media-agent-mcp 2.7.0__tar.gz → 2.7.2__tar.gz

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: media-agent-mcp
-Version: 2.7.0
+Version: 2.7.2
 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
@@ -26,6 +26,7 @@ Requires-Dist: loguru>=0.7.3
 Requires-Dist: imageio-ffmpeg>=0.4.0
 Requires-Dist: Flask>=3.0.0
 Requires-Dist: pydub>=0.25.1
+Requires-Dist: gunicorn>=22.0.0
 Requires-Dist: audioop-lts; python_version >= "3.13"
 Requires-Dist: google-genai>=1.33.0
 Requires-Dist: mcp==1.11.0

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "media-agent-mcp"
-version = "2.7.0"
+version = "2.7.2"
 description = "A Model Context Protocol server for media processing with AI tools"
 readme = "README.md"
 requires-python = ">=3.12"
@@ -31,6 +31,7 @@ dependencies = [
     "imageio-ffmpeg>=0.4.0",
     "Flask>=3.0.0",
     "pydub>=0.25.1",
+    "gunicorn>=22.0.0",
     "audioop-lts; python_version >= '3.13'",
     "google-genai>=1.33.0",
     "mcp==1.11.0",

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/src/media_agent_mcp/ai_models/omni_human.py

@@ -122,8 +122,8 @@ if __name__ == '__main__':
     time1 = time.time()
 
     print(generate_video_from_omni_human(
-        image_url="https://ark-content-generation-v2-ap-southeast-1.tos-ap-southeast-1.volces.com/seedream-3-0-t2i/0217564979274996d76cd7acfd572da46abbcedd46e0f23a31db6.jpeg?X-Tos-Algorithm=TOS4-HMAC-SHA256&X-Tos-Credential=AKLTYWJkZTExNjA1ZDUyNDc3YzhjNTM5OGIyNjBhNDcyOTQ%2F20250829%2Fap-southeast-1%2Ftos%2Frequest&X-Tos-Date=20250829T200528Z&X-Tos-Expires=86400&X-Tos-Signature=abf8c7d516eecd4543691ffee5c3c5bcc7e38d5d047caf7f2fcfe3e41ab897d3&X-Tos-SignedHeaders=host",
-        audio_url="https://carey.tos-ap-southeast-1.bytepluses.com/media_agent/2025-08-30/a72d04349e81454c8065674792256a1d.mp3"
+        image_url="https://carey.tos-ap-southeast-1.bytepluses.com/Art%20Portrait/Art%20Portrait/Art%20Portrait/Art%20Portrait%20(1).jpg",
+        audio_url="https://carey.tos-ap-southeast-1.bytepluses.com/media_agent/2025-09-02/66620bed2a5f4b2cbc641559ff93a2ed.mp3"
     ))
 
     print(time.time() - time1)

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/src/media_agent_mcp/async_server.py

@@ -449,8 +449,6 @@ def main():
                         help='Host for SSE transport (default: 127.0.0.1)')
     parser.add_argument('--port', type=int, default=8000,
                         help='Port for SSE transport (default: 8000)')
-    parser.add_argument('--version', action='store_true',
-                        help='Show version information')
     parser.add_argument('--run-be', action='store_true',
                         help='Run the backend server')
     parser.add_argument('--be-host', type=str, default='0.0.0.0',
@@ -460,10 +458,6 @@ def main():
 
     args = parser.parse_args()
 
-    if args.version:
-        print("Async Media Agent MCP Server v0.1.0")
-        return
-
     if args.run_be:
         logger.info(f"Starting backend server on {args.be_host}:{args.be_port}...")
         be_path = os.path.abspath(os.path.join(os.path.dirname(__file__), 'be'))
@@ -473,22 +467,20 @@ def main():
         try:
             subprocess.run(
                 [
-                    sys.executable,
-                    "-c",
-                    (
-                        "import sys, app; "
-                        "flask_app = getattr(app, '_flask_app', getattr(app, 'app', None)); "
-                        "assert flask_app is not None, 'No Flask application instance found in app.py'; "
-                        "flask_app.run(host=sys.argv[1], port=int(sys.argv[2]), debug=False)"
-                    ),
-                    args.be_host,
-                    str(args.be_port),
+                    "gunicorn",
+                    "--workers",
+                    "4",
+                    "--bind",
+                    f"{args.be_host}:{args.be_port}",
+                    "--log-level",
+                    "debug",
+                    "app:app",
                 ],
                 cwd=be_path,
                 check=True
             )
         except FileNotFoundError:
-            logger.error("`uvicorn` command not found. Please ensure it is installed in your environment.")
+            logger.error("`gunicorn` command not found. Please ensure it is installed in your environment.")
         except subprocess.CalledProcessError as e:
             logger.error(f"Failed to start backend server: {e}")
         return
@@ -528,7 +520,7 @@ def main():
                 mcp.sse_app(),
                 host=args.host,
                 port=args.port,
-                timeout_keep_alive=1200
+                timeout_keep_alive=300
             )
         else:
             # Default stdio transport

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/src/media_agent_mcp/be/app.py

@@ -6,6 +6,7 @@ from flask import Flask
 # Register blueprints from modularized routes
 from media_agent_mcp.be.routes_media import media_bp
 from media_agent_mcp.be.routes_subtitles import subtitles_bp
+from media_agent_mcp.be.routes_omni_human import omni_human_bp
 
 
 logger = logging.getLogger(__name__)
@@ -14,6 +15,7 @@ app = Flask(__name__)
 # Register blueprints (keep original paths unchanged)
 app.register_blueprint(media_bp)
 app.register_blueprint(subtitles_bp)
+app.register_blueprint(omni_human_bp)
 
 
 if __name__ == "__main__":

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/src/media_agent_mcp/be/routes_media.py

@@ -5,6 +5,7 @@ import subprocess
 import tempfile
 from pathlib import Path
 from typing import List, Optional, Dict, Any
+from concurrent.futures import ThreadPoolExecutor
 
 from flask import Blueprint, jsonify, request, send_file, after_this_request
 
@@ -160,13 +161,15 @@ def concat_videos():
                     pass
             return response
 
-        logger.info('Downloading videos')
+        logger.info('Downloading videos concurrently')
         video_paths = []
-        for url in video_urls:
-            dl = download_video_from_url(url)
+        with ThreadPoolExecutor(max_workers=min(len(video_urls), 5)) as executor:
+            download_results = list(executor.map(download_video_from_url, video_urls))
+
+        for dl in download_results:
             if dl.get("status") == "error":
                 return jsonify(dl), 400
-            path = Path(dl["data"]["file_path"])  # type: ignore[index]
+            path = Path(dl["data"]["file_path"])
             temp_files.append(path)
             video_paths.append(path)
 

media_agent_mcp-2.7.2/src/media_agent_mcp/be/routes_omni_human.py

@@ -0,0 +1,34 @@
+from flask import Blueprint, request, jsonify
+from media_agent_mcp.video.omni_human import generate_video_from_omni_human
+
+omni_human_bp = Blueprint("omni_human", __name__)
+
+@omni_human_bp.post("/generate-video-from-omni-human")
+def generate_video():
+    """
+    Generates a video from an image and audio using the Omni Human API.
+    """
+    try:
+        data = request.get_json(silent=True) or {}
+        image_url = data.get("image_url")
+        audio_url = data.get("audio_url")
+
+        if not image_url or not audio_url:
+            return jsonify({
+                "status": "error",
+                "data": None,
+                "message": "Fields image_url and audio_url are required"
+            }), 400
+
+        video_url = generate_video_from_omni_human(image_url, audio_url)
+
+        return jsonify({
+            "status": "success",
+            "data": {
+                "video_url": video_url
+            },
+            "message": "Video generated successfully"
+        })
+
+    except Exception as e:
+        return jsonify({"status": "error", "data": None, "message": str(e)}), 500

media_agent_mcp-2.7.2/src/media_agent_mcp/video/omni_human.py

@@ -0,0 +1,118 @@
+import hashlib
+import json
+import os
+import random
+import time
+from typing import Dict, Any
+import requests
+
+
+def _generate_signature(nonce: int, timestamp: int, security_key: str) -> str:
+    """
+    Generates a signature for the API request.
+    """
+    keys = [str(nonce), str(security_key), str(timestamp)]
+    keys.sort()
+    key_str = "".join(keys).encode("utf-8")
+    signature = hashlib.sha1(key_str).hexdigest()
+    return signature.lower()
+
+
+def _submit_task(image_url: str, audio_url: str, api_key: str, security_key: str) -> str:
+    """
+    Submits a video generation task.
+    """
+    submit_task_url = "https://cv-api.byteintlapi.com/api/common/v2/submit_task"
+    timestamp = int(time.time())
+    nonce = random.randint(0, (1 << 31) - 1)
+    signature = _generate_signature(nonce, timestamp, security_key)
+
+    params = {
+        "api_key": api_key,
+        "timestamp": str(timestamp),
+        "nonce": str(nonce),
+        "sign": signature,
+    }
+    headers = {"Content-Type": "application/json"}
+    body = {
+        "req_key": "realman_avatar_picture_omni_cv",
+        "image_url": image_url,
+        "audio_url": audio_url,
+    }
+
+    response = requests.post(submit_task_url, params=params, headers=headers, json=body)
+    response.raise_for_status()
+    data = response.json()
+    if data["code"] != 10000:
+        raise Exception(f"Failed to submit task: {data['message']}")
+    return data["data"]["task_id"]
+
+
+def _get_task_result(task_id: str, api_key: str, security_key: str) -> Dict[str, Any]:
+    """
+    Gets the result of a video generation task.
+    """
+    get_result_url = "https://cv-api.byteintlapi.com/api/common/v2/get_result"
+    timestamp = int(time.time())
+    nonce = random.randint(0, (1 << 31) - 1)
+    signature = _generate_signature(nonce, timestamp, security_key)
+
+    params = {
+        "api_key": api_key,
+        "timestamp": str(timestamp),
+        "nonce": str(nonce),
+        "sign": signature,
+    }
+    headers = {"Content-Type": "application/json"}
+    body = {
+        "req_key": "realman_avatar_picture_omni_cv",
+        "task_id": task_id,
+    }
+
+    response = requests.post(get_result_url, params=params, headers=headers, json=body)
+    print(response.text)
+    response.raise_for_status()
+    return response.json()
+
+
+def generate_video_from_omni_human(image_url: str, audio_url: str) -> str:
+    """
+    Generates a video from an image and audio using the Omni Human API.
+
+    Args:
+        image_url: The URL of the portrait image.
+        audio_url: The URL of the audio.
+
+    Returns:
+        The URL of the generated video.
+    """
+    api_key = os.environ.get("OMNI_HUMAN_AK")
+    security_key = os.environ.get("OMNI_HUMAN_SK")
+
+    if not api_key or not security_key:
+        raise ValueError("OMNI_HUMAN_AK and OMNI_HUMAN_SK environment variables must be set")
+
+    task_id = _submit_task(image_url, audio_url, api_key, security_key)
+    print('Submitted task, task_id:', task_id)
+    while True:
+        result = _get_task_result(task_id, api_key, security_key)
+        if result["code"] != 10000:
+            raise Exception(f"Failed to get task result: {result['message']}")
+
+        status = result.get("data", {}).get("status")
+        if status == "done":
+            # Parse resp_data JSON string to get video_url
+            resp_data_str = result["data"].get("resp_data", "{}")
+            try:
+                resp_data = json.loads(resp_data_str)
+                video_url = resp_data.get("video_url")
+                if video_url:
+                    return video_url
+                else:
+                    raise Exception(f"No video_url found in response: {resp_data}")
+            except json.JSONDecodeError as e:
+                raise Exception(f"Failed to parse resp_data JSON: {e}")
+        elif status in ["failed", "error"]:
+            raise Exception(f"Video generation failed: {result}")
+
+        time.sleep(5)

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/src/media_agent_mcp.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: media-agent-mcp
-Version: 2.7.0
+Version: 2.7.2
 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
@@ -26,6 +26,7 @@ Requires-Dist: loguru>=0.7.3
 Requires-Dist: imageio-ffmpeg>=0.4.0
 Requires-Dist: Flask>=3.0.0
 Requires-Dist: pydub>=0.25.1
+Requires-Dist: gunicorn>=22.0.0
 Requires-Dist: audioop-lts; python_version >= "3.13"
 Requires-Dist: google-genai>=1.33.0
 Requires-Dist: mcp==1.11.0

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/src/media_agent_mcp.egg-info/SOURCES.txt

@@ -25,6 +25,7 @@ src/media_agent_mcp/be/__init__.py
 src/media_agent_mcp/be/app.py
 src/media_agent_mcp/be/pyproject.toml
 src/media_agent_mcp/be/routes_media.py
+src/media_agent_mcp/be/routes_omni_human.py
 src/media_agent_mcp/be/routes_subtitles.py
 src/media_agent_mcp/be/utils.py
 src/media_agent_mcp/be/uv.lock
@@ -32,6 +33,7 @@ src/media_agent_mcp/be/__pycache__/__init__.cpython-312.pyc
 src/media_agent_mcp/be/__pycache__/app.cpython-312.pyc
 src/media_agent_mcp/be/__pycache__/routes_media.cpython-312.pyc
 src/media_agent_mcp/be/__pycache__/routes_omni.cpython-312.pyc
+src/media_agent_mcp/be/__pycache__/routes_omni_human.cpython-312.pyc
 src/media_agent_mcp/be/__pycache__/routes_subtitles.cpython-312.pyc
 src/media_agent_mcp/be/__pycache__/utils.cpython-312.pyc
 src/media_agent_mcp/be/fonts/en/EduNSWACTCursive-VariableFont_wght.ttf
@@ -48,6 +50,7 @@ src/media_agent_mcp/media_selectors/video_selector.py
 src/media_agent_mcp/storage/__init__.py
 src/media_agent_mcp/storage/tos_client.py
 src/media_agent_mcp/video/__init__.py
+src/media_agent_mcp/video/omni_human.py
 src/media_agent_mcp/video/processor.py
 src/media_agent_mcp/video/stack.py
 src/media_agent_mcp/video/subtitle.py

{media_agent_mcp-2.7.0 → media_agent_mcp-2.7.2}/src/media_agent_mcp.egg-info/requires.txt

@@ -13,6 +13,7 @@ loguru>=0.7.3
 imageio-ffmpeg>=0.4.0
 Flask>=3.0.0
 pydub>=0.25.1
+gunicorn>=22.0.0
 google-genai>=1.33.0
 mcp==1.11.0