_SAS_TOKEN}"
+
 # Help
 uv run python scripts/speeches.py --help
 
 # Create a new transcription job
-uv run python scripts/speeches.py create-transcription "https://example.com/audio.wav" --locale "ja-JP" --name "My Transcription"
+uv run python scripts/speeches.py create-transcription "$URL" \
+  --locale "ja-JP" \
+  --name "My Transcription"
 
 # Get transcription job status
-uv run python scripts/speeches.py get-transcription JOB_ID
+uv run python scripts/speeches.py get-transcription "$JOB_ID"
+
+# Wait for transcription completion
+uv run python scripts/speeches.py wait-for-completion "$JOB_ID" --timeout 300 --interval 10
 
 # Get transcription files
-uv run python scripts/speeches.py get-transcription-files JOB_ID
+uv run python scripts/speeches.py get-transcription-files "$JOB_ID"
 
 # Get transcription result
-uv run python scripts/speeches.py get-transcription-result "https://example.com/result.json" --save "result.json"
+uv run python scripts/speeches.py get-transcription-result "https://<contentUrl>" --save "result.json"
 
 # List all transcription jobs
 uv run python scripts/speeches.py list-transcriptions
 
-# Wait for transcription completion
-uv run python scripts/speeches.py wait-for-completion JOB_ID --timeout 300 --interval 10
-
 # Delete transcription job
-uv run python scripts/speeches.py delete-transcription JOB_ID
+uv run python scripts/speeches.py delete-transcription "$JOB_ID"
 
 # Delete transcription job (without confirmation)
-uv run python scripts/speeches.py delete-transcription JOB_ID --force
+uv run python scripts/speeches.py delete-transcription "$JOB_ID" --force
 ```
 
 ## MCP
@@ -157,3 +164,7 @@ az resource update \
 - [FastAPI のテレメトリデータを Azure Application Insights に送る](https://qiita.com/hoto17296/items/2f366dfabdbe3d1d4e97)
 - [【Azure Functions】 - Application Insights のログが表示されない問題](https://zenn.dev/headwaters/articles/ff19f7e1b99b44)
 - [opentelemetry-instrumentation-fastapi (python) から OpenTelemetry に入門する](https://zenn.dev/taxin/articles/opentelemetry-fast-api-instrumentation-basics)
+
+### Azure AI Speech
+
+- [バッチ文字起こしとは](https://learn.microsoft.com/ja-jp/azure/ai-services/speech-service/batch-transcription)

scripts/speeches.py

@@ -22,7 +22,6 @@ def create_transcription(
     content_urls: list[str] = typer.Argument(..., help="転写するファイルのURL（複数指定可能）"),
     locale: str = typer.Option("ja-JP", "--locale", "-l", help="言語設定"),
     display_name: str = typer.Option(None, "--name", "-n", help="転写ジョブの表示名"),
-    model: str = typer.Option(None, "--model", "-m", help="使用するモデル"),
 ):
     """新しい転写ジョブを作成する"""
     console.print("[bold green]転写ジョブを作成します[/bold green]")
@@ -34,7 +33,6 @@ def create_transcription(
             content_urls=content_urls,
             locale=locale,
             display_name=display_name or "CLI Batch Transcription",
-            model=model,
         )
 
         response = speech_repo.create_transcription_job(request)
@@ -95,6 +93,8 @@ def get_transcription_files(
         table.add_column("名前", style="cyan")
         table.add_column("種類", style="green")
         table.add_column("リンク", style="yellow")
+        for file in files:
+            print(file)
 
         for file in files:
             table.add_row(

template_fastapi/models/speech.py

@@ -58,11 +58,12 @@ class BatchTranscriptionRequest(BaseModel):
 
     model_config = ConfigDict(extra="ignore")
 
-    content_urls: list[str]
+    content_urls: list[str] = [
+        "https://<storage_account_name>.blob.core.windows.net/<container_name>/<file1.m4a>?<sas_token>",
+        "https://<storage_account_name>.blob.core.windows.net/<container_name>/<file2.m4a>?<sas_token>",
+    ]
     locale: str = "ja-JP"
-    display_name: str | None = None
-    model: str | None = None
-    properties: dict[str, Any] | None = None
+    display_name: str | None = "My Batch Transcription"
 
 
 class BatchTranscriptionResponse(BaseModel):

template_fastapi/repositories/speeches.py

@@ -23,11 +23,10 @@ class SpeechRepository:
     """音声認識データを管理するリポジトリクラス"""
 
     def __init__(self):
-        self.speech_key = azure_speech_settings.azure_speech_key
-        self.speech_region = azure_speech_settings.azure_speech_region
-        self.speech_endpoint = azure_speech_settings.azure_speech_endpoint
+        self.speech_key = azure_speech_settings.azure_ai_speech_api_key
+        self.speech_endpoint = azure_speech_settings.azure_ai_speech_endpoint
         self.api_version = "v3.2-preview.2"
-        self.base_url = f"{self.speech_endpoint}speechtotext/{self.api_version}"
+        self.base_url = urljoin(self.speech_endpoint, f"speechtotext/{self.api_version}/")
 
         # セッションの設定
         self.session = requests.Session()
@@ -55,14 +54,10 @@ def create_transcription_job(self, request: BatchTranscriptionRequest) -> BatchT
             "contentUrls": request.content_urls,
             "locale": request.locale,
             "displayName": request.display_name or "Batch Transcription",
-            "model": request.model,
-            "properties": request.properties or {},
+            "model": None,  # Whisperモデルを使用する場合はNone
+            "properties": {},
         }
 
-        # Whisperモデルのデフォルト設定
-        if not request.model:
-            payload["model"] = None  # Whisperモデルを使用する場合
-
         try:
             response = self.session.post(url, headers=self.headers, data=json.dumps(payload), timeout=30)
             response.raise_for_status()
@@ -150,7 +145,6 @@ def delete_transcription_job(self, job_id: str) -> bool:
     def list_transcription_jobs(self) -> list[TranscriptionJob]:
         """転写ジョブの一覧を取得する"""
         url = urljoin(self.base_url, "transcriptions")
-
         try:
             response = self.session.get(url, headers=self.headers, timeout=30)
             response.raise_for_status()

template_fastapi/settings/azure_speech.py

@@ -4,9 +4,8 @@
 
 
 class Settings(BaseSettings):
-    azure_speech_key: str = "<YOUR_SPEECH_KEY>"
-    azure_speech_region: str = "<YOUR_SPEECH_REGION>"
-    azure_speech_endpoint: str = "https://<YOUR_SPEECH_REGION>.api.cognitive.microsoft.com/"
+    azure_ai_speech_api_key: str = "<YOUR_AZURE_AI_SPEECH_API_KEY>"
+    azure_ai_speech_endpoint: str = "https://<speech-api-name>.cognitiveservices.azure.com/"
 
     model_config = SettingsConfigDict(
         env_file=".env",