support wan30

Author: mose-x.zm

dashscope/aigc/video_synthesis.py

@@ -47,9 +47,12 @@ class MediaType:
         REFERENCE_IMAGE = "reference_image"
         REFERENCE_VIDEO = "reference_video"
         REFERENCE_VOICE = "reference_voice"
+        REFERENCE_AUDIO = "reference_audio"
         VIDEO = "video"
         FIRST_CLIP = "first_clip"
         DRIVING_AUDIO = "driving_audio"
+        FILE = "file"
+        LINK = "link"
 
     @classmethod
     def call(  # type: ignore[override]
@@ -84,6 +87,7 @@ def call(  # type: ignore[override]
         ratio: Optional[str] = None,
         shot_type: Optional[str] = None,
         audio_setting: Optional[str] = None,
+        enable_thinking: Optional[bool] = None,
         **kwargs,
     ) -> VideoSynthesisResponse:
         """Call video synthesis service and get result.
@@ -126,6 +130,7 @@ def call(  # type: ignore[override]
             shot_type (str, optional): Shot type for video generation.
             audio_setting (str, optional): Audio setting for video,
                 "auto" or "origin".
+            enable_thinking (bool, optional): Whether to use Agentic PE.
             **kwargs: Additional parameters passed to the API.
 
         Raises:
@@ -152,6 +157,8 @@ def call(  # type: ignore[override]
             kwargs["shot_type"] = shot_type
         if audio_setting is not None:
             kwargs["audio_setting"] = audio_setting
+        if enable_thinking is not None:
+            kwargs["enable_thinking"] = enable_thinking
         return super().call(  # type: ignore[return-value]
             model,
             prompt,
@@ -263,7 +270,8 @@ def _get_input(  # pylint: disable=too-many-branches
         if media:
             for i, m_file in enumerate(media):
                 if isinstance(m_file, dict):
-                    if m_file.get("url"):
+                    # exclude when media is a link
+                    if m_file.get("url") and m_file.get("type") != "link":
                         tasks.append(
                             {
                                 "type": "media",
@@ -397,6 +405,7 @@ def async_call(  # pylint: disable=arguments-renamed  # type: ignore[override] #
         ratio: Optional[str] = None,
         shot_type: Optional[str] = None,
         audio_setting: Optional[str] = None,
+        enable_thinking: Optional[bool] = None,
         **kwargs,
     ) -> VideoSynthesisResponse:
         """Create a video synthesis task, and return task information.
@@ -435,6 +444,7 @@ def async_call(  # pylint: disable=arguments-renamed  # type: ignore[override] #
             shot_type (str, optional): Shot type for video generation.
             audio_setting (str, optional): Audio setting for video,
                 "auto" or "origin".
+            enable_thinking (bool, optional): Whether to use Agentic PE.
             **kwargs: Additional parameters passed to the API.
 
         Raises:
@@ -462,6 +472,8 @@ def async_call(  # pylint: disable=arguments-renamed  # type: ignore[override] #
             kwargs["shot_type"] = shot_type
         if audio_setting is not None:
             kwargs["audio_setting"] = audio_setting
+        if enable_thinking is not None:
+            kwargs["enable_thinking"] = enable_thinking
         task_group, function = _get_task_group_and_task(__name__)
 
         inputs, kwargs, task = cls._get_input(
@@ -656,6 +668,7 @@ async def call(  # type: ignore[override] # pylint: disable=arguments-renamed  #
         ratio: str = None,
         shot_type: str = None,
         audio_setting: str = None,
+        enable_thinking: bool = None,
         **kwargs,
     ) -> VideoSynthesisResponse:
         """Call video synthesis service and get result.
@@ -694,6 +707,7 @@ async def call(  # type: ignore[override] # pylint: disable=arguments-renamed  #
             shot_type (str, optional): Shot type for video generation.
             audio_setting (str, optional): Audio setting for video,
                 "auto" or "origin".
+            enable_thinking (bool, optional): Whether to use Agentic PE.
             **kwargs: Additional parameters passed to the API.
 
         Raises:
@@ -720,6 +734,8 @@ async def call(  # type: ignore[override] # pylint: disable=arguments-renamed  #
             kwargs["shot_type"] = shot_type
         if audio_setting is not None:
             kwargs["audio_setting"] = audio_setting
+        if enable_thinking is not None:
+            kwargs["enable_thinking"] = enable_thinking
         task_group, f = _get_task_group_and_task(__name__)
         # pylint: disable=protected-access
         inputs, kwargs, task = VideoSynthesis._get_input(
@@ -792,6 +808,8 @@ async def async_call(  # type: ignore[override] # pylint: disable=arguments-rena
         resolution: str = None,
         ratio: str = None,
         shot_type: str = None,
+        audio_setting: str = None,
+        enable_thinking: bool = None,
         **kwargs,
     ) -> VideoSynthesisResponse:
         """Create a video synthesis task, and return task information.
@@ -828,6 +846,8 @@ async def async_call(  # type: ignore[override] # pylint: disable=arguments-rena
             resolution (str, optional): Output resolution.
             ratio (str, optional): Aspect ratio, e.g. "16:9".
             shot_type (str, optional): Shot type for video generation.
+            enable_thinking (bool, optional): Whether to use Agentic PE.
+            audio_setting (str, optional): Audio setting for video,
             **kwargs: Additional parameters passed to the API.
 
         Raises:
@@ -853,6 +873,10 @@ async def async_call(  # type: ignore[override] # pylint: disable=arguments-rena
             kwargs["ratio"] = ratio
         if shot_type is not None:
             kwargs["shot_type"] = shot_type
+        if audio_setting is not None:
+            kwargs["audio_setting"] = audio_setting
+        if enable_thinking is not None:
+            kwargs["enable_thinking"] = enable_thinking
         task_group, function = _get_task_group_and_task(__name__)
 
         # pylint: disable=protected-access