import asyncio

from yolo_gs.pipeline.handler.DrawDirectionProcessor import DrawDirectionProcessor
from yolo_gs.pipeline.handler.DrawGraffitiProcessor import DrawGraffitiProcessor
from yolo_gs.pipeline.handler.DrawObjectBoxProcessor import DrawObjectBoxProcessor
from yolo_gs.pipeline.handler.GraffitiProcessor import GraffitiProcessor
from yolo_gs.pipeline.handler.ResultLogger import ResultLogger
from yolo_gs.pipeline.handler.RetrogradeProcessor import RetrogradeProcessor
from yolo_gs.pipeline.handler.graffiti_visualizer import GraffitiVisualizer
import cv2

from yolo_gs.pipeline.data_source import video_yolo_iterator
from yolo_gs.pipeline.pipeline import Pipeline

# ===================== 测试运行 =====================
if __name__ == "__main__":
    # 1. 创建管道
    pipeline = Pipeline()

    # 2. 添加同步处理器（按执行顺序执行，每个处理器的输出作为下一个处理器的输入）
    pipeline.add_processor(GraffitiProcessor(name="涂鸦计算器"))
    pipeline.add_processor(RetrogradeProcessor(name="逆行检测器"))
    pipeline.add_processor(DrawDirectionProcessor())
    pipeline.add_processor(DrawGraffitiProcessor())
    pipeline.add_processor(DrawObjectBoxProcessor())
    pipeline.add_processor(ResultLogger(name="结果日志器"))


    # 4. 构建视频+YOLO跟踪迭代器（替换为你的视频路径，或0使用摄像头）
    video_path = "123.mp4"  # 或 0 (摄像头)
    try:
        # 使用跟踪迭代器，可以获取对象ID
        data_iter = video_yolo_iterator(video_path, model_path="yolo11s.pt",tracker="botsort.yaml", device="0")

        # 5. 运行管道，迭代处理数据
        for processed_data in pipeline.run(data_iter):
            # 显示帧（按q退出）
            frame = processed_data.frame
            if frame is not None:
                # 直接显示经过管道处理后的帧，其中已经包含了绘制处理器的效果
                cv2.imshow("YOLO Tracking Pipeline", frame)
                if cv2.waitKey(1) & 0xFF == ord('q'):
                    break
    except Exception as e:
        print(f"运行出错: {e}")
    finally:
        cv2.destroyAllWindows()