media-agent-mcp 0.3.11__tar.gz → 0.4.0__tar.gz

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: media-agent-mcp
-Version: 0.3.11
+Version: 0.4.0
 Summary: A Model Context Protocol server for media processing with AI tools
 Author-email: Media Agent Team <team@mediaagent.com>
 Keywords: mcp,ai,media,video,image,processing
@@ -22,6 +22,7 @@ Requires-Dist: numpy>=1.24.0
 Requires-Dist: python-dotenv>=1.0.0
 Requires-Dist: volcengine-python-sdk>=1.0.0
 Requires-Dist: volcengine>=1.0.194
+Requires-Dist: openai>=1.97.1
 
 # Media Agent MCP
 

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "media-agent-mcp"
-version = "0.3.11"
+version = "0.4.0"
 description = "A Model Context Protocol server for media processing with AI tools"
 readme = "README.md"
 requires-python = ">=3.12"
@@ -27,6 +27,7 @@ dependencies = [
     "python-dotenv>=1.0.0",
     "volcengine-python-sdk>=1.0.0",
     "volcengine>=1.0.194",
+    "openai>=1.97.1",
 ]
 
 [project.scripts]

media_agent_mcp-0.4.0/src/media_agent_mcp/ai_models/openaiedit.py

@@ -0,0 +1,89 @@
+import os
+from typing import Dict, Any
+import openai
+import requests
+from PIL import Image
+from io import BytesIO
+import tempfile
+import os
+from urllib.parse import urlparse
+
+def openaiedit(image_url: str, prompt: str, size: str = "1024x1024") -> Dict[str, Any]:
+    """
+    Perform image editing using the OpenAI Images API.
+
+    :param image_url: URL of the input image.
+    :param prompt: The editing prompt.
+    :param size: The size of the generated images. Must be one of "256x256", "512x512", or "1024x1024".
+    :return: JSON response with status, data (image URL), and message.
+    """
+    try:
+        client = openai.OpenAI(
+            api_key='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'
+        )
+
+        # Download the image
+        response = requests.get(image_url)
+        response.raise_for_status()
+
+        # Save image to a temporary file
+        parsed_url = urlparse(image_url)
+        file_ext = os.path.splitext(parsed_url.path)[1]
+        if not file_ext:
+            # Fallback if extension is not in URL path
+            content_type = response.headers.get('content-type')
+            if content_type and 'image' in content_type:
+                file_ext = '.' + content_type.split('/')[1]
+            else:
+                file_ext = '.png' # default
+
+        temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=file_ext)
+        try:
+            temp_file.write(response.content)
+            temp_file.close()
+
+            # Call OpenAI API with the local file
+            with open(temp_file.name, "rb") as f:
+                response = client.images.edit(
+                    model="gpt-image-1",
+                    image=f,
+                    prompt=prompt,
+                    n=1,
+                    size=size
+                )
+        finally:
+            os.unlink(temp_file.name) # Clean up the temporary file
+
+        image_url = response.data[0].url
+
+        return {
+            "status": "success",
+            "data": {"image_url": image_url},
+            "message": "Image edited successfully."
+        }
+    except openai.APIError as e:
+        return {
+            "status": "error",
+            "data": None,
+            "message": f"OpenAI API Error: {e}"
+        }
+    except requests.RequestException as e:
+        return {
+            "status": "error",
+            "data": None,
+            "message": f"Error downloading image: {e}"
+        }
+    except Exception as e:
+        return {
+            "status": "error",
+            "data": None,
+            "message": f"An unexpected error occurred: {e}"
+        }
+
+if __name__ == '__main__':
+    # Make sure to set your OPENAI_API_KEY environment variable
+    # For example: export OPENAI_API_KEY='your-api-key'
+    image_url = 'https://carey.tos-ap-southeast-1.bytepluses.com/Art%20Portrait/Art%20Portrait/Art%20Portrait/Art%20Portrait%20(1).jpg'
+    prompt = 'A cute baby sea otter cooking a meal'
+    result = openaiedit(image_url, prompt)
+    print(result)

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/src/media_agent_mcp/ai_models/seededit.py

@@ -1,7 +1,20 @@
 import os
+import tempfile
 from typing import Dict, Any
+
+import requests
 from volcengine.visual.VisualService import VisualService
 
+try:
+    from ..storage.tos_client import upload_to_tos
+except (ImportError, ValueError):
+    # This fallback is for running the script directly
+    import sys
+    import os
+
+    sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..', '..')))
+    from media_agent_mcp.storage.tos_client import upload_to_tos
+
 
 def parse_seededit_response(response) -> Dict[str, Any]:
     """
@@ -44,7 +57,7 @@ def parse_seededit_response(response) -> Dict[str, Any]:
 
 def seededit(image_url, prompt, charactor_keep=False, return_url=True, scale=1, seed=-1) -> Dict[str, Any]:
     """
-    Perform image editing using the VisualService.
+    Perform image editing using the VisualService, then upload the result to TOS.
 
     :param image_url: URL of the input image.
     :param prompt: The editing prompt.
@@ -52,7 +65,7 @@ def seededit(image_url, prompt, charactor_keep=False, return_url=True, scale=1,
     :param return_url: Whether to return image URL or base64 string.
     :param scale: Text influence scale (0.1-1.0).
     :param seed: Random seed for reproducibility.
-    :return: JSON response with status, data (image URL), and message.
+    :return: JSON response with status, data (TOS URL), and message.
     """
     try:
         visual_service = VisualService()
@@ -81,7 +94,43 @@ def seededit(image_url, prompt, charactor_keep=False, return_url=True, scale=1,
         }
 
         response = visual_service.cv_process(form)
-        return parse_seededit_response(response)
+        parsed_response = parse_seededit_response(response)
+
+        if parsed_response['status'] == 'success':
+            generated_image_url = parsed_response['data']['image_url']
+
+            # Download the image from the URL
+            try:
+                image_response = requests.get(generated_image_url, stream=True)
+                image_response.raise_for_status()  # Raise an exception for bad status codes
+            except requests.exceptions.RequestException as e:
+                return {"status": "error", "data": None, "message": f"Failed to download image: {e}"}
+
+            # Create a temporary file to save the image
+            with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as temp_file:
+                for chunk in image_response.iter_content(chunk_size=8192):
+                    temp_file.write(chunk)
+                temp_file_path = temp_file.name
+
+            try:
+                # Upload the temporary file to TOS
+                tos_response = upload_to_tos(temp_file_path)
+                # The URL is in tos_response['data']['url']
+                # The final response should be compatible with other tools
+                if tos_response['status'] == 'success':
+                    return {
+                        "status": "success",
+                        "data": {"image_url": tos_response['data']['url']},
+                        "message": "Image edited and uploaded to TOS successfully"
+                    }
+                else:
+                    return tos_response
+            finally:
+                # Clean up the temporary file
+                os.remove(temp_file_path)
+        else:
+            return parsed_response
+
     except Exception as e:
         return {
             "status": "error",

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/src/media_agent_mcp/async_server.py

@@ -48,16 +48,12 @@ from mcp.server.fastmcp import FastMCP
 
 # Import async wrappers
 from media_agent_mcp.async_wrapper import (
-    async_video_concat_tool,
-    async_video_last_frame_tool,
-    async_seedream_generate_image_tool,
-    async_seedance_generate_video_tool,
-    async_seededit_tool,
-    async_vlm_vision_task_tool,
-    async_image_selector_tool,
-    async_video_selector_tool,
-    async_tos_save_content_tool,
-    cleanup_executor
+    async_video_concat_tool, async_video_last_frame_tool, 
+    async_seedream_generate_image_tool, async_seedance_generate_video_tool, 
+    async_seededit_tool, async_vlm_vision_task_tool, 
+    async_image_selector_tool, async_video_selector_tool, 
+    async_tos_save_content_tool, cleanup_executor,
+    async_openaiedit_tool
 )
 
 # Configure logging
@@ -138,22 +134,35 @@ async def seedance_generate_video_tool(prompt: str, first_frame_image: str,
 
 @mcp.tool()
 @async_retry()
-async def seededit_tool(image_url: str, prompt: str, seed: int = -1, 
-                             scale: float = 0.5, charactor_keep: bool = False) -> dict:
+async def seededit_tool(image_url: str, prompt: str) -> dict:
     """
-    Asynchronously edit an image using Seededit model.
+    Asynchronously edit an image using the OpenAI Images API.
     
     Args:
         image_url: Input image URL for editing
         prompt: Text prompt for image editing
-        seed: Random seed for reproducibility (-1 for random)
-        scale: Influence degree of text description (0-1)
-        charactor_keep: whether to keep the main character in this image, if you wanna change the main character, please keep False
         
     Returns:
         Dictionary with status, data, and message
     """
-    return await async_seededit_tool(image_url, prompt, seed, scale, charactor_keep)
+    return await async_seededit_tool(image_url, prompt)
+
+
+@mcp.tool()
+@async_retry()
+async def openaiedit_tool(image_url: str, prompt: str, size: str = "1024x1024") -> dict:
+    """
+    Asynchronously edit an image using the OpenAI Images API.
+    
+    Args:
+        image_url: Input image URL for editing
+        prompt: Text prompt for image editing
+        size: The size of the generated images. Must be one of "256x256", "512x512", or "1024x1024".
+        
+    Returns:
+        Dictionary with status, data, and message
+    """
+    return await async_openaiedit_tool(image_url, prompt, size)
 
 
 @mcp.tool()

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/src/media_agent_mcp/async_wrapper.py

@@ -18,6 +18,7 @@ from media_agent_mcp.video import concat_videos, extract_last_frame
 from media_agent_mcp.ai_models.seedream import generate_image
 from media_agent_mcp.ai_models.seedance import generate_video
 from media_agent_mcp.ai_models.seededit import seededit
+from media_agent_mcp.ai_models.openaiedit import openaiedit
 from media_agent_mcp.media_selectors.image_selector import select_best_image
 from media_agent_mcp.media_selectors.video_selector import select_best_video
 
@@ -128,16 +129,11 @@ def _sync_seedance_generate_video(prompt: str, first_frame_image: str,
 
 @async_wrapper
 @json_response_wrapper
-def _sync_seededit(image_url: str, prompt: str, seed: int = -1, 
-                  scale: float = 0.5, charactor_keep: bool = False) -> str:
+def _sync_seededit(image_url: str, prompt: str) -> str:
     """Synchronous image editing wrapper."""
     return seededit(
         image_url=image_url,
-        prompt=prompt,
-        charactor_keep=charactor_keep,
-        return_url=True,
-        scale=scale,
-        seed=seed
+        prompt=prompt
     )
 
 
@@ -238,10 +234,10 @@ async def async_seedance_generate_video_tool(prompt: str, first_frame_image: str
     return await _sync_seedance_generate_video(prompt, first_frame_image, last_frame_image, duration, resolution)
 
 
-async def async_seededit_tool(image_url: str, prompt: str, seed: int = -1, 
-                             scale: float = 0.5, charactor_keep: bool = False) -> str:
+async def async_seededit_tool(image_url: str, prompt: str) -> str:
     """Async image editing tool."""
-    return await _sync_seededit(image_url, prompt, seed, scale, charactor_keep)
+    result = await _sync_seededit(image_url, prompt)
+    return json.loads(result)
 
 
 async def async_vlm_vision_task_tool(messages: List) -> str:
@@ -258,6 +254,17 @@ async def async_video_selector_tool(video_paths: List[str], prompt: str) -> str:
     """Async video selector tool."""
     return await _sync_video_selector(video_paths, prompt)
 
+@async_wrapper
+@json_response_wrapper
+def _sync_openaiedit(image_url: str, prompt: str, size: str = "1024x1024") -> str:
+    """Synchronous image editing wrapper for openaiedit."""
+    return openaiedit(image_url=image_url, prompt=prompt, size=size)
+
+async def async_openaiedit_tool(image_url: str, prompt: str, size: str = "1024x1024") -> str:
+    """Async image editing tool for openaiedit."""
+    result = await _sync_openaiedit(image_url, prompt, size)
+    return json.loads(result)
+
 
 async def async_tos_save_content_tool(content: str, file_extension: str = "txt", 
                                      object_key: Optional[str] = None) -> str:

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/src/media_agent_mcp/media_selectors/image_selector.py

@@ -76,6 +76,7 @@ def select_best_image(image_urls: List[str], prompt: str) -> dict:
         except Exception as e:
             logger.error(f"[VLM]Error selecting image: {e}")
             return {
+                "status": "error",
                 "choice": None,
                 "reason": f"Error selecting image: {str(e)}",
                 "url": image_urls[0] if image_urls else None
@@ -94,6 +95,7 @@ def select_best_image(image_urls: List[str], prompt: str) -> dict:
         except Exception as e:
             logger.error(f"[VLM]Error parsing response: {e}")
             return {
+                "status": "error",
                 "choice": None,
                 "reason": f"Error parsing response: {str(e)}",
                 "url": image_urls[0]
@@ -102,6 +104,7 @@ def select_best_image(image_urls: List[str], prompt: str) -> dict:
     except Exception as e:
         logger.error(f"Error selecting image: {e}")
         return {
+            "status": "error",
             "choice": None,
             "reason": f"Error selecting image: {str(e)}",
             "url": image_urls[0] if image_urls else None

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/src/media_agent_mcp/media_selectors/video_selector.py

@@ -116,6 +116,7 @@ def select_best_video(video_urls: List[str], prompt: str) -> dict:
         except Exception as e:
             logger.error(f"[VLM]Error selecting image: {e}")
             return {
+                "status": "error",
                 "choice": None,
                 "reason": f"Error selecting image: {str(e)}",
                 "url": video_urls[0] if video_urls else None
@@ -134,6 +135,7 @@ def select_best_video(video_urls: List[str], prompt: str) -> dict:
         except Exception as e:
             logger.error(f"[VLM]Error parsing response: {e}")
             return {
+                "status": "error",
                 "choice": None,
                 "reason": f"Error parsing response: {str(e)}",
                 "url": video_urls[0]
@@ -142,6 +144,7 @@ def select_best_video(video_urls: List[str], prompt: str) -> dict:
     except Exception as e:
         logger.error(f"Error selecting image: {e}")
         return {
+            "status": "error",
             "choice": None,
             "reason": f"Error selecting image: {str(e)}",
             "url": video_urls[0] if video_urls else None

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/src/media_agent_mcp.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: media-agent-mcp
-Version: 0.3.11
+Version: 0.4.0
 Summary: A Model Context Protocol server for media processing with AI tools
 Author-email: Media Agent Team <team@mediaagent.com>
 Keywords: mcp,ai,media,video,image,processing
@@ -22,6 +22,7 @@ Requires-Dist: numpy>=1.24.0
 Requires-Dist: python-dotenv>=1.0.0
 Requires-Dist: volcengine-python-sdk>=1.0.0
 Requires-Dist: volcengine>=1.0.194
+Requires-Dist: openai>=1.97.1
 
 # Media Agent MCP
 

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/src/media_agent_mcp.egg-info/SOURCES.txt

@@ -11,6 +11,7 @@ src/media_agent_mcp.egg-info/entry_points.txt
 src/media_agent_mcp.egg-info/requires.txt
 src/media_agent_mcp.egg-info/top_level.txt
 src/media_agent_mcp/ai_models/__init__.py
+src/media_agent_mcp/ai_models/openaiedit.py
 src/media_agent_mcp/ai_models/seed16.py
 src/media_agent_mcp/ai_models/seedance.py
 src/media_agent_mcp/ai_models/seededit.py

{media_agent_mcp-0.3.11 → media_agent_mcp-0.4.0}/src/media_agent_mcp.egg-info/requires.txt

@@ -9,3 +9,4 @@ numpy>=1.24.0
 python-dotenv>=1.0.0
 volcengine-python-sdk>=1.0.0
 volcengine>=1.0.194
+openai>=1.97.1