botrun-flow-lang 5.12.263__py3-none-any.whl → 6.2.21__py3-none-any.whl

botrun_flow_lang/langgraph_agents/agents/util/local_files.py

@@ -1,419 +1,419 @@
-import requests
-import mimetypes
-import time
-import random
-import os
-from tempfile import NamedTemporaryFile
-from pathlib import Path
-from io import BytesIO
-from .img_util import get_img_content_type
-from botrun_flow_lang.services.storage.storage_factory import storage_store_factory
-
-
-def get_file_content_type(file_path: str) -> str:
-    """根據檔案類型取得對應的 MIME type
-    如果是圖片檔案，會使用 get_img_content_type 來檢測實際的圖片格式
-    如果是其他檔案，則使用副檔名來判斷
-
-    Args:
-        file_path: 檔案路徑
-
-    Returns:
-        str: MIME type，如果無法判斷則返回 application/octet-stream
-    """
-    # 先用副檔名判斷是否為圖片
-    content_type, _ = mimetypes.guess_type(file_path)
-
-    # 如果副檔名顯示是圖片，使用 get_img_content_type 進行實際檢測
-    if content_type and content_type.startswith("image/"):
-        try:
-            return get_img_content_type(file_path)
-        except (ValueError, FileNotFoundError):
-            # 如果圖片檢測失敗，回傳原本用副檔名判斷的結果
-            return content_type
-
-    # 非圖片檔案或無法判斷類型，回傳原本的結果
-    return content_type or "application/octet-stream"
-
-
-def upload_and_get_tmp_public_url(
-    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
-) -> str:
-    """上傳檔案到 GCS，並取得公開存取的 URL
-    如果是圖片檔案，會根據實際的圖片格式來調整檔案副檔名
-
-    Args:
-        file_path: 本地檔案路徑
-        botrun_flow_lang_url: botrun flow lang API 的 URL，預設為空字串
-        user_id: 使用者 ID
-
-    Returns:
-        str: 上傳後的公開存取 URL
-    """
-    # 外層 try-except: 處理第一次嘗試與重試邏輯
-    try:
-        return _perform_tmp_file_upload(file_path, botrun_flow_lang_url, user_id)
-    except Exception as e:
-        import traceback
-
-        # 第一次嘗試失敗，記錄錯誤但不立即返回
-        print(f"First attempt failed: {str(e)}")
-        traceback.print_exc()
-
-        # 隨機等待 7-15 秒
-        retry_delay = random.randint(7, 20)
-        print(f"Retrying in {retry_delay} seconds...")
-        time.sleep(retry_delay)
-
-        # 第二次嘗試
-        try:
-            print("Retry attempt...")
-            return _perform_tmp_file_upload(file_path, botrun_flow_lang_url, user_id)
-        except Exception as retry_e:
-            # 第二次嘗試也失敗，記錄錯誤並返回錯誤訊息
-            print(f"Retry attempt failed: {str(retry_e)}")
-            traceback.print_exc()
-            return "Error uploading file"
-
-
-def _perform_tmp_file_upload(
-    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
-) -> str:
-    """執行實際的上傳操作
-
-    Args:
-        file_path: 本地檔案路徑
-        botrun_flow_lang_url: botrun flow lang API 的 URL
-        user_id: 使用者 ID
-
-    Returns:
-        str: 上傳後的公開存取 URL
-    """
-    try:
-        # 如果沒有提供 API URL，使用預設值
-        if not botrun_flow_lang_url or not user_id:
-            raise ValueError("botrun_flow_lang_url and user_id are required")
-
-        # 取得檔案的 MIME type
-        content_type = get_file_content_type(file_path)
-
-        # 從檔案路徑取得檔案名稱
-        file_name = Path(file_path).name
-
-        # 如果是圖片檔案，根據實際的 content type 調整副檔名
-        if content_type.startswith("image/"):
-            # 取得檔案名稱（不含副檔名）
-            name_without_ext = Path(file_name).stem
-            # 根據 content type 決定新的副檔名
-            ext_map = {
-                "image/jpeg": ".jpg",
-                "image/png": ".png",
-                "image/gif": ".gif",
-                "image/webp": ".webp",
-            }
-            new_ext = ext_map.get(content_type, Path(file_name).suffix)
-            file_name = f"{name_without_ext}{new_ext}"
-
-        # 準備 API endpoint
-        url = f"{botrun_flow_lang_url}/api/tmp-files/{user_id}"
-
-        # 準備檔案
-        files = {
-            "file": (file_name, open(file_path, "rb"), content_type),
-            "file_name": (None, file_name),
-            "content_type": (None, content_type),
-        }
-
-        # 發送請求
-        response = requests.post(url, files=files)
-        response.raise_for_status()  # 如果請求失敗會拋出異常
-
-        # 從回應中取得 URL
-        result = response.json()
-        return result.get("url", "")
-
-    except Exception as e:
-        # 這裡把異常往上拋，讓外層的重試邏輯處理
-        raise e
-
-
-def _upload_text_file_with_utf8(
-    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
-) -> str:
-    """上傳文字檔案到 GCS，並設定正確的 UTF-8 charset
-
-    Args:
-        file_path: 本地檔案路徑
-        botrun_flow_lang_url: botrun flow lang API 的 URL
-        user_id: 使用者 ID
-
-    Returns:
-        str: 上傳後的公開存取 URL
-    """
-    try:
-        # 檢查必要參數
-        if not botrun_flow_lang_url or not user_id:
-            raise ValueError("botrun_flow_lang_url and user_id are required")
-
-        # 設定正確的 MIME type with UTF-8 charset
-        content_type = "text/plain; charset=utf-8"
-
-        # 從檔案路徑取得檔案名稱
-        file_name = Path(file_path).name
-
-        # 準備 API endpoint
-        url = f"{botrun_flow_lang_url}/api/tmp-files/{user_id}"
-
-        # 準備檔案
-        files = {
-            "file": (file_name, open(file_path, "rb"), content_type),
-            "file_name": (None, file_name),
-            "content_type": (None, content_type),
-            "public": (None, False),
-        }
-
-        # 發送請求
-        response = requests.post(url, files=files)
-        response.raise_for_status()  # 如果請求失敗會拋出異常
-
-        # 從回應中取得 URL
-        result = response.json()
-        return result.get("url", "")
-
-    except Exception as e:
-        # 這裡把異常往上拋，讓外層的重試邏輯處理
-        raise e
-
-
-async def generate_tmp_text_file(text_content: str) -> str:
-    """
-    Generate a temporary text file from content and upload it to GCS.
-
-    Args:
-        text_content: Text content to write to the file
-        user_id: User ID for file upload
-
-    Returns:
-        str: Storage path for the text file or error message starting with "Error: "
-    """
-    try:
-        # Generate a unique filename
-        import uuid
-
-        file_name = f"tmp_{uuid.uuid4().hex[:8]}.txt"
-
-        # Create file object from text content
-        file_content_bytes = text_content.encode("utf-8")
-        file_object = BytesIO(file_content_bytes)
-
-        # Set content type
-        content_type = "text/plain; charset=utf-8"
-
-        # Build storage path
-        storage_path = f"tmp/{file_name}"
-
-        # Store file to GCS
-        storage = storage_store_factory()
-        success, _ = await storage.store_file(
-            storage_path, file_object, public=False, content_type=content_type
-        )
-
-        if not success:
-            return "Error: Failed to store file"
-
-        return storage_path
-
-    except Exception as e:
-        return f"Error: {str(e)}"
-
-
-async def read_tmp_text_file(storage_path: str) -> str:
-    """
-    Read a temporary text file from GCS storage.
-
-    Args:
-        storage_path: Storage path of the file in GCS (e.g., "tmp/user_id/filename.txt")
-
-    Returns:
-        str: Text content of the file or error message starting with "Error: "
-    """
-    try:
-        storage = storage_store_factory()
-        file_object = await storage.retrieve_file(storage_path)
-
-        if not file_object:
-            return "Error: File not found"
-
-        # Decode the file content as UTF-8 text
-        text_content = file_object.getvalue().decode("utf-8")
-        return text_content
-
-    except Exception as e:
-        return f"Error: {str(e)}"
-
-
-async def upload_html_and_get_public_url(
-    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
-) -> str:
-    """上傳 HTML 檔案到 GCS，並取得公開存取的 URL
-    此函數僅接受 .html 檔案
-
-    Args:
-        file_path: 本地 HTML 檔案路徑
-        botrun_flow_lang_url: botrun flow lang API 的 URL，預設為空字串
-        user_id: 使用者 ID
-
-    Returns:
-        str: 上傳後的公開存取 URL
-    """
-    # 檢查檔案是否為 HTML 檔案
-    if not file_path.lower().endswith(".html"):
-        raise ValueError("Only HTML files are allowed")
-
-    # 外層 try-except: 處理第一次嘗試與重試邏輯
-    try:
-        return await _perform_html_upload(file_path, botrun_flow_lang_url, user_id)
-    except Exception as e:
-        import traceback
-
-        # 第一次嘗試失敗，記錄錯誤但不立即返回
-        print(f"First attempt failed: {str(e)}")
-        traceback.print_exc()
-
-        # 隨機等待 7-15 秒
-        retry_delay = random.randint(7, 20)
-        print(f"Retrying in {retry_delay} seconds...")
-        time.sleep(retry_delay)
-
-        # 第二次嘗試
-        try:
-            print("Retry attempt...")
-            return await _perform_html_upload(file_path, botrun_flow_lang_url, user_id)
-        except Exception as retry_e:
-            # 第二次嘗試也失敗，記錄錯誤並返回錯誤訊息
-            print(f"Retry attempt failed: {str(retry_e)}")
-            traceback.print_exc()
-            return "Error uploading HTML file"
-
-
-async def _perform_html_upload(
-    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
-) -> str:
-    """執行 HTML 檔案的上傳操作
-
-    Args:
-        file_path: 本地 HTML 檔案路徑
-        botrun_flow_lang_url: botrun flow lang API 的 URL (不再使用，保留向後相容性)
-        user_id: 使用者 ID
-
-    Returns:
-        str: 上傳後的公開存取 URL
-    """
-    try:
-        # 檢查必要參數
-        if not user_id:
-            raise ValueError("user_id is required")
-
-        # 檢查檔案是否為 HTML 檔案
-        if not file_path.lower().endswith(".html"):
-            raise ValueError("Only HTML files are allowed")
-
-        # 取得檔案的 MIME type
-        content_type = "text/html"
-
-        # 從檔案路徑取得檔案名稱
-        file_name = Path(file_path).name
-
-        # 讀取檔案內容
-        with open(file_path, "rb") as f:
-            file_content = f.read()
-
-        # 使用內部函數直接上傳，避免 HTTP 自我呼叫
-        from botrun_flow_lang.api.storage_api import _upload_html_file_internal
-
-        # 在同步函數中運行異步函數
-        return await _upload_html_file_internal(
-            user_id, file_content, file_name, content_type
-        )
-
-    except Exception as e:
-        import traceback
-
-        print(f"Error uploading HTML file: {str(e)}")
-        traceback.print_exc()
-        # 這裡把異常往上拋，讓外層的重試邏輯處理
-        raise e
-
-
-async def download_image_from_url(image_url: str) -> tuple[BytesIO, str]:
-    """
-    從 URL 下載圖片到記憶體
-
-    Args:
-        image_url: 圖片的 URL
-
-    Returns:
-        tuple[BytesIO, str]: (圖片內容, content_type)
-
-    Raises:
-        Exception: 下載失敗時拋出例外
-    """
-    try:
-        import httpx
-
-        async with httpx.AsyncClient(timeout=30.0) as client:
-            response = await client.get(image_url, follow_redirects=True)
-            response.raise_for_status()
-
-            # 從 response headers 取得 content type
-            content_type = response.headers.get("content-type", "image/png")
-
-            # 將圖片內容存入 BytesIO
-            image_data = BytesIO(response.content)
-
-            return image_data, content_type
-
-    except Exception as e:
-        raise Exception(f"Failed to download image from URL: {str(e)}")
-
-
-async def upload_image_and_get_public_url(
-    image_url: str, botrun_flow_lang_url: str = "", user_id: str = ""
-) -> str:
-    """
-    從 URL 下載圖片並上傳到 GCS /img 目錄，取得永久公開 URL
-
-    Args:
-        image_url: 圖片來源 URL（如 DALL-E 生成的臨時 URL）
-        botrun_flow_lang_url: botrun flow lang API 的 URL
-        user_id: 使用者 ID
-
-    Returns:
-        str: 上傳後的永久公開存取 URL
-    """
-    try:
-        # 1. 從 URL 下載圖片
-        image_data, content_type = await download_image_from_url(image_url)
-
-        # 2. 生成唯一檔案名稱
-        import uuid
-        from datetime import datetime
-
-        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
-        file_name = f"dalle_{timestamp}_{uuid.uuid4().hex[:8]}.png"
-
-        # 3. 使用內部函數上傳
-        from botrun_flow_lang.api.storage_api import _upload_img_file_internal
-
-        public_url = await _upload_img_file_internal(
-            user_id, image_data.getvalue(), file_name, content_type
-        )
-
-        return public_url
-
-    except Exception as e:
-        import traceback
-
-        print(f"Error uploading image: {str(e)}")
-        traceback.print_exc()
-        raise e
+import requests
+import mimetypes
+import time
+import random
+import os
+from tempfile import NamedTemporaryFile
+from pathlib import Path
+from io import BytesIO
+from .img_util import get_img_content_type
+from botrun_flow_lang.services.storage.storage_factory import storage_store_factory
+
+
+def get_file_content_type(file_path: str) -> str:
+    """根據檔案類型取得對應的 MIME type
+    如果是圖片檔案，會使用 get_img_content_type 來檢測實際的圖片格式
+    如果是其他檔案，則使用副檔名來判斷
+
+    Args:
+        file_path: 檔案路徑
+
+    Returns:
+        str: MIME type，如果無法判斷則返回 application/octet-stream
+    """
+    # 先用副檔名判斷是否為圖片
+    content_type, _ = mimetypes.guess_type(file_path)
+
+    # 如果副檔名顯示是圖片，使用 get_img_content_type 進行實際檢測
+    if content_type and content_type.startswith("image/"):
+        try:
+            return get_img_content_type(file_path)
+        except (ValueError, FileNotFoundError):
+            # 如果圖片檢測失敗，回傳原本用副檔名判斷的結果
+            return content_type
+
+    # 非圖片檔案或無法判斷類型，回傳原本的結果
+    return content_type or "application/octet-stream"
+
+
+def upload_and_get_tmp_public_url(
+    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
+) -> str:
+    """上傳檔案到 GCS，並取得公開存取的 URL
+    如果是圖片檔案，會根據實際的圖片格式來調整檔案副檔名
+
+    Args:
+        file_path: 本地檔案路徑
+        botrun_flow_lang_url: botrun flow lang API 的 URL，預設為空字串
+        user_id: 使用者 ID
+
+    Returns:
+        str: 上傳後的公開存取 URL
+    """
+    # 外層 try-except: 處理第一次嘗試與重試邏輯
+    try:
+        return _perform_tmp_file_upload(file_path, botrun_flow_lang_url, user_id)
+    except Exception as e:
+        import traceback
+
+        # 第一次嘗試失敗，記錄錯誤但不立即返回
+        print(f"First attempt failed: {str(e)}")
+        traceback.print_exc()
+
+        # 隨機等待 7-15 秒
+        retry_delay = random.randint(7, 20)
+        print(f"Retrying in {retry_delay} seconds...")
+        time.sleep(retry_delay)
+
+        # 第二次嘗試
+        try:
+            print("Retry attempt...")
+            return _perform_tmp_file_upload(file_path, botrun_flow_lang_url, user_id)
+        except Exception as retry_e:
+            # 第二次嘗試也失敗，記錄錯誤並返回錯誤訊息
+            print(f"Retry attempt failed: {str(retry_e)}")
+            traceback.print_exc()
+            return "Error uploading file"
+
+
+def _perform_tmp_file_upload(
+    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
+) -> str:
+    """執行實際的上傳操作
+
+    Args:
+        file_path: 本地檔案路徑
+        botrun_flow_lang_url: botrun flow lang API 的 URL
+        user_id: 使用者 ID
+
+    Returns:
+        str: 上傳後的公開存取 URL
+    """
+    try:
+        # 如果沒有提供 API URL，使用預設值
+        if not botrun_flow_lang_url or not user_id:
+            raise ValueError("botrun_flow_lang_url and user_id are required")
+
+        # 取得檔案的 MIME type
+        content_type = get_file_content_type(file_path)
+
+        # 從檔案路徑取得檔案名稱
+        file_name = Path(file_path).name
+
+        # 如果是圖片檔案，根據實際的 content type 調整副檔名
+        if content_type.startswith("image/"):
+            # 取得檔案名稱（不含副檔名）
+            name_without_ext = Path(file_name).stem
+            # 根據 content type 決定新的副檔名
+            ext_map = {
+                "image/jpeg": ".jpg",
+                "image/png": ".png",
+                "image/gif": ".gif",
+                "image/webp": ".webp",
+            }
+            new_ext = ext_map.get(content_type, Path(file_name).suffix)
+            file_name = f"{name_without_ext}{new_ext}"
+
+        # 準備 API endpoint
+        url = f"{botrun_flow_lang_url}/api/tmp-files/{user_id}"
+
+        # 準備檔案
+        files = {
+            "file": (file_name, open(file_path, "rb"), content_type),
+            "file_name": (None, file_name),
+            "content_type": (None, content_type),
+        }
+
+        # 發送請求
+        response = requests.post(url, files=files)
+        response.raise_for_status()  # 如果請求失敗會拋出異常
+
+        # 從回應中取得 URL
+        result = response.json()
+        return result.get("url", "")
+
+    except Exception as e:
+        # 這裡把異常往上拋，讓外層的重試邏輯處理
+        raise e
+
+
+def _upload_text_file_with_utf8(
+    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
+) -> str:
+    """上傳文字檔案到 GCS，並設定正確的 UTF-8 charset
+
+    Args:
+        file_path: 本地檔案路徑
+        botrun_flow_lang_url: botrun flow lang API 的 URL
+        user_id: 使用者 ID
+
+    Returns:
+        str: 上傳後的公開存取 URL
+    """
+    try:
+        # 檢查必要參數
+        if not botrun_flow_lang_url or not user_id:
+            raise ValueError("botrun_flow_lang_url and user_id are required")
+
+        # 設定正確的 MIME type with UTF-8 charset
+        content_type = "text/plain; charset=utf-8"
+
+        # 從檔案路徑取得檔案名稱
+        file_name = Path(file_path).name
+
+        # 準備 API endpoint
+        url = f"{botrun_flow_lang_url}/api/tmp-files/{user_id}"
+
+        # 準備檔案
+        files = {
+            "file": (file_name, open(file_path, "rb"), content_type),
+            "file_name": (None, file_name),
+            "content_type": (None, content_type),
+            "public": (None, False),
+        }
+
+        # 發送請求
+        response = requests.post(url, files=files)
+        response.raise_for_status()  # 如果請求失敗會拋出異常
+
+        # 從回應中取得 URL
+        result = response.json()
+        return result.get("url", "")
+
+    except Exception as e:
+        # 這裡把異常往上拋，讓外層的重試邏輯處理
+        raise e
+
+
+async def generate_tmp_text_file(text_content: str) -> str:
+    """
+    Generate a temporary text file from content and upload it to GCS.
+
+    Args:
+        text_content: Text content to write to the file
+        user_id: User ID for file upload
+
+    Returns:
+        str: Storage path for the text file or error message starting with "Error: "
+    """
+    try:
+        # Generate a unique filename
+        import uuid
+
+        file_name = f"tmp_{uuid.uuid4().hex[:8]}.txt"
+
+        # Create file object from text content
+        file_content_bytes = text_content.encode("utf-8")
+        file_object = BytesIO(file_content_bytes)
+
+        # Set content type
+        content_type = "text/plain; charset=utf-8"
+
+        # Build storage path
+        storage_path = f"tmp/{file_name}"
+
+        # Store file to GCS
+        storage = storage_store_factory()
+        success, _ = await storage.store_file(
+            storage_path, file_object, public=False, content_type=content_type
+        )
+
+        if not success:
+            return "Error: Failed to store file"
+
+        return storage_path
+
+    except Exception as e:
+        return f"Error: {str(e)}"
+
+
+async def read_tmp_text_file(storage_path: str) -> str:
+    """
+    Read a temporary text file from GCS storage.
+
+    Args:
+        storage_path: Storage path of the file in GCS (e.g., "tmp/user_id/filename.txt")
+
+    Returns:
+        str: Text content of the file or error message starting with "Error: "
+    """
+    try:
+        storage = storage_store_factory()
+        file_object = await storage.retrieve_file(storage_path)
+
+        if not file_object:
+            return "Error: File not found"
+
+        # Decode the file content as UTF-8 text
+        text_content = file_object.getvalue().decode("utf-8")
+        return text_content
+
+    except Exception as e:
+        return f"Error: {str(e)}"
+
+
+async def upload_html_and_get_public_url(
+    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
+) -> str:
+    """上傳 HTML 檔案到 GCS，並取得公開存取的 URL
+    此函數僅接受 .html 檔案
+
+    Args:
+        file_path: 本地 HTML 檔案路徑
+        botrun_flow_lang_url: botrun flow lang API 的 URL，預設為空字串
+        user_id: 使用者 ID
+
+    Returns:
+        str: 上傳後的公開存取 URL
+    """
+    # 檢查檔案是否為 HTML 檔案
+    if not file_path.lower().endswith(".html"):
+        raise ValueError("Only HTML files are allowed")
+
+    # 外層 try-except: 處理第一次嘗試與重試邏輯
+    try:
+        return await _perform_html_upload(file_path, botrun_flow_lang_url, user_id)
+    except Exception as e:
+        import traceback
+
+        # 第一次嘗試失敗，記錄錯誤但不立即返回
+        print(f"First attempt failed: {str(e)}")
+        traceback.print_exc()
+
+        # 隨機等待 7-15 秒
+        retry_delay = random.randint(7, 20)
+        print(f"Retrying in {retry_delay} seconds...")
+        time.sleep(retry_delay)
+
+        # 第二次嘗試
+        try:
+            print("Retry attempt...")
+            return await _perform_html_upload(file_path, botrun_flow_lang_url, user_id)
+        except Exception as retry_e:
+            # 第二次嘗試也失敗，記錄錯誤並返回錯誤訊息
+            print(f"Retry attempt failed: {str(retry_e)}")
+            traceback.print_exc()
+            return "Error uploading HTML file"
+
+
+async def _perform_html_upload(
+    file_path: str, botrun_flow_lang_url: str = "", user_id: str = ""
+) -> str:
+    """執行 HTML 檔案的上傳操作
+
+    Args:
+        file_path: 本地 HTML 檔案路徑
+        botrun_flow_lang_url: botrun flow lang API 的 URL (不再使用，保留向後相容性)
+        user_id: 使用者 ID
+
+    Returns:
+        str: 上傳後的公開存取 URL
+    """
+    try:
+        # 檢查必要參數
+        if not user_id:
+            raise ValueError("user_id is required")
+
+        # 檢查檔案是否為 HTML 檔案
+        if not file_path.lower().endswith(".html"):
+            raise ValueError("Only HTML files are allowed")
+
+        # 取得檔案的 MIME type
+        content_type = "text/html"
+
+        # 從檔案路徑取得檔案名稱
+        file_name = Path(file_path).name
+
+        # 讀取檔案內容
+        with open(file_path, "rb") as f:
+            file_content = f.read()
+
+        # 使用內部函數直接上傳，避免 HTTP 自我呼叫
+        from botrun_flow_lang.api.storage_api import _upload_html_file_internal
+
+        # 在同步函數中運行異步函數
+        return await _upload_html_file_internal(
+            user_id, file_content, file_name, content_type
+        )
+
+    except Exception as e:
+        import traceback
+
+        print(f"Error uploading HTML file: {str(e)}")
+        traceback.print_exc()
+        # 這裡把異常往上拋，讓外層的重試邏輯處理
+        raise e
+
+
+async def download_image_from_url(image_url: str) -> tuple[BytesIO, str]:
+    """
+    從 URL 下載圖片到記憶體
+
+    Args:
+        image_url: 圖片的 URL
+
+    Returns:
+        tuple[BytesIO, str]: (圖片內容, content_type)
+
+    Raises:
+        Exception: 下載失敗時拋出例外
+    """
+    try:
+        import httpx
+
+        async with httpx.AsyncClient(timeout=30.0) as client:
+            response = await client.get(image_url, follow_redirects=True)
+            response.raise_for_status()
+
+            # 從 response headers 取得 content type
+            content_type = response.headers.get("content-type", "image/png")
+
+            # 將圖片內容存入 BytesIO
+            image_data = BytesIO(response.content)
+
+            return image_data, content_type
+
+    except Exception as e:
+        raise Exception(f"Failed to download image from URL: {str(e)}")
+
+
+async def upload_image_and_get_public_url(
+    image_url: str, botrun_flow_lang_url: str = "", user_id: str = ""
+) -> str:
+    """
+    從 URL 下載圖片並上傳到 GCS /img 目錄，取得永久公開 URL
+
+    Args:
+        image_url: 圖片來源 URL（如 DALL-E 生成的臨時 URL）
+        botrun_flow_lang_url: botrun flow lang API 的 URL
+        user_id: 使用者 ID
+
+    Returns:
+        str: 上傳後的永久公開存取 URL
+    """
+    try:
+        # 1. 從 URL 下載圖片
+        image_data, content_type = await download_image_from_url(image_url)
+
+        # 2. 生成唯一檔案名稱
+        import uuid
+        from datetime import datetime
+
+        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+        file_name = f"dalle_{timestamp}_{uuid.uuid4().hex[:8]}.png"
+
+        # 3. 使用內部函數上傳
+        from botrun_flow_lang.api.storage_api import _upload_img_file_internal
+
+        public_url = await _upload_img_file_internal(
+            user_id, image_data.getvalue(), file_name, content_type
+        )
+
+        return public_url
+
+    except Exception as e:
+        import traceback
+
+        print(f"Error uploading image: {str(e)}")
+        traceback.print_exc()
+        raise e