import cv2
import argparse
from ultralytics import YOLO
import supervision as sv
import numpy as np
import requests
import time
import pandas as pd
import os
from ultralytics.yolo.utils.plotting import Annotator, colors, save_one_box

ZONE_POLYGON = np.array([
    [0, 0],
    [0.3, 0],
    [0.3, 1],
    [0, 1]
])


def parse_arguments() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description="YOLOv8 live")
    parser.add_argument(
        "--webcam-resolution",
        default=[800, 600],
        nargs=2,
        type=int
    )
    args = parser.parse_args()
    return args

def main():
    args = parse_arguments()
    frame_width, frame_height = args.webcam_resolution
    # print(frame_width,frame_height)
    cap = cv2.VideoCapture(0)
    cap.set(cv2.CAP_PROP_FRAME_WIDTH, frame_width)
    cap.set(cv2.CAP_PROP_FRAME_HEIGHT, frame_height)

    # model = YOLO("yolov8n.pt")
    #使用模組
    model = YOLO("best.pt")

    #設定辨識方框參數
    box_annotator = sv.BoxAnnotator(
        thickness=2,
        text_thickness=2,
        text_scale=1
    )

    #設定區塊參數
    zone_polygon = (ZONE_POLYGON * np.array(args.webcam_resolution)).astype(int)
    zone = sv.PolygonZone(polygon=zone_polygon, frame_resolution_wh=tuple(args.webcam_resolution))
    zone_annotator = sv.PolygonZoneAnnotator(
        zone=zone,
        color=sv.Color.blue(),
        thickness=2,
        text_thickness=4,
        text_scale=2
    )
    # count 計算數量
    # FPS_count 計算FPS
    # start_time 用於計算FPS時間
    count = 0
    FPS_count = 0
    font = cv2.FONT_HERSHEY_SIMPLEX
    color = (255,0,0)

    start_time = time.time()

    while True:
        ret, frame = cap.read()
        result = model(frame, agnostic_nms=True,save_crop=False,save_conf=False)[0]
        detections = sv.Detections.from_yolov8(result)

        #辨識到的名稱與準確度
        labels = [
            f"{model.model.names[class_id]} {confidence:0.2f}"
            for _, confidence, class_id, _
            in detections
        ]

        #辨識名稱
        labels_name = [
            f"{model.model.names[class_id]}"
            for _, confidence, class_id, _
            in detections
        ]
        #辨識準確度
        labels_confidence = [
            f"{confidence:0.2f}"
            for _, confidence, class_id, _
            in detections
        ]

        #畫面顯示辨識框
        frame = box_annotator.annotate(
            scene=frame,
            detections=detections,
            labels=labels
        )

        #抓取辨識框的資料
        boxes_confidence = result.boxes.conf
        boxes_confidence = boxes_confidence * 100
        print(boxes_confidence)

        #顯示區塊內結果
        mask = zone.trigger(detections=detections)

        # 區塊內有辨識到會寫True，判斷等於'Fatwolf'時，判斷辨識框的準確度是否大於等於50，
        # 如果是就計算次數+1且把辨識框內的圖片擷取下來並記錄到txt。
        if mask.any() == True:
            if labels_name == 0:
                continue
            elif labels_name == ['Fatwolf']:
                print('fatwolf')
                #count += 1
                #x1,y1,x2,y2辨識框的座標
                x1 = result.boxes.xyxy[0][0]
                y1 = result.boxes.xyxy[0][1]
                x2 = result.boxes.xyxy[0][2]
                y2 = result.boxes.xyxy[0][3]
                if int(boxes_confidence) >= 50:
                    count += 1
                    roi2 = frame[int(y1) + 4:int(y2) - 2, int(x1) + 4:int(x2) - 2]
                    now_time = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime(time.time()))
                    save_pic_name = now_time+'_'+str(count) + '.jpg'
                    print("存圖片：",save_pic_name, "可信度：",int(boxes_confidence))
                    cv2.imwrite(save_pic_name, roi2)
                    path = 'output.txt'
                    with open(path, 'a') as f:
                        f.write(now_time+'_'+str(count) + '.jpg'+' 可信度：'+str(boxes_confidence)+'\n')
                else:
                    continue
            elif labels_name == ['Bottle']:
                print('Bottle')
        # 畫面添加區塊顯示
        frame = zone_annotator.annotate(scene=frame)
        # 即時顯示計算區塊內辨識到的次數
        cv2.putText(frame, 'Count: {}'.format(count), (10, 50), font, 1, color, 2, cv2.LINE_AA)
        # 計算FPS實際幀數並即時顯示
        FPS_count += 1
        fps = FPS_count / (time.time() - start_time)
        cv2.putText(frame, 'FPS: {:.2f}'.format(fps), (frame.shape[1]-200, 50), font, 1, color, 2, cv2.LINE_AA)

        cv2.imshow("yolov8", frame)

        key = cv2.waitKey(1)
        if key == ord('q'):
            break


if __name__ == "__main__":
    main()