update

tools/Interface.py

@@ -0,0 +1,175 @@
+import abc
+# import os
+# import pdb
+# import pickle
+import sys
+
+import cv2
+import numpy as np
+from tools.config import gvalue
+
+sys.path.append('./ytracking')
+sys.path.append('./contrast')
+# from ytracking.tracking.dotrack import init_tracker, VideoTracks, boxes_add_fid
+from ytracking.tracking.have_tracking import have_tracked
+from ytracking.track_ import *
+from contrast.logic import datacollection, similarityResult, similarity
+from PIL import Image
+from tools.config import gvalue
+
+class AiInterface(metaclass=abc.ABCMeta):
+    @abc.abstractmethod
+    def getTrackingBox(self):
+        pass
+
+    @abc.abstractmethod
+    def getSimilarity(self):
+        pass
+
+
+class AiClass(AiInterface):
+    def __init__(self):
+        pass
+
+    def get_xyxy_coordinates(self, box, frame_id_img):
+        """
+        计算并返回边界框的坐标。
+        """
+        try:
+            x1 = max(0, int(box[0]))
+            x2 = min(frame_id_img.shape[1], int(box[2]))
+            y1 = max(0, int(box[1]))
+            y2 = min(frame_id_img.shape[0], int(box[3]))
+            return x1, y1, x2, y2
+        except IndexError as e:
+            raise ValueError("边界框坐标超出图像尺寸") from e
+
+    def getTrackingBox(self, bboxes, features_dict, camera_id, frame_id_img, save_imgs_dir):
+        """
+        根据提供的边界框和帧图像，返回图像列表和轨迹ID列表。
+        """
+
+        image_lists = {}
+        track_id_list = []
+
+        gt = Profile()
+        with gt:
+            vts = have_tracked(bboxes, features_dict, camera_id)
+
+        nn = 0
+        for res in vts.Residual:
+            for box in res.boxes:
+                try:
+                    box = [int(i) for i in box.tolist()]
+                    print('box[7] >>>> {}'.format(box[7]))
+                    x1, y1, x2, y2 = self.get_xyxy_coordinates(box, frame_id_img[box[7]])
+                    gvalue.track_y_lists.append(y1)
+                    c_img = frame_id_img[box[7]][y1:y2, x1:x2][:, :, ::-1]
+
+                    # c_img = frame_id_img[box[7]][box[1]:box[3], box[0]:box[2]][:, :, ::-1]
+                    img_pil = Image.fromarray(c_img.astype('uint8'), 'RGB')
+
+                    img_pil.save(os.sep.join([save_imgs_dir, str(nn) + '.jpg']))
+                    nn += 1
+
+                    track_id = str(box[4])
+                    track_id_list.append(track_id)
+                    if track_id not in image_lists:
+                        image_lists[track_id] = []
+                    image_lists[track_id].append(img_pil)
+                except Exception as e:
+                    print("y1: {}, y2: {}, x1:{} x2:{}".format(box[2], box[3], box[0], box[1]))
+                    print("x:{}, y:{}".format(frame_id_img[box[7]].shape[1], frame_id_img[box[7]].shape[0]))
+                    print(f"处理边界框时发生错误: {e}")
+                    continue
+
+        all_image_list = list(image_lists.values())
+        trackIdList = list(set(track_id_list))
+
+        return all_image_list, trackIdList
+
+    @staticmethod
+    def process_topn_data(source_data):
+        if source_data is None:
+            return None
+
+        if not isinstance(source_data, dict):
+            raise ValueError("输入数据必须是字典类型")
+
+        if not source_data:
+            return {}
+
+        total = {}
+        carId_barcode_trackId_list = []
+        data_category = []
+
+        for category, category_data in source_data.items():
+            carId_barcode_trackId_list.append(category)
+            for car_id, similarity in category_data.items():
+                data_category.append({'carId_barcode_trackId_n': car_id, 'similarity': similarity})
+
+        total['carId_barcode_trackId'] = carId_barcode_trackId_list
+        total['data'] = data_category
+
+        return total
+
+    @staticmethod
+    def process_top10_data(source_data):
+        if source_data is None:
+            return None
+
+        if not isinstance(source_data, dict):
+            raise ValueError("输入数据必须是字典类型")
+
+        if not source_data:
+            return {}
+
+        total = {}
+        data_category = []
+
+        for category, category_data in source_data.items():
+            trackid = category.split('_')[-1]
+            barcode = category.split('_')[-2]
+            for car_id, similarity in category_data.items():
+                data_category.append({'barcode': car_id, 'similarity': similarity, 'trackid': trackid})
+
+        total['barcode'] = barcode
+        total['data'] = data_category
+        return total
+
+    def getSimilarity(self, model, queueImgs):
+        data_collection = datacollection()
+        similarityRes = similarityResult()
+        data_collection.barcode_flag = queueImgs['barcode_flag']
+        data_collection.add_flag = queueImgs['add_flag']
+        data_collection.barcode_list = queueImgs['barcode_list'].strip("'").split(',')
+        data_collection.queImgsDict = queueImgs
+
+        similarityRes = similarity().getSimilarity(model, data_collection, similarityRes)
+        # print('similarityRes.top10: ------------------ {}'.format(similarityRes.top10))
+        if similarityRes.top1:
+            similarityRes.top1 = {"barcode": list(similarityRes.top1.keys())[0],
+                                  "similarity": list(similarityRes.top1.values())[0]}
+        # similarityRes.tempLibList = gvalue.tempLibList
+        # print('-------------------------', gvalue.tempLibLists)
+        if gvalue.tempLibLists.get(gvalue.mac_id) is not None:
+            similarityRes.tempLibList = gvalue.tempLibLists[gvalue.mac_id]
+        else:
+            similarityRes.tempLibList = []
+        similarityresult = {
+            'top10': AiClass.process_top10_data(similarityRes.top10),
+            'top1': similarityRes.top1,
+            'topn': AiClass.process_topn_data(similarityRes.topn),
+            'tempLibList': similarityRes.tempLibList,
+            'sequenceId': queueImgs['sequenceId'],
+        }
+        return similarityresult
+
+
+if __name__ == '__main__':
+    AI = AiClass()
+
+    # track_boxes, frame_id_img = run()
+    # AI.getTrackingBox(track_boxes, frame_id_img)
+    # print('=== test ===')
+    # AI.getSimilarity(cfg.queueImgs)

tools/config.py

@@ -0,0 +1,93 @@
+
+# from yacs.config import CfgNode as CfgNode
+import torchvision.transforms as T
+import torch
+import os
+
+class globalVal:
+    tempLibList = []
+    tempLibLists = {}
+    track_y_lists = []
+    mac_id = None
+    back_return_similarity = []
+    back_add_similarity = []
+    front_return_similarity = []
+    front_add_similarity = []
+    comprehensive_similarity = []
+
+class config:
+    save_videos_dir = 'videos'
+
+    #url 
+    # push_url = 'http://api.test2.ieemoo.cn/emoo-api/intelligence/addVideoPathBySequenceId.do'
+    push_url = 'https://api.test2.ieemoo.cn/emoo-api/intelligence/addVideoPathBySequenceId.do'  # 闲时上传
+    get_config_url = 'https://api.test2.ieemoo.cn/emoo-api/intelligence/addVideoPathByStoreId.do'  # 闲时上传相应配置
+    storidPth = 'tools/storeId.txt'
+
+    #obs update
+    obs_access_key_id = 'LHXJC7GIC2NNUUHHTNVI'
+    obs_secret_access_key = 'sVWvEItrFKWPp5DxeMvX8jLFU69iXPpzkjuMX3iM'
+    obs_server = 'https://obs.cn-east-3.myhuaweicloud.com'
+    obs_bucketName = 'ieemoo-ai'
+
+    keys = ['x', 'y', 'w', 'h', 'track_id', 'score', 'cls', 'frame_index']
+    
+    obs_root_dir = 'ieemoo_ai_data'
+
+    #contrast config
+    host = "192.168.1.28"
+    port = "19530"
+    embedding_size = 256
+    img_size = 224
+    test_transform = T.Compose([
+        T.ToTensor(),
+        T.Resize((224, 224)),
+        T.ConvertImageDtype(torch.float32),
+        T.Normalize(mean=[0.5], std=[0.5]),
+    ])
+
+    # test_model = "./tools/ckpts/MobilenetV3Large_noParallel_2624.pth"
+    test_model = "./tools/ckpts/resnet18_0721_best.pth"
+    tracking_model = "./tools/ckpts/best_158734_cls11_noaug10.pt"
+    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+    httpHost = '0.0.0.0'
+    httpPort = 8088
+
+    #tracking config
+    botsort = './ytracking/tracking/trackers/cfg/botsort.yaml'
+    incart = './tools/Template_images/incart.png'
+    outcart = './tools/Template_images/outcart.png'
+    cartboarder = './tools/Template_images/cartboarder.png'
+    edgeline = './tools/Template_images/edgeline.png'
+    cartedge = './tools/Template_images/cartedge.png'
+    incart_ftmp = './tools/Template_images/incart_ftmp.png'
+
+
+    action_type = {
+        "1": 'purchase',
+        '2': 'jettison',
+        '3': 'unswept_purchase',
+        '4': 'unswept_jettison'
+    }
+    camera_id = {
+        '0': 'back',
+        '1': 'front',
+    }
+    recognize_result = {
+        '01': 'uncatalogued',
+        '02': 'fail',
+        '03': 'exception',
+        '04': 'pass',
+    }
+
+    # reid config
+    backbone = 'resnet18'  # [resnet18, mobilevit_s, mobilenet_v2, mobilenetv3]
+    batch_size = 8
+
+    model_path = './tools/ckpts/best_resnet18_0515.pth'
+
+    temp_video_name = None
+
+cfg = config()
+gvalue = globalVal()

tools/getResult.py

@@ -0,0 +1,78 @@
+from tools.Interface import AiInterface, AiClass
+# from Interface import AiInterface, AiClass
+from config import cfg
+import numpy as np
+from PIL import Image
+import pdb
+from track_ import *
+import time
+
+'''
+跟踪与对比输入
+'''
+ai_obj = AiClass()
+
+
+def deal_similarity_data(message):
+    # print('message --- > {}'.format(message))
+    car_mac = message['videoIds'].split('_')[-3]
+    similar = {'add_flag': True if message['action'] in {'1', '3'} else False,
+               'barcode_flag': True if not message['barcode'] == 'null' else False}
+    if similar['add_flag'] and similar['barcode_flag']:  # 加购有barcode
+        for Id, image_list in zip(message['trackIdList'],
+                                  message['images']):
+            similar[car_mac + '_' + message['barcode'] + '_' + Id] = image_list
+        similar['barcode_list'] = message['barcodeList']
+    elif similar['add_flag'] and (not similar['barcode_flag']):  # 加购无barcode
+        for Id, image_list in zip(message['trackIdList'],
+                                  message['images']):
+            similar[car_mac + '_' + Id] = image_list
+        similar['barcode_list'] = message['barcodeList']
+    elif (not similar['add_flag']) and similar['barcode_flag']:  # 退购有barcode
+        for Id, image_list in zip(message['trackIdList'],
+                                  message['images']):
+            similar[car_mac + '_' + message['barcode'] + '_' + Id] = image_list
+        similar['barcode_list'] = message['barcodeList']
+    else:  # 退购无barcode
+        for Id, image_list in zip(message['trackIdList'],
+                                  message['images']):
+            similar[car_mac + '_' + Id] = image_list
+        similar['barcode_list'] = message['barcodeList']
+    similar['sequenceId'] = message['sequenceId']
+    return similar
+
+
+def get_similarity_result(obj: AiInterface, videopth, model, camera_id, message):
+    """
+    获取相似度结果。
+
+    :param videopth:
+    :param obj: AiInterface 对象，用于获取跟踪框和相似度数据。
+    :param message: dict, 可选参数，包含跟踪数据和相似度处理结果。
+    :return: 相似度结果。
+    """
+    try:
+        track_boxes, features_dict, frame_id_img = run(model, source=videopth)
+        allimages, trackIdList = obj.getTrackingBox(track_boxes, features_dict, camera_id, frame_id_img)
+        message['trackIdList'] = trackIdList
+        message['images'] = allimages
+        message = deal_similarity_data(message)
+        similarityRes = obj.getSimilarity(model, message)
+    except ValueError as ve:
+        print('ve >>>> ', ve)
+        similarityRes = {'top10': {},
+                         'top1': {},
+                         'topn': {},
+                         'tempLibList': [],
+                         'sequenceId': message['sequenceId']
+                         }
+    return similarityRes
+
+
+if __name__ == '__main__':
+    message = {
+        'action': '1',
+        'barcode': '084501446314',
+        'sequenceId': 'test'
+    }
+    get_similarity_result(ai_obj, message)

tools/getbox.py

@@ -0,0 +1,153 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Tue May 21 15:25:23 2024
+
+@author: ieemoo-zl003
+"""
+
+import os
+import numpy as np
+
+# 替换为你的目录路径
+files_path = 'D:/Project/ieemoo/kmeans/comparisonData/deletedBarcode_10_0709_am/err_pair/6902088146356/20240709-105804_6902088146356'
+
+
+def str_to_float_arr(s):
+    # 移除字符串末尾的逗号（如果存在）
+    if s.endswith(','):
+        s = s[:-1]
+
+        # 使用split()方法分割字符串，然后将每个元素转化为float
+    float_array = [float(x) for x in s.split(",")]
+    return float_array
+
+
+def extract_tracker_input_boxes_feats(file_name):
+    framesId = []
+    boxes = []
+    feats = []
+
+
+    frame_id = 0
+    with open(file_name, 'r', encoding='utf-8') as file:
+        for line in file:
+            line = line.strip()  # 去除行尾的换行符和可能的空白字符
+
+            # 跳过空行
+            if not line:
+                continue
+            if line.find("frameId") >= 0:
+                frame_id = line[line.find("frameId:") + 8:].strip()
+
+            # 检查是否以'box:'或'feat:'开始
+            if line.find("box:") >= 0 and line.find("output_box:") < 0:
+                boxes.append(line[line.find("box:") + 4:].strip())  # 去掉'box:'并去除可能的空白字符
+                framesId.append(frame_id)
+
+            if line.find("feat:") >= 0:
+                feats.append(line[line.find("feat:") + 5:].strip())  # 去掉'box:'并去除可能的空白字符
+
+    return boxes, feats, framesId
+
+
+def find_string_in_array(arr, target):
+    """
+    在字符串数组中找到目标字符串对应的行（索引）。
+
+    参数:
+    arr -- 字符串数组
+    target -- 要查找的目标字符串
+
+    返回:
+    目标字符串在数组中的索引。如果未找到，则返回-1。
+    """
+    parts = target.split(',')
+    box_substrings = ','.join(parts[:4])
+    conf_substring = ','.join(parts[5:6])
+    for i, s in enumerate(arr):
+        if s.find(box_substrings) >= 0 and s.find(conf_substring[:7]) >= 0:
+            return i
+    return -1
+
+
+def extract_tracker_output_boxes_feats(read_file_name):
+    input_boxes, input_feats, framesId = extract_tracker_input_boxes_feats(read_file_name)
+
+    boxes = []
+    feats = []
+    with open(read_file_name, 'r', encoding='utf-8') as file:
+        for line in file:
+            line = line.strip()  # 去除行尾的换行符和可能的空白字符
+
+            # 跳过空行
+            if not line:
+                continue
+
+                # 检查是否以'output_box:'开始
+            if line.find("output_box:") >= 0:
+                boxes_str = line[line.find("output_box:") + 11:].strip()
+                boxes.append(boxes_str)  # 去掉'output_box:'并去除可能的空白字符
+                index = find_string_in_array(input_boxes, boxes_str)
+                feat_f = str_to_float_arr(input_feats[index])
+                norm_f = np.linalg.norm(feat_f)
+                feat_f = feat_f / norm_f
+                feats.append(feat_f)
+    return input_boxes, input_feats, boxes, feats, framesId
+
+
+def extract_tracking_output_boxes_feats(read_file_name):
+    tracker_boxes, tracker_feats, input_boxes, input_feats, framesId = extract_tracker_output_boxes_feats(
+        read_file_name)
+    boxes = []
+    feats = []
+    boxes_frames_id = []
+    tracking_flag = False
+    tracking_num_cnt = 0
+    with open(read_file_name, 'r', encoding='utf-8') as file:
+        for line in file:
+            line = line.strip()  # 去除行尾的换行符和可能的空白字符
+
+            # 跳过空行
+            if not line:
+                continue
+
+            if tracking_flag:
+                if line.find("tracking_") >= 0:
+                    tracking_flag = False
+                    tracking_num_cnt = tracking_num_cnt + 1
+                else:
+                    boxes.append(line)
+                    index = find_string_in_array(input_boxes, line)
+                    feats.append(input_feats[index])
+
+                    if tracking_num_cnt == 0:
+                        index = find_string_in_array(tracker_boxes, line)
+                        boxes_frames_id.append(framesId[index])
+            # 检查是否以tracking_'开始
+            if line.find("tracking_") >= 0:
+                tracking_flag = True
+
+    return tracker_boxes, tracker_feats, input_boxes, input_feats, boxes, feats, boxes_frames_id
+
+def find_index_feats(files_path):
+    # 遍历目录下的所有文件和目录
+    all_boxes, boboxes_frames_ids, Boxes, framesIds =[],[],[],[]
+    for filename in os.listdir(files_path):
+        # 构造完整的文件路径
+        file_path = os.path.join(files_path, filename)
+        # 判断是否是文件
+        if os.path.isfile(file_path):
+            # 打开文件
+            if filename.endswith('data') and (not 'tracking' in filename):
+                tracker_boxes, tracker_feats, input_boxes, input_feats, boxes, feats, boxes_frames_id = extract_tracking_output_boxes_feats(file_path)
+                box, feats, framesId = extract_tracker_input_boxes_feats(file_path)
+                Boxes += box
+                framesIds += framesId
+                all_boxes += boxes[:len(boxes_frames_id)]
+                boboxes_frames_ids += boxes_frames_id
+    # print(all_boxes)
+    # print(boboxes_frames_ids)
+    return all_boxes, boboxes_frames_ids, tracker_boxes, Boxes, framesIds
+
+if __name__ == '__main__':
+    find_index_feats(files_path)

tools/initModel.py

@@ -0,0 +1,45 @@
+import torch
+from ytracking.models.common import DetectMultiBackend
+from ytracking.utils.torch_utils import select_device
+from tools.config import cfg
+from contrast.model.resnet_pre import resnet18
+from ytracking.tracking.utils import Boxes, IterableSimpleNamespace, yaml_load
+from ytracking.tracking.trackers import BOTSORT, BYTETracker
+# import mediapipe as mp
+# from pymilvus import (
+#     connections,
+#     utility,
+#     FieldSchema, CollectionSchema, DataType,
+#     Collection,
+#     Milvus
+# )
+
+class Models:
+    def __init__(self):
+        self.yoloModel = None
+        self.reidModel = None
+        self.similarityModel = None
+        self.Milvus = None
+        self.device = 'cpu'
+
+    def initSimilarityModel(self):
+        # model = MobileNetV3_Large().to(cfg.device)
+        model = resnet18().to(cfg.device)
+        # model.load_state_dict(torch.load(cfg.test_model, map_location=cfg.device))
+        model.load_state_dict(torch.load(cfg.model_path, map_location=cfg.device))
+        model.eval()
+        return model
+
+    def initYoloModel(self):
+        device = select_device(self.device)
+        model = DetectMultiBackend(cfg.tracking_model, device=device, dnn=False, fp16=False)
+        return model
+
+    def initModel(self):
+        self.yoloModel = self.initYoloModel()
+        self.similarityModel = self.initSimilarityModel()
+
+models = Models()
+
+if __name__ == "__main__":
+    Models().initModel()

tools/operate_usearch.py

@@ -0,0 +1,153 @@
+import os
+import numpy as np
+from usearch.index import Index
+import json
+import statistics
+
+
+def create_index():
+    index = Index(
+        ndim=256,
+        metric='cos',
+        # dtype='f32',
+        dtype='f16',
+        connectivity=32,
+        expansion_add=40,#128,
+        expansion_search=10,#64,
+        multi=True
+    )
+    return index
+
+def compare_feature(features1, features2, model = '1'):
+    """
+    :param model 比对策略
+        '0':模拟一个轨迹的图像(所有的图像、或者挑选的若干图像)与标准库,先求每个图片与标准库的最大值，再求所有图片对应最大值的均值
+        '1':带对比的所有相似度的均值
+        '2':比对1:1的最大值
+    :param feature1:
+    :param feature2:
+    :return:
+    """
+    similarity_group, similarity_groups = [], []
+    if model == '0':
+        for feature1 in features1:
+            for feature2 in features2[0]:
+                similarity = np.dot(feature1, feature2) / (np.linalg.norm(feature1) * np.linalg.norm(feature2))
+                similarity_group.append(similarity)
+            similarity_groups.append(max(similarity_group))
+            similarity_group = []
+        return sum(similarity_groups)/len(similarity_groups)
+
+    elif model == '1':
+        feature2 = features2[0]
+        for feature1 in features1:
+            for num in range(len(feature2)):
+                similarity = np.dot(feature1, feature2[num]) / (np.linalg.norm(feature1) * np.linalg.norm(feature2[num]))
+                similarity_group.append(similarity)
+            similarity_groups.append(sum(similarity_group) / len(similarity_group))
+            similarity_group = []
+        # return sum(similarity_groups)/len(similarity_groups), max(similarity_groups)
+        if len(similarity_groups) == 0:
+            return -1
+        return sum(similarity_groups)/len(similarity_groups)
+    elif model == '2':
+        feature2 = features2[0]
+        for feature1 in features1:
+            for num in range(len(feature2)):
+                similarity = np.dot(feature1, feature2[num]) / (np.linalg.norm(feature1) * np.linalg.norm(feature2[num]))
+                similarity_group.append(similarity)
+        return max(similarity_group)
+
+
+
+def get_barcode_feature(data):
+    barcode = data['key']
+    features = data['value']
+    return [barcode] * len(features), features
+
+
+def analysis_file(file_path):
+    """
+    :param file_path:
+    :return:
+    """
+    barcodes, features = [], []
+    with open(file_path, 'r', encoding='utf-8') as f:
+        data = json.load(f)
+    for dic in data['total']:
+        barcode, feature = get_barcode_feature(dic)
+        barcodes.append(barcode)
+        features.append(feature)
+    return barcodes, features
+
+
+def create_base_index(index_file_pth=None,
+                      barcodes=None,
+                      features=None,
+                      save_index_name=None):
+    index = create_index()
+    if index_file_pth is not None:
+        # save_index_name = index_file_pth.split('json')[0] + 'usearch'
+        save_index_name = index_file_pth.split('json')[0] + 'data'
+        barcodes, features = analysis_file(index_file_pth)
+    else:
+        assert barcodes is not None and features is not None, 'barcodes and features must be not None'
+    for barcode, feature in zip(barcodes, features):
+        index.add(np.array(barcode), np.array(feature))
+    index.save(save_index_name)
+
+def get_feature_index(index_file_pth=None,
+                      barcodes=None):
+    assert index_file_pth is not None, 'index_file_pth must be not None'
+    index = Index.restore(index_file_pth, view=True)
+    feature_lists = index.get(np.array(barcodes))
+    print("memory {}     size  {}".format(index.memory_usage, index.size))
+    return feature_lists
+
+def search_in_index(query=None,
+                    barcode=None,  # barcode -> int or np.ndarray
+                    index_name=None,
+                    temp_index=False,  # 是否为临时库
+                    model='0',
+                    ):
+    if temp_index:
+        assert index_name is not None, 'index_name must be not None'
+        index = Index.restore(index_name, view=True)
+        if barcode is not None:  # 1:1对比测试
+            feature_lists = index.get(np.array(barcode))
+            results = compare_feature(query, feature_lists)
+        else:
+            results = index.search(query, count=5)
+        return results
+    else:  # 标准库
+        assert index_name is not None, 'index_name must be not None'
+        index = Index.restore(index_name, view=True)
+        if barcode is not None:  # 1:1对比测试
+            feature_lists = index.get(np.array(barcode))
+            results = compare_feature(query, feature_lists, model)
+        else:
+            results = index.search(query, count=10)
+        return results
+
+def delete_index(index_name=None, key=None, index=None):
+    assert key is not None, 'key must be not None'
+    if index is None:
+        assert index_name is not None, 'index_name must be not None'
+        index = Index.restore(index_name, view=True)
+        index.remove(index_name)
+    else:
+        index.remove(key)
+
+if __name__ == '__main__':
+    # index_file_pth = '../search_library/data_0923.json'
+    # create_base_index(index_file_pth)
+
+    # index_file_pth = '../search_library/test_index_10_normal_0717.usearch'
+    # # index_file_pth = '../search_library/data_10_normal_0718.index'
+    # search_in_index(query='693', index_name=index_file_pth, barcode='6934024590466')
+
+     # check index data file
+    index_file_pth = '../search_library/data_0923.data'
+    # # get_feature_index(index_file_pth, ['6901070602818'])
+    get_feature_index(index_file_pth, ['6934230050105'])
+

tools/storeId.txt

@@ -0,0 +1 @@
+32011001

tools/uploadvideos.py

@@ -0,0 +1,159 @@
+from obs import ObsClient
+import obs
+import sys
+import os
+from tools.config import cfg
+
+import requests
+import argparse
+
+class uploadVideos:
+    def __init__(self):
+        self.obsClient, self.headers = self.InitObsClient()
+
+    def InitObsClient(self):
+        # 创建ObsClient实例
+        obsClient = ObsClient(
+            access_key_id=cfg.obs_access_key_id,  # 配置访问密钥ID
+            secret_access_key=cfg.obs_secret_access_key,  # 配置访问密钥
+            server=cfg.obs_server)  # 配置服务器地址
+        headers = obs.SetObjectMetadataHeader()  # 设置对象元数据头
+        headers.cacheControl = "no-cache"  # 设置缓存控制为不缓存
+        return obsClient, headers  # 返回ObsClient实例和元数据头
+
+    def upload(self, video_name, video_dir=None):
+        # 检查是否已初始化ObsClient实例
+        if not self.obsClient:
+            raise ValueError("请先初始化ObsClient实例")
+
+        class uploadResult:
+            def __init__(self, video_path=None, squenceId=None):
+                self.video_path = video_path
+                self.squenceId = squenceId
+
+        uploadRes = uploadResult()
+
+        # 解析视频名称获取相关信息
+        information = video_name.split('.')[0].split('_')
+        action_category = cfg.action_type[information[-1]]  # 动作类别
+        camera_id = cfg.camera_id[information[-3]]  # 摄像头ID
+        recognize_result = cfg.recognize_result[information[0]]  # 识别结果
+        time = information[1].split('-')[0]  # 时间
+        squenceId = information[1]
+
+
+        # 构建OSS对象键
+        objectkey = os.path.join(cfg.obs_root_dir, time, action_category, camera_id, recognize_result, video_name)
+        if video_dir is None:
+            file_path = os.sep.join([cfg.save_videos_dir, squenceId, video_name])  # 本地文件路径
+        else:
+            file_path = os.sep.join([video_dir, squenceId, video_name])
+
+        # 上传文件到OSS
+        resp = self.obsClient.putFile(cfg.obs_bucketName, objectkey, file_path)
+
+        uploadRes.video_path = resp['body']['objectUrl']
+        uploadRes.squenceId = squenceId
+        os.remove(file_path)
+        return uploadRes
+
+    def get_information(self, video_squence, camera_type):
+        """获取视频信息"""
+        videos_path = []
+        videos_dir = os.sep.join([cfg.save_videos_dir, video_squence])
+        for video_name in os.listdir(videos_dir):
+            if video_squence in video_name:
+                if camera_type == video_name.split('_')[-3]:  # 摄像头位置ID
+                    videos_path.append(self.upload(video_name).video_path)
+                elif camera_type == '2':
+                    videos_path.append(self.upload(video_name).video_path)
+        return {"videos_path": videos_path}
+
+    @staticmethod
+    def exception_action(queueImgs):
+        """删除视频"""
+        # 解析视频名称获取相关信息
+        video_squence = queueImgs['videoIds'].split(',')[0]
+        for video_name in os.listdir(cfg.save_videos_dir):
+            if video_squence in video_name:
+                os.rename(os.sep.join([cfg.save_videos_dir, video_name]),
+                          os.sep.join([cfg.save_videos_dir, '03_' + video_name]))
+
+class VideoUploader:
+    @staticmethod
+    def read_config_file(file_path):
+        """安全地读取配置文件"""
+        try:
+            with open(file_path, 'r') as file:
+                lines = file.readlines()
+        except IOError as e:
+            print(f"读取配置文件错误: {e}")
+            return []
+        return [line.strip() for line in lines if line.strip() != '']
+
+    @staticmethod
+    def upload_videos_for_ids(video_ids, video_dir):
+        """根据ID上传视频"""
+        tempdata = []
+        print('----------->', video_dir)
+        for root, dirs, files in os.walk(video_dir):
+            for name in files:
+                print(name)
+                name_s = name.split('.')[0]  # 避免重复分割
+                parts = name_s.split('_')
+                if len(parts) < 7:
+                    continue
+                for data in video_ids:
+                    if parts[-1] == data['action'] and (parts[-3] == data['type'] or data['type'] == '2'):
+                        try:
+                            upload_rs = uploadVideos().upload(name, video_dir)
+                            if upload_rs:
+                                video_path = upload_rs.video_path
+                                sequence_id = upload_rs.squenceId
+                                tempdata.append({
+                                    "squenceId": sequence_id,
+                                    "video_path": [video_path]
+                                })
+                                break  # 找到匹配项即跳出循环
+                        except Exception as e:
+                            print(f"上传视频 {name} 时出错: {e}")
+        return tempdata
+
+
+    @staticmethod
+    def timedUpload(rootPth = '/home/lc/project/ieemoo'):
+        """定时上传视频"""
+        storidPth = os.sep.join([rootPth, 'tools', 'storeId.txt'])
+        save_videos_dir = os.sep.join([rootPth, 'videos'])
+        config_lines = VideoUploader.read_config_file(storidPth)
+        if not config_lines:
+            print("未找到有效配置。")
+
+        print('配置行 --- >', config_lines)
+        soreid_list = [{"storeId": line} for line in config_lines]
+        try:
+            rep = requests.post(url=cfg.get_config_url, data=soreid_list[0])
+            rep.raise_for_status()  # 检查响应状态
+            video_ids = rep.json().get('data', [])
+        except requests.RequestException as e:
+            print(f"获取配置信息失败: {e}")
+
+        if video_ids:
+            tempdata = VideoUploader.upload_videos_for_ids(video_ids, save_videos_dir)
+            if tempdata:
+                tmpdata = {'videosPth': str(tempdata)}
+                try:
+                    requests.post(url=cfg.push_url, data=tmpdata)
+                    print("推送数据成功")
+                except requests.RequestException as e:
+                    print(f"推送数据失败: {e}")
+        else:
+            tmpdata = {'videosPth': str([])}
+            try:
+                requests.post(url=cfg.push_url, data=tmpdata)
+            except requests.RequestException as e:
+                print(f"空数据推送失败: {e}")
+        # print(tmpdata)
+
+if __name__ == '__main__':
+    VideoUploader.timedUpload()