vision-agent 0.2.185__py3-none-any.whl → 0.2.187__py3-none-any.whl

vision_agent/tools/tools.py

@@ -1057,23 +1057,25 @@ def video_temporal_localization(
     prompt: str,
     frames: List[np.ndarray],
     model: str = "qwen2vl",
-    chunk_length: Optional[float] = None,
-    chunk_length_seconds: Optional[float] = None,
     chunk_length_frames: Optional[int] = 2,
 ) -> List[float]:
-    """'video_temporal_localization' is a tool that can find objects in a video given a question about it.
-    It returns a list of floats with a value of 1.0 if the object to be found is present in the chunk of video being analyzed.
+    """'video_temporal_localization' will run qwen2vl on each chunk_length_frames
+    value selected for the video. It can detect multiple objects independently per
+    chunk_length_frames given a text prompt such as a referring expression
+    but does not track objects across frames.
+    It returns a list of floats with a value of 1.0 if the objects are found in a given
+    chunk_length_frames of the video.
 
     Parameters:
         prompt (str): The question about the video
         frames (List[np.ndarray]): The reference frames used for the question
-        model (str): The model to use for the inference. Valid values are 'qwen2vl', 'gpt4o', 'internlm-xcomposer'
-        chunk_length (Optional[float]): length of each chunk in seconds
-        chunk_length_seconds (Optional[float]): alternative length for chunk in seconds
+        model (str): The model to use for the inference. Valid values are
+            'qwen2vl', 'gpt4o', 'internlm-xcomposer'
         chunk_length_frames (Optional[int]): length of each chunk in frames
 
     Returns:
-        List[float]: A list of floats with a value of 1.0 if the object to be found is present in the chunk of video
+        List[float]: A list of floats with a value of 1.0 if the objects to be found
+            are present in the chunk_length_frames of the video.
 
     Example
     -------
@@ -1088,10 +1090,6 @@ def video_temporal_localization(
         "model": model,
         "function_name": "video_temporal_localization",
     }
-    if chunk_length is not None:
-        payload["chunk_length"] = chunk_length
-    if chunk_length_seconds is not None:
-        payload["chunk_length_seconds"] = chunk_length_seconds
     if chunk_length_frames is not None:
         payload["chunk_length_frames"] = chunk_length_frames
 
@@ -1790,9 +1788,8 @@ def flux_image_inpainting(
             where 1 indicates areas to be inpainted and 0 indicates areas to be preserved.
 
     Returns:
-        np.ndarray:
-            The generated image(s) as a numpy array in RGB format
-            with values ranging from 0 to 255.
+        np.ndarray: The generated image(s) as a numpy array in RGB format with values
+            ranging from 0 to 255.
 
     -------
     Example:
@@ -1874,6 +1871,9 @@ def extract_frames_and_timestamps(
         >>> extract_frames("path/to/video.mp4")
         [{"frame": np.ndarray, "timestamp": 0.0}, ...]
     """
+    if isinstance(fps, str):
+        # fps could be a string when it's passed in from a web endpoint deployment
+        fps = float(fps)
 
     def reformat(
         frames_and_timestamps: List[Tuple[np.ndarray, float]],
@@ -1937,6 +1937,7 @@ def save_json(data: Any, file_path: str) -> None:
                 return bool(obj)
             return json.JSONEncoder.default(self, obj)
 
+    Path(file_path).parent.mkdir(parents=True, exist_ok=True)
     with open(file_path, "w") as f:
         json.dump(data, f, cls=NumpyEncoder)
 
@@ -1979,6 +1980,7 @@ def save_image(image: np.ndarray, file_path: str) -> None:
     -------
         >>> save_image(image)
     """
+    Path(file_path).parent.mkdir(parents=True, exist_ok=True)
     from IPython.display import display
 
     if not isinstance(image, np.ndarray) or (
@@ -2009,6 +2011,9 @@ def save_video(
         >>> save_video(frames)
         "/tmp/tmpvideo123.mp4"
     """
+    if isinstance(fps, str):
+        # fps could be a string when it's passed in from a web endpoint deployment
+        fps = float(fps)
     if fps <= 0:
         raise ValueError(f"fps must be greater than 0 got {fps}")
 
@@ -2025,6 +2030,8 @@ def save_video(
         output_video_path = tempfile.NamedTemporaryFile(
             delete=False, suffix=".mp4"
         ).name
+    else:
+        Path(output_video_path).parent.mkdir(parents=True, exist_ok=True)
 
     output_video_path = video_writer(frames, fps, output_video_path)
     _save_video_to_result(output_video_path)
@@ -2351,6 +2358,7 @@ FUNCTION_TOOLS = [
     closest_box_distance,
     qwen2_vl_images_vqa,
     qwen2_vl_video_vqa,
+    video_temporal_localization,
 ]
 
 UTIL_TOOLS = [

vision_agent/utils/video.py

@@ -58,6 +58,9 @@ def video_writer(
     fps: float = _DEFAULT_INPUT_FPS,
     filename: Optional[str] = None,
 ) -> str:
+    if isinstance(fps, str):
+        # fps could be a string when it's passed in from a web endpoint deployment
+        fps = float(fps)
     if filename is None:
         filename = tempfile.NamedTemporaryFile(delete=False, suffix=".mp4").name
     container = av.open(filename, mode="w")
@@ -92,6 +95,9 @@ def frames_to_bytes(
         fps: the frames per second of the video
         file_ext: the file extension of the video file
     """
+    if isinstance(fps, str):
+        # fps could be a string when it's passed in from a web endpoint deployment
+        fps = float(fps)
     with tempfile.NamedTemporaryFile(delete=True, suffix=file_ext) as temp_file:
         video_writer(frames, fps, temp_file.name)
 
@@ -120,6 +126,9 @@ def extract_frames_from_video(
             from the start of the video. E.g. 12.125 means 12.125 seconds from the start of
             the video. The frames are sorted by the timestamp in ascending order.
     """
+    if isinstance(fps, str):
+        # fps could be a string when it's passed in from a web endpoint deployment
+        fps = float(fps)
 
     cap = cv2.VideoCapture(video_uri)
     orig_fps = cap.get(cv2.CAP_PROP_FPS)

{vision_agent-0.2.185.dist-info → vision_agent-0.2.187.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vision-agent
-Version: 0.2.185
+Version: 0.2.187
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai
@@ -63,10 +63,10 @@ code to solve the task for them. Check out our discord for updates and roadmaps!
 ## Table of Contents
 - [🚀Quick Start](#quick-start)
 - [📚Documentation](#documentation)
-- [🔍🤖VisionAgent](#vision-agent-basic-usage)
+- [🔍🤖VisionAgent](#visionagent-basic-usage)
 - [🛠️Tools](#tools)
 - [🤖LMMs](#lmms)
-- [💻🤖VisionAgent Coder](#vision-agent-coder)
+- [💻🤖VisionAgent Coder](#visionagent-coder)
 - [🏗️Additional Backends](#additional-backends)
 
 ## Quick Start

{vision_agent-0.2.185.dist-info → vision_agent-0.2.187.dist-info}/RECORD

@@ -20,7 +20,7 @@ vision_agent/tools/__init__.py,sha256=KVP4_6qxOb2lpFdQgQtyDfdkMLL1O6wVZNK19MXp-x
 vision_agent/tools/meta_tools.py,sha256=by7TIbH7lsLIayX_Pe2mS1iw8aeLn2T8yqAo8SkB9Kg,32074
 vision_agent/tools/prompts.py,sha256=V1z4YJLXZuUl_iZ5rY0M5hHc_2tmMEUKr0WocXKGt4E,1430
 vision_agent/tools/tool_utils.py,sha256=VPGqGJ2ZYEJA6AW7K9X7hQv6vRlMtAQcybE4izdToCw,8196
-vision_agent/tools/tools.py,sha256=us3fOV3JIqFB9WidEX6NT65HwJbIxhh59RRvUcMIshI,83251
+vision_agent/tools/tools.py,sha256=rsvQ7cz2xiGiJZme8yb-r-omSarhtC0zapUSt3_pmuo,83541
 vision_agent/tools/tools_types.py,sha256=8hYf2OZhI58gvf65KGaeGkt4EQ56nwLFqIQDPHioOBc,2339
 vision_agent/utils/__init__.py,sha256=7fMgbZiEwbNS0fBOS_hJI5PuEYBblw36zLi_UjUzvj4,244
 vision_agent/utils/exceptions.py,sha256=booSPSuoULF7OXRr_YbC4dtKt6gM_HyiFQHBuaW86C4,2052
@@ -28,8 +28,8 @@ vision_agent/utils/execute.py,sha256=2sIQn45llOENMyrKu3TPINVRLLbOvvZ6SVHFCB9MQUo
 vision_agent/utils/image_utils.py,sha256=rm9GfXvD4JrjnqKrP_f2gfq4SzmqYC0IdC1kKwdn6sk,11303
 vision_agent/utils/sim.py,sha256=ZuSS07TUXFGjipmiQoY8TKRmSes7XXCdtU9PI8PC1sw,5609
 vision_agent/utils/type_defs.py,sha256=BE12s3JNQy36QvauXHjwyeffVh5enfcvd4vTzSwvEZI,1384
-vision_agent/utils/video.py,sha256=fOPR48-SuwMbE5eB5rc2F7lVo6k1mVHn26eEJ0QCslc,5602
-vision_agent-0.2.185.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-vision_agent-0.2.185.dist-info/METADATA,sha256=Wgo1bRpQ3MgqxIDpBiN0Tj0YAUBwRtYCQ7DmhJwgKpY,18330
-vision_agent-0.2.185.dist-info/WHEEL,sha256=7Z8_27uaHI_UZAc4Uox4PpBhQ9Y5_modZXWMxtUi4NU,88
-vision_agent-0.2.185.dist-info/RECORD,,
+vision_agent/utils/video.py,sha256=tRcGp4vEnaDycigL1hBO9k0FBPtDH35fCQciVr9GqYI,6013
+vision_agent-0.2.187.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+vision_agent-0.2.187.dist-info/METADATA,sha256=SYs-27G_7CqSSP8tNatzvnca9BTA-gupcRSVSezmxsw,18328
+vision_agent-0.2.187.dist-info/WHEEL,sha256=7Z8_27uaHI_UZAc4Uox4PpBhQ9Y5_modZXWMxtUi4NU,88
+vision_agent-0.2.187.dist-info/RECORD,,