gcs

app.py

@@ -67,20 +67,62 @@ def init_gcs_client(service_account_key_string):
     gcs_client = storage.Client(credentials=credentials, project=credentials_dict['project_id'])
     return gcs_client
 
-def upload_file_to_gcs(gcs_client, bucket_name, source_file_name, destination_blob_name, service_account_key_file):
-    """上传文件到 GCS"""
+def gcs_create_bucket_folder_if_not_exists(gcs_client, bucket_name, folder_name):
+    """检查是否存在特定名称的文件夹（前缀），如果不存在则创建一个标记文件来模拟文件夹"""
+    bucket = gcs_client.bucket(bucket_name)
+    blob = bucket.blob(folder_name)
+    if not blob.exists():
+        blob.upload_from_string('', content_type='application/x-www-form-urlencoded;charset=UTF-8')
+        print(f"GCS Folder '{folder_name}' created.")
+    else:
+        print(f"GCS Folder '{folder_name}' already exists.")
+
+def gcs_check_folder_exists(gcs_client, bucket_name, folder_name):
+    """检查 GCS 存储桶中是否存在指定的文件夹"""
+    bucket = gcs_client.bucket(bucket_name)
+    blobs = list(bucket.list_blobs(prefix=folder_name))
+    return len(blobs) > 0
+
+def gcs_check_file_exists(gcs_client, bucket_name, file_name):
+    """
+    检查 GCS 存储桶中是否存在指定的文件
+    file_name 格式：{folder_name}/{file_name}
+    """
+    bucket = gcs_client.bucket(bucket_name)
+    blob = bucket.blob(file_name)
+    return blob.exists()
+
+def upload_file_to_gcs(gcs_client, bucket_name, destination_blob_name, source_file_name):
+    """上传文件到指定的 GCS 存储桶"""
     bucket = gcs_client.bucket(bucket_name)
     blob = bucket.blob(destination_blob_name)
     blob.upload_from_filename(source_file_name)
-    print(f"{source_file_name} uploaded to {destination_blob_name}.")
+    print(f"File {source_file_name} uploaded to {destination_blob_name}.")
 
-def download_file_from_gcs(gcs_client, bucket_name, source_blob_name, destination_file_name, service_account_key_file):
-    """从 GCS 下载文件"""
+def download_blob_to_string(gcs_client, bucket_name, source_blob_name):
+    """从 GCS 下载文件内容到字符串"""
     bucket = gcs_client.bucket(bucket_name)
     blob = bucket.blob(source_blob_name)
-    blob.download_to_filename(destination_file_name)
-    print(f"{source_blob_name} downloaded to {destination_file_name}.")
+    return blob.download_as_text()
 
+def make_blob_public(gcs_client, bucket_name, blob_name):
+    """将指定的 GCS 对象设置为公共可读"""
+    bucket = gcs_client.bucket(bucket_name)
+    blob = bucket.blob(blob_name)
+    blob.make_public()
+    print(f"Blob {blob_name} is now publicly accessible at {blob.public_url}")
+
+def copy_all_files_from_drive_to_gcs(drive_service, gcs_client, drive_folder_id, bucket_name, gcs_folder_name):
+    # Get all files from the folder
+    query = f"'{drive_folder_id}' in parents and trashed = false"
+    response = drive_service.files().list(q=query).execute()
+    files = response.get('files', [])
+    for file in files:
+        # Copy each file to GCS
+        file_id = file['id']
+        file_name = file['name']
+        gcs_destination_path = f"{gcs_folder_name}/{file_name}"
+        copy_file_from_drive_to_gcs(drive_service, gcs_client, file_id, bucket_name, gcs_destination_path)
 
 # # ====drive====初始化Google Drive服务
 def init_drive_service():
@@ -278,6 +320,8 @@ def get_transcript(video_id):
 
 def process_transcript_and_screenshots(video_id):
     print("====process_transcript_and_screenshots====")
+
+    # Drive
     service = init_drive_service()
     parent_folder_id = '1GgI4YVs0KckwStVQkLa1NZ8IpaEMurkL'
     folder_id = create_folder_if_not_exists(service, video_id, parent_folder_id)
@@ -316,6 +360,16 @@ def process_transcript_and_screenshots(video_id):
     update_file_on_drive(service, file_id, updated_transcript_text)
     print("逐字稿已更新，包括截图链接")
 
+    # init gcs client
+    gcs_client = init_gcs_client(GCS_KEY)
+    bucket_name = 'video_ai_assistant'
+    # 检查 folder 是否存在
+    is_gcs_exists = gcs_check_folder_exists(gcs_client, bucket_name, video_id)
+    if not is_gcs_exists:
+        gcs_create_bucket_folder_if_not_exists(gcs_client, bucket_name, video_id)
+        copy_all_files_from_drive_to_gcs(service, gcs_client, folder_id, bucket_name, video_id)
+        print("Drive file 已上传到GCS")
+
     return transcript
 
 def process_youtube_link(link):