pybaseutils 2.4.9__tar.gz → 2.4.11__tar.gz

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pybaseutils
-Version: 2.4.9
+Version: 2.4.11
 Summary: pybaseutils
 Home-page: https://github.com/PanJinquan/base-utils
 Author: PanJinquan

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils/__init__.py

@@ -8,4 +8,4 @@
 # --------------------------------------------------------
 """
 
-__version__ = '2.4.9'
+__version__ = '2.4.11'

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils/build_utils/pyarmor_utils.py

@@ -38,7 +38,7 @@ def build_pyarmor_project(root,
     :param exclude_dirs: 不需要处理的文件
     :return:
     """
-    app = os.path.dirname(entry)
+    dir = os.path.dirname(entry)
     file_utils.copy_dir(root, build, exclude=exclude_dirs)
     os.system(f'pyarmor --version')
     os.system(f'rm -rf .pyarmor_config')
@@ -47,9 +47,16 @@ def build_pyarmor_project(root,
     if manifest:
         os.system(f'pyarmor config --manifest="{manifest}"')
     os.system(f'pyarmor build --output={build} --force')
-    # os.system(f'cd {build} && cp -r pytransform {os.path.join(build, app)}')
-    file_utils.move_dir(os.path.join(build, "pytransform"), os.path.join(build, app, "pytransform"))
-    file_utils.remove_dir(os.path.join(build, "pytransform"))
+    src = os.path.join(build, "pytransform")
+    dst = os.path.join(build, dir)
+    if src != os.path.join(dst, "pytransform"):
+        cmd = f"cp -rf {src} {dst}"
+        print(cmd)
+        os.system(cmd)
+        os.system(f"rm -rf {src}")
+    # os.system(f'cd {build} && cp -r pytransform {os.path.join(build, dir)}')
+    # file_utils.move_dir(os.path.join(build, "pytransform"), os.path.join(build, dir, "pytransform"))
+    # file_utils.remove_dir(os.path.join(build, "pytransform"))
     # 将pyarmor编译结果保存到
     # file_utils.copy_dir(os.path.join(root, "dist"), build, exclude=[])
 

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils/converter/build_labelme.py

@@ -13,7 +13,7 @@ from pybaseutils import image_utils, file_utils, json_utils
 
 
 def save_labelme(out_root, image_file, points, names, class_dict={}, group=None, image=None, prefix="",
-                 index=0, vis=False, delay=0):
+                 index=0, shape_type="polygon", vis=False, delay=0):
     """
     :param out_root:   输出根目录
     :param image_file: 图片路径
@@ -24,6 +24,7 @@ def save_labelme(out_root, image_file, points, names, class_dict={}, group=None,
     :param image:  图像
     :param prefix: 前缀，如果提供，则重新命名
     :param index:  提供前缀需要重新名称
+    :param shape_type:  polygon,rotation
     :return:
     """
     if class_dict: names = [class_dict.get(n, n) for n in names]
@@ -43,7 +44,7 @@ def save_labelme(out_root, image_file, points, names, class_dict={}, group=None,
     image_id, postfix = file_utils.split_postfix(image_name)
     json_file = file_utils.create_dir(out_root, "images", f"{image_id}.json")
     file_path = file_utils.create_dir(out_root, "images", f"{image_name}")
-    maker_labelme(json_file, points, names, image_name, group=group, image_size=(w, h), image_bs64=None)
+    maker_labelme(json_file, points, names, image_name, group=group, image_size=(w, h), image_bs64=None,shape_type=shape_type)
     if isinstance(image, np.ndarray):
         cv2.imwrite(file_path, image)
     else:
@@ -52,7 +53,8 @@ def save_labelme(out_root, image_file, points, names, class_dict={}, group=None,
         print("points is empty,file={}".format(file_path))
 
 
-def maker_labelme(json_file, points, labels, image_name, image_size, group=None, image_bs64=None, keypoints=[]):
+def maker_labelme(json_file, points, labels, image_name, image_size, group=None, image_bs64=None, keypoints=[],
+                  shape_type="polygon"):
     """
     制作label数据格式
     :param json_file: 保存json文件路径
@@ -76,7 +78,7 @@ def maker_labelme(json_file, points, labels, image_name, image_size, group=None,
         if not isinstance(point[0], list): point = [point]
         kpts = keypoints[i] if keypoints else []
         item = {"label": label, "score": None, "keypoints": kpts, "line_color": None, "fill_color": None,
-                "group_id": group[i], "points": point, "shape_type": "polygon", "flags": {}, "description": ""}
+                "group_id": group[i], "points": point, "shape_type": shape_type, "flags": {}, "description": ""}
         shapes.append(item)
     data = {
         "version": "3.16.7", "flags": {},

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils/converter/convert_voc2labelme.py

@@ -23,6 +23,7 @@ def convert_voc2labelme(filename=None,
                         class_dict={},
                         prefix="",
                         max_num=-1,
+                        shape_type='polygon',
                         vis=True):
     """
     将voc格式转换为labelme格式
@@ -52,7 +53,8 @@ def convert_voc2labelme(filename=None,
         image_file = data_info["image_file"]
         points = image_utils.boxes2polygons(boxes)
         build_labelme.save_labelme(out_root, image_file=image_file, points=points, names=names,
-                                   class_dict=class_dict, image=image, prefix=prefix, index=i, vis=vis)
+                                   class_dict=class_dict, image=image, prefix=prefix, index=i, shape_type=shape_type,
+                                   vis=vis)
 
 
 if __name__ == "__main__":

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils/cvutils/camera.py

@@ -8,7 +8,7 @@ class CameraCapture(object):
     def __init__(self, video: str or int = 0, size=(1920, 1080), fps=30):
         """
         :param video: 视频设备路径或索引（如 0 或 "/dev/video0"）
-        :param size: 视频分辨率 (宽, 高)
+        :param size: 视频分辨率 (宽, 高)，(1280,720),(1920,1080)
         :param fps: 视频帧率
         """
         self.size = size
@@ -47,9 +47,9 @@ class CameraCapture(object):
         if len(buf) != self.size[0] * self.size[1] * 3:
             self.stop()
             return False, None  # 读取失败或结束
-        # 将字节流转换为 numpy 数组
-        frame = np.frombuffer(buf, dtype=np.uint8).reshape((self.size[1], self.size[0], 3))
-        return True, frame
+        # TODO 将字节流转换为numpy数组图像(bgr)
+        bgr = np.frombuffer(buf, dtype=np.uint8).reshape((self.size[1], self.size[0], 3))
+        return True, bgr
 
     def release(self):
         self.stop()
@@ -66,7 +66,7 @@ class CameraCapture(object):
             self.pipe.communicate()  # 再次尝试清理
             print(err)
 
-    def display(self):
+    def display(self, title="camera", delay=30):
         while True:
             t1 = time.time()
             ret, frame = self.read()
@@ -75,9 +75,9 @@ class CameraCapture(object):
                 break
             t21 = (t2 - t1) * 1000
             print(f"image shape: {frame.shape},耗时: {t21:.3f}ms")
-            cv2.imshow('FFmpeg Low Latency', frame)
-            if cv2.waitKey(30) & 0xFF == ord('q'):
-                break
+            cv2.namedWindow(title, flags=cv2.WINDOW_NORMAL)
+            cv2.imshow(title, frame)
+            cv2.waitKey(delay)
 
 
 if __name__ == '__main__':

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils/cvutils/video_utils.py

@@ -36,7 +36,8 @@ def get_usb_camera(video=-1, max_index=50):
     camera = -1
     for i in range(max_index):
         cap = cv2.VideoCapture(i)
-        if cap.isOpened():
+        ret, frame = cap.read()
+        if cap.isOpened() and ret:
             camera = i
             cap.release()
             break

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils/image_utils.py

@@ -1772,7 +1772,7 @@ def draw_image_arrowed_lines(image,
     return image
 
 
-def draw_image_polylines(image, points, color=(0, 0, 255)):
+def draw_image_polylines(image, points, color=(0, 0, 255), closed=True, thickness=2):
     """
     # points是三维坐标，分别表示(多边形个数，多边形坐标点x,多边形坐标点y)=(num_polylines,num_point,2)
     points = np.asarray([[[100, 100], [200, 100], [400, 200]],
@@ -1784,8 +1784,8 @@ def draw_image_polylines(image, points, color=(0, 0, 255)):
     参数4 color：必选参数。用于设置多边形的颜色
     参数5 lineType：可选参数。用于设置线段的类型，可选8（8邻接连接线-默认）、4（4邻接连接线）和cv2.LINE_AA 为抗锯齿
     """
-    points = np.asarray(points, dtype=np.int32)
-    image = cv2.polylines(image, points, isClosed=True, color=color, thickness=2)  # 画任意多边形
+    pts = [np.array(p, dtype=np.int32) for p in points]
+    image = cv2.polylines(image, pts, isClosed=closed, color=color, thickness=thickness)  # 画任意多边形
     return image
 
 

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils/log.py

@@ -13,11 +13,11 @@ from loguru import logger
 
 LOG_FORMAT = {
     "simple":   "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}| {message}</level>",
-    "name":     "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{name} {line:4}| {message}</level>",  # 打印文件名
-    "module":   "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{module} {line:4}| {message}</level>",  # 打印模块名
-    # "function": "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{function} {line:4}| {message}</level>",  # 打印函数
-    "function": "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{module}.{function} {line:4}| {message}</level>",  # 打印函数
-    "all":      "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{name}.{module}.{function} {line:4}| {message}</level>",  # 打印函数
+    "name":     "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{name} {line}| {message}</level>",  # 打印文件名
+    "module":   "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{module} {line}| {message}</level>",  # 打印模块名
+    # "function": "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{function} {line}| {message}</level>",  # 打印函数
+    "function": "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{module}.{function} {line}| {message}</level>",  # 打印函数
+    "all":      "<level>{time:YYYY-MM-DD HH:mm:ss}|{level:7}|{name}.{module}.{function} {line}| {message}</level>",  # 打印函数
 }
 
 

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/pybaseutils.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pybaseutils
-Version: 2.4.9
+Version: 2.4.11
 Summary: pybaseutils
 Home-page: https://github.com/PanJinquan/base-utils
 Author: PanJinquan

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/test_py/aije/demo_labelme_crop_ailt.py

@@ -108,11 +108,10 @@ if __name__ == '__main__':
                    ]
     # '*长条标签'
     datasets = [
-        '/home/PKing/nasdata/dataset-dmai/AILT/ailt-det/dataset-20260120/images',
-        '/home/PKing/nasdata/dataset-dmai/AILT/ailt-det/dataset-20260130/images',
-        '/home/PKing/nasdata/dataset-dmai/AILT/ailt-det/dataset-20260206/images',
+        '/home/PKing/nasdata/dataset-dmai/AILT/ailt-det/dataset-20260227/images',
+        '/home/PKing/nasdata/dataset-dmai/AILT/ailt-det/dataset-20260211/images',
     ]
-    datasets = "/home/PKing/nasdata/dataset-dmai/AILT/ailt-det/文件列表.txt"
+    # datasets = "/home/PKing/nasdata/dataset-dmai/AILT/ailt-det/文件列表.txt"
     datasets = file_utils.read_data(datasets, split=None)
     # dataroot = '/home/PKing/nasdata/dataset-dmai/AIJE/dataset/aije-v2-det'
     # datasets = get_sub_dataset(dataroot)

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/test_py/converter/convert_voc2labelme.py

@@ -19,8 +19,8 @@ if __name__ == "__main__":
     Sedan     # 轿车
     Truck     # 卡车
     """
-    data_root = "/home/PKing/nasdata/tmp/tmp/car-det/dataset-v3"
-    out_root = data_root + "/labelme"
+    data_root = "/home/PKing/nasdata/tmp/tmp/pen/dataset-pen-finger-tip-005/train"
+    out_root = data_root
     class_dict = {"car": "car",
                   "person": "pedestrian",
                   "bicycle": "bike",
@@ -30,7 +30,8 @@ if __name__ == "__main__":
                   }
     convert_voc2labelme.convert_voc2labelme(data_root=data_root,
                                             out_root=out_root,
-                                            prefix="image2025",
+                                            prefix="",
                                             class_name=None,
-                                            class_dict=class_dict,
+                                            class_dict={},
+                                            shape_type="rotation",
                                             vis=False)

pybaseutils-2.4.11/test_py/demo1.py

@@ -0,0 +1,43 @@
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+
+class VariationalAutoencoder(nn.Module):
+    def __init__(self, input_dim=784, hidden_dim=128, latent_dim=20):
+        super().__init__()
+        self.latent_dim = latent_dim
+        # TODO 编码器: 输出分布参数
+        self.encoder = nn.Sequential(nn.Linear(input_dim, hidden_dim), nn.ReLU())
+        self.fc_mu_ = nn.Linear(hidden_dim, latent_dim)  # 均值
+        self.fc_var = nn.Linear(hidden_dim, latent_dim)  # 对数方差
+        # TODO 解码器
+        self.decoder = nn.Sequential(nn.Linear(latent_dim, hidden_dim), nn.ReLU(), nn.Linear(hidden_dim, input_dim),nn.Sigmoid())
+
+    def encode(self, x):
+        h = self.encoder(x) # VAE编码器：将输入x映射到潜在空间的分布参数（均值mu和对数方差log_var）
+        return self.fc_mu_(h), self.fc_var(h)
+
+    def decode(self, z):
+        return self.decoder(z)
+
+    def reparameterize(self, mu, log_var):
+        # 重参数化技巧：从N(0,1)采样，然后变换
+        std = torch.exp(0.5 * log_var)
+        eps = torch.randn_like(std)
+        return mu + eps * std
+
+    def forward(self, x):
+        mu, log_var = self.encode(x)  # 输出分布参数
+        z = self.reparameterize(mu, log_var)  # 采样
+        x_recon = self.decode(z)  # 解码重建
+        return x_recon, mu, log_var
+
+    def loss_function(self, x_recon, x, mu, log_var):
+        # 1. 重建损失
+        recon_loss = F.binary_cross_entropy(x_recon, x, reduction='sum')
+        # 2. KL散度：让后验分布接近标准正态分布
+        # KL(N(mu, sigma^2) || N(0, 1)) = -0.5 * (1 + log(sigma^2) - mu^2 - sigma^2)
+        kl_loss = -0.5 * torch.sum(1 + log_var - mu.pow(2) - log_var.exp())
+        total_loss = recon_loss + kl_loss # 总损失
+        return total_loss, recon_loss, kl_loss

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/test_py/demo2.py

@@ -61,9 +61,6 @@ def video_iterator(video):
 
 
 if __name__ == '__main__':
-    from pybaseutils.cvutils import video_utils
-
-    t= datetime.now().strftime("%Y%m%d_%H%M%S_%f")
-    video = "/home/PKing/Videos/人体检测.mp4"
-    video = 0
-    video_capture(video)
+    x = 123
+    x = None
+    print(f"{str(x):7}")

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/test_py/demo_video2frame.py

@@ -19,11 +19,12 @@ def video2frames_demo(root, out, prefix="", vis=False):
     for video_file in files:
         print(video_file)
         # video_utils.video2frames_similarity(video_file, out_dir=out, func=None, interval=20, thresh=0.3, vis=True)
-        video_utils.video2frames(video_file, out_dir=out, task=None, interval=10, vis=vis, prefix=prefix)
+        video_utils.video2frames(video_file, out_dir=out, task=None, interval=5, vis=vis, prefix=prefix)
 
 
 if __name__ == "__main__":
     root = "/home/PKing/edudata/dataset/AIJE/【TOP】技能人才系统_数据集管理/06-联通岗评/02-训练数据/联通入户布线_20260130_180400（补充打胶专题）/20260130_180400.mp4"
+    root = "/media/PKing/dev1/project/ai-cv-service/data/video1.avi"
     out = os.path.join(os.path.dirname(root), "frames")
     prefix = "image"
     # out = os.path.join(root, "frames")

{pybaseutils-2.4.9 → pybaseutils-2.4.11}/test_py/test_camera.py

@@ -2,6 +2,7 @@ import numpy as np
 import cv2
 import time
 from pybaseutils.cvutils.camera import CameraCapture
+from pybaseutils.cvutils.video_utils import get_usb_camera
 
 
 def display(cap: CameraCapture):
@@ -19,10 +20,11 @@ def display(cap: CameraCapture):
 
 
 if __name__ == '__main__':
-    fps = 5
+    fps = 15
     width = 1920
     height = 1080
     # video = 0  # Windows 下可能是 0 或 "video=Integrated Webcam"
     video = "/home/PKing/Videos/demo-src.mp4"  # Windows 下可能是 0 或 "video=Integrated Webcam"
+    video = get_usb_camera()
     cap = CameraCapture(video=video, size=(width, height), fps=fps)
     display(cap)

pybaseutils-2.4.9/test_py/demo1.py

@@ -1,46 +0,0 @@
-# -*- coding: utf-8 -*-
-import os
-import sys
-
-import numpy as np
-import pandas as pd
-
-
-class AppState:
-    def __init__(self, ):
-        self.code = 1
-        self.db_cl_names = []
-        self._cl_name = self.db_cl_names[0] if self.db_cl_names else None
-
-    @property
-    def cl_name(self):
-        if not self._cl_name:
-            self._cl_name = self.db_cl_names[0] if self.db_cl_names else None
-        return self._cl_name
-
-    @cl_name.setter
-    def cl_name(self, v):
-        assert v in self.db_cl_names, f'cl_name {v} not in db_cl_names {self.db_cl_names}'
-
-        self._cl_name = v
-
-
-def is_sorted_numpy(arr: np.ndarray) -> bool:
-    """
-    使用 numpy 判断数组是否为升序。
-    对于大型数值数组，这是最快的方法。
-    注意：此函数输入应为 numpy 数组。
-    """
-    # np.diff 计算相邻元素之差，如果升序，则所有差值都 >= 0
-    # np.all() 检查所有条件是否为真
-    return np.all(np.diff(arr) >= 0)
-
-
-if __name__ == '__main__':
-    import numpy as np
-
-    # --- 示例 ---
-    # 假设 data 是一个 numpy 数组
-    np_data = np.array([1, 2, 2, 3, 5, 8, 0])
-    result = is_sorted_numpy(np_data)
-    print(f"Numpy 数组 {np_data} 是否为升序？ {result}")  # 输出: True