pybaseutils 2.4.34__tar.gz → 2.4.35__tar.gz

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/PKG-INFO

@@ -1,13 +1,19 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: pybaseutils
-Version: 2.4.34
+Version: 2.4.35
 Summary: pybaseutils
 Home-page: https://github.com/PanJinquan/base-utils
 Author: PanJinquan
 Author-email: 390737991@qq.com
 License: MIT
-Platform: UNKNOWN
 License-File: LICENCE
+Dynamic: author
+Dynamic: author-email
+Dynamic: description
+Dynamic: home-page
+Dynamic: license
+Dynamic: license-file
+Dynamic: summary
 
 base-utils
 ==========
@@ -142,5 +148,3 @@ base_utils(C++)
        DEBUG_IMSHOW("image", image);
        return 0;
    }
-
-

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/pybaseutils/__init__.py

@@ -8,4 +8,4 @@
 # --------------------------------------------------------
 """
 
-__version__ = '2.4.34'
+__version__ = '2.4.35'

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/pybaseutils/base64_utils.py

@@ -11,13 +11,16 @@ import sys
 import os
 
 import PIL.Image as Image
+from PIL import ImageFile
 import cv2
 import base64
 import numpy as np
 from typing import Any
+from io import BytesIO
 
 IMG_PREFIX = "image/jpg"  # 图片base64字符串前缀
 precision = 6  # 小数点精度
+ImageFile.LOAD_TRUNCATED_IMAGES = True  # 允许加载截断的图像
 
 
 def isbase64(data: str, prefix=IMG_PREFIX):
@@ -36,10 +39,14 @@ def base642image(bs64, prefix=IMG_PREFIX, use_rgb=False) -> np.ndarray:
     if prefix == bs64[0:len(prefix)]:
         bs64 = bs64[len(prefix):]
     bs64 = bytes(bs64, 'utf-8')
-    image = base64.b64decode(bs64)
-    image = np.frombuffer(image, np.uint8)
-    image = cv2.imdecode(image, flags=cv2.IMREAD_UNCHANGED)
+    imgbf = base64.b64decode(bs64)
+    # image = np.frombuffer(imgbf, np.uint8)
+    # image = cv2.imdecode(image, flags=cv2.IMREAD_UNCHANGED)
     # image = cv2.imdecode(image, cv2.IMREAD_COLOR)
+    # if image is None:
+    image = Image.open(BytesIO(imgbf)) # 修复截断的图像
+    image = np.array(image)
+    image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
     if use_rgb:
         image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
     return image

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/pybaseutils/color_utils.py

@@ -10,6 +10,8 @@
 import cv2
 import matplotlib.pyplot as plt
 import numpy as np
+from collections import defaultdict, Counter
+from pybaseutils import file_utils
 
 VOC_NAMES = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat",
              "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person",
@@ -28,6 +30,30 @@ COCO_NAMES = ['background', 'person', 'bicycle', 'car', 'motorcycle', 'airplane'
               'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors',
               'teddy bear', 'hair drier', 'toothbrush']
 
+color_table = [(0, 0, 0), (0, 255, 0), (255, 0, 0), (0, 0, 255), (0, 255, 255), (255, 0, 255), (255, 255, 0),
+               (128, 0, 0), (0, 128, 0), (128, 128, 0),
+               (0, 0, 128), (128, 0, 128), (0, 128, 128), (128, 128, 128),
+               (64, 0, 0), (192, 0, 0), (64, 128, 0), (192, 128, 0),
+               (64, 0, 128), (192, 0, 128), (64, 128, 128), (192, 128, 128),
+               (0, 64, 0), (128, 64, 0), (0, 192, 0), (128, 192, 0), (0, 64, 128)] * 100
+
+
+class ColorHash():
+    def __init__(self):
+        self.color = dict()
+        self.cnums = len(color_table)
+
+    def get_color(self, label: str):
+        """
+        :param label: str or list
+        :return:
+        """
+        h = file_utils.get_hash(texts=[str(label)])
+        if h not in self.color:
+            i = len(self.color) + 1
+            self.color[h] = color_table[i % self.cnums]
+        return self.color[h]
+
 
 def get_colormap(data_type="custom"):
     if data_type == 'pascal':

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/pybaseutils/file_utils.py

@@ -148,6 +148,25 @@ def get_data_hash(data: np.ndarray):
     return hash
 
 
+def get_hash(image: np.ndarray = None, texts=[]) -> str:
+    """根据image和texts数据生成唯一ID(hash)"""
+    if texts is None: texts = []
+    # 使用更快的哈希算法（如果安全性要求不高）
+    hasher = hashlib.sha256()  # 或 hashlib.md5() for speed
+    # 文本处理
+    if texts:
+        texts = sorted([t.strip() for t in texts if t and t.strip()])
+        if texts: hasher.update('|'.join(texts).encode('utf-8'))
+    # 图像处理
+    if image is not None and isinstance(image, np.ndarray):
+        if not image.flags.c_contiguous:  image = np.ascontiguousarray(image)  # 确保连续性（视图操作）
+        # 使用memoryview避免拷贝
+        hasher.update(memoryview(image))
+    # 如果完全没有数据，返回空哈希
+    v = hasher.hexdigest() if hasher.digest_size > 0 else hashlib.sha256(b'empty').hexdigest()
+    return v
+
+
 def replace_elements(items, src, dst, ignore=True):
     """
     将列表中，值为src改为dst
@@ -264,6 +283,18 @@ def save_json(json_file, json_data):
 write_json_path = save_json
 
 
+def save_jsonl(jsonl_file, data_list: list):
+    """
+    data_list转为jsonl保持
+    :param jsonl_file:
+    :param data_list:
+    :return:
+    """
+    with open(jsonl_file, 'w', encoding='utf-8') as f:
+        for item in data_list:
+            f.write(json.dumps(item, ensure_ascii=False) + '\n')
+
+
 def load_json_files(files: list, max_workers=8):
     """
     根据url读取文件

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/pybaseutils/image_utils.py

@@ -43,6 +43,8 @@ color_map = [(0, 0, 0), (56, 56, 255), (151, 157, 255), (31, 112, 255), (29, 178
              (168, 153, 44), (255, 194, 0), (147, 69, 52), (255, 115, 100),
              (236, 24, 0), (255, 56, 132), (133, 0, 82), (255, 56, 203), (200, 149, 255), (199, 55, 255)] * 100
 
+color_hash = color_utils.ColorHash()
+
 root = os.path.dirname(__file__)
 coco_skeleton = [[1, 3], [1, 0], [2, 4], [2, 0], [0, 5], [0, 6], [5, 7], [7, 9], [6, 8],
                  [8, 10], [5, 11], [6, 12], [11, 12], [11, 13], [13, 15], [12, 14], [14, 16]]
@@ -1167,7 +1169,7 @@ def draw_image_boxes_labels_texts(image, boxes, labels, texts, color=None, thick
     for i, (label, bbox, name) in enumerate(zip(labels, boxes, texts)):
         bbox = [int(b) for b in bbox]
         if color_type == "class":
-            c = color if color else colors[(int(label) + 1) % 20]  # 相同类别相同颜色
+            c = color if color else color_hash.get_color(label)  # 相同类别相同颜色
         else:
             c = color if color else colors[(i + 1) % 20]  # 每个实例不同颜色
         image = draw_image_bbox_text(image, bbox, str(name), color=c, thickness=thickness, fontScale=fontScale,
@@ -1472,13 +1474,12 @@ def draw_text(image, point, text, color=(255, 0, 0), fontScale=0, thickness=0, d
     """
     point = (int(point[0]), int(point[1]))
     thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
-    fontFace = cv2.FONT_HERSHEY_SIMPLEX
-    # fontFace=cv2.FONT_HERSHEY_SIMPLEX
+    fontFace = cv2.FONT_HERSHEY_COMPLEX
     if drawType == "custom" or drawType == "en":
-        text_size, baseline = cv2.getTextSize(str(text), fontFace, fontScale, abs(thickness))
-        text_loc = (point[0], point[1] + text_size[1])
+        text_wh, baseline = cv2.getTextSize(str(text), fontFace, fontScale, abs(thickness))
+        text_loc = (point[0], point[1] + text_wh[1])
         cv2.rectangle(image, (text_loc[0] - 2 // 2, text_loc[1] - 2 - baseline),
-                      (text_loc[0] + text_size[0], text_loc[1] + text_size[1]), color=color, thickness=-1)
+                      (text_loc[0] + text_wh[0], text_loc[1] + text_wh[1]), color=color, thickness=-1)
         # draw score value
         cv2.putText(image, str(text), (text_loc[0], text_loc[1] + baseline), fontFace, fontScale, (255, 255, 255),
                     abs(thickness), 2)
@@ -1489,7 +1490,7 @@ def draw_text(image, point, text, color=(255, 0, 0), fontScale=0, thickness=0, d
     return image
 
 
-def draw_image_bbox_text(image, bbox, text, color, thickness=2, fontScale=0.8, alpha=0, drawType="custom", top=True):
+def draw_image_bbox_text_bk(image, bbox, text, color, thickness=2, fontScale=0.8, alpha=0, drawType="custom", top=True):
     """
     :param image:
     :param bbox:
@@ -1508,32 +1509,100 @@ def draw_image_bbox_text(image, bbox, text, color, thickness=2, fontScale=0.8, a
     thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
     # text_loc = (bbox[0], bbox[1]) if top else (bbox[0], bbox[3])
     if not text: drawType = "simple"
-    text_size, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_SIMPLEX, fontScale, abs(thickness))
+    text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX, fontScale, abs(thickness))
     if top:
-        # text_loc = (bbox[0], bbox[1] + text_size[1])  # 在左上角下方绘制文字
+        # text_loc = (bbox[0], bbox[1] + text_wh[1])  # 在左上角下方绘制文字
         text_loc = (bbox[0], bbox[1] - baseline // 2)  # 在左上角上方绘制文字
     else:
-        text_loc = (bbox[0], bbox[3] + text_size[1] + baseline // 2)
+        text_loc = (bbox[0], bbox[3] + text_wh[1] + baseline // 2)
     if drawType == "chinese" or drawType == "ch":
         if not top: text_loc = (text_loc[0], text_loc[1] - baseline)
         cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness)
         cv2_putText(image, str(text), text_loc, color=color, fontScale=fontScale, thickness=abs(thickness))
     elif drawType == "simple" or drawType == "en":
         cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness, 8, 0)
-        cv2.putText(image, str(text), text_loc, cv2.FONT_HERSHEY_SIMPLEX, fontScale, color, abs(thickness))
+        cv2.putText(image, str(text), text_loc, cv2.FONT_HERSHEY_COMPLEX, fontScale, color, abs(thickness))
     elif drawType == "custom":
+        text_loc = (bbox[0], bbox[1] + text_wh[1] + baseline // 2)  # 在左上角下方绘制文字
+        # text_loc = (bbox[0], bbox[1] + baseline)  # 在左上角上方绘制文字
         cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness)
-        bg1 = (text_loc[0], text_loc[1] - text_size[1] - baseline // 2)
-        bg2 = (text_loc[0] + text_size[0], text_loc[1] + baseline // 2)  # 底纹等于字体的宽度
-        # bg2 = (max(bbox[2], text_loc[0] + text_size[0]), text_loc[1] + baseline // 2) # 底纹等于box的宽度
+        bg1 = (text_loc[0], text_loc[1] - text_wh[1] - baseline // 2)
+        bg2 = (text_loc[0] + text_wh[0], text_loc[1] + baseline // 2)  # 底纹等于字体的宽度
+        # bg2 = (max(bbox[2], text_loc[0] + text_wh[0]), text_loc[1] + baseline // 2) # 底纹等于box的宽度
         cv2.rectangle(image, bg1, bg2, color, thickness)  # 先绘制框，再填充
         cv2.rectangle(image, bg1, bg2, color, -1)
         # draw score value
-        cv2.putText(image, str(text), (text_loc[0], text_loc[1]), cv2.FONT_HERSHEY_SIMPLEX, fontScale,
+        cv2.putText(image, str(text), (text_loc[0], text_loc[1]), cv2.FONT_HERSHEY_COMPLEX, fontScale,
                     (255, 255, 255), abs(thickness))
     return image
 
 
+def draw_image_bbox_text(image, bbox, text, color, thickness=2, fontScale=0.8, alpha=0, drawType="custom", top=True):
+    """
+    :param image:
+    :param bbox:
+    :param color:
+    :param text:
+    :param drawType:
+    :param top:
+    :return:
+    """
+    text = str(text)
+    if alpha > 0:
+        bgimg = draw_image_bbox_text(image.copy(), bbox, text=text, color=color, thickness=thickness,
+                                     fontScale=fontScale, drawType=drawType, top=top, alpha=0)
+        image[:] = cv2.addWeighted(bgimg, 1 - alpha, image, alpha, 0)
+        return image
+    if drawType == "ch" or drawType == "chinese":
+        image = draw_ch_text(image, bbox, text, color, thickness=thickness, fontScale=fontScale)
+    elif drawType == "simple":
+        text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX, fontScale, abs(thickness))
+        text_loc = (bbox[0], bbox[1] - baseline // 2)  # 在左上角上方绘制文字
+        cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness, 8, 0)
+        cv2.putText(image, str(text), text_loc, cv2.FONT_HERSHEY_COMPLEX, fontScale, color, abs(thickness))
+    else:  # custom ,en
+        image = draw_en_text(image, bbox, text, color, thickness=thickness, fontScale=fontScale)
+    return image
+
+
+def draw_ch_text(image, bbox, text, color, thickness=2, fontScale=0.8):
+    thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
+    text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX, fontScale, abs(thickness))
+    text_wh = (text_wh[0], int(text_wh[1] * 0.7))  # 与draw_en_text区别，底纹缩小80%
+    text_loc = (bbox[0], bbox[1] - baseline // 2)  # 在左上角上方绘制文字
+    text_bgy = text_loc[1] - text_wh[1] - baseline // 2
+    if text_bgy < 0:  # 如果文字超出图片范围，切换到下方绘制
+        text_loc = (bbox[0], bbox[1] + text_wh[1] + baseline // 2)  # 在左上角下方绘制文字
+    cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness)
+    bg1 = (text_loc[0], text_loc[1] - text_wh[1] - baseline // 2)
+    bg2 = (text_loc[0] + text_wh[0], text_loc[1] + baseline // 2)  # 底纹等于字体的宽度
+    # bg2 = (max(bbox[2], text_loc[0] + text_wh[0]), text_loc[1] + baseline // 2) # 底纹等于box的宽度
+    cv2.rectangle(image, bg1, bg2, color, thickness)  # 先绘制框，再填充
+    cv2.rectangle(image, bg1, bg2, color, -1)
+    # draw score value
+    cv2_putText(image, str(text), text_loc, color=(255, 255, 255), fontScale=fontScale, thickness=abs(thickness))
+    return image
+
+
+def draw_en_text(image, bbox, text, color, thickness=2, fontScale=0.8):
+    thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
+    text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX, fontScale, abs(thickness))
+    text_loc = (bbox[0], bbox[1] - baseline // 2)  # 在左上角上方绘制文字
+    text_bgy = text_loc[1] - text_wh[1] - baseline // 2
+    if text_bgy < 0:  # 如果文字超出图片范围，切换到下方绘制
+        text_loc = (bbox[0], bbox[1] + text_wh[1] + baseline // 2)  # 在左上角下方绘制文字
+    cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness)
+    bg1 = (text_loc[0], text_loc[1] - text_wh[1] - baseline // 2)
+    bg2 = (text_loc[0] + text_wh[0], text_loc[1] + baseline // 2)  # 底纹等于字体的宽度
+    # bg2 = (max(bbox[2], text_loc[0] + text_wh[0]), text_loc[1] + baseline // 2) # 底纹等于box的宽度
+    cv2.rectangle(image, bg1, bg2, color, thickness)  # 先绘制框，再填充
+    cv2.rectangle(image, bg1, bg2, color, -1)
+    # draw score value
+    cv2.putText(image, str(text), (text_loc[0], text_loc[1]), fontFace=cv2.FONT_HERSHEY_COMPLEX, fontScale=fontScale,
+                color=(255, 255, 255), thickness=abs(thickness))
+    return image
+
+
 def draw_text_line(image, point, text_line: str, bg_color=(255, 0, 0), thickness=0, fontScale=0, drawType="custom"):
     """
     :param image:
@@ -1543,14 +1612,13 @@ def draw_text_line(image, point, text_line: str, bg_color=(255, 0, 0), thickness
     :return:
     """
     thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
-    fontFace = cv2.FONT_HERSHEY_SIMPLEX
-    # fontFace=cv2.FONT_HERSHEY_SIMPLEX
+    fontFace = cv2.FONT_HERSHEY_COMPLEX
     text_line = text_line.split("\n")
-    # text_size, baseline = cv2.getTextSize(str(text_line), fontFace, fontScale, thickness)
-    text_size, baseline = cv2.getTextSize(str(text_line), fontFace, fontScale, thickness)
+    # text_wh, baseline = cv2.getTextSize(str(text_line), fontFace, fontScale, thickness)
+    text_wh, baseline = cv2.getTextSize(str(text_line), fontFace, fontScale, thickness)
     for i, text in enumerate(text_line):
         if text:
-            draw_point = (point[0], point[1] + (text_size[1] + 2 + baseline) * i)
+            draw_point = (point[0], point[1] + (text_wh[1] + 2 + baseline) * i)
             image = draw_text(image, draw_point, text, bg_color, thickness=thickness, drawType=drawType)
     return image
 
@@ -1599,11 +1667,11 @@ def get_font_type(size, font=""):
 
 def cv2_putText(img, text, point, fontFace=None, fontScale=0.8, color=(255, 0, 0), thickness=None):
     # cv2.putText(img, str(text), point, fontFace, fontScale, color=color, thickness=thickness)
-    text_size, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_SIMPLEX, fontScale, abs(thickness))
+    text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX, fontScale, abs(thickness))
     pilimg = Image.fromarray(img)  # Image.fromarray()将数组类型转成图片格式，与np.array()相反
     draw = ImageDraw.Draw(pilimg)  # PIL图片上打印汉字
-    size = text_size[1]  # 字体大小
-    point = (point[0], point[1] - text_size[1] + baseline // 2)
+    size = text_wh[1]  # 字体大小
+    point = (point[0], point[1] - text_wh[1] + baseline // 2)
     font = get_font_type(size=size)
     draw.text(point, text, color, font)
     img[:] = np.asarray(pilimg)
@@ -2333,7 +2401,7 @@ def draw_yaw_pitch_roll_in_right_axis(image,
     cv2.arrowedLine(image, (int(cx), int(cy)), (int(x2), int(y2)), color_pitch_y, 2, tipLength=0.2)
     cv2.arrowedLine(image, (int(cx), int(cy)), (int(x3), int(y3)), color_roll_z, 2, tipLength=0.2)
     if vis:
-        fontFace = cv2.FONT_HERSHEY_SIMPLEX
+        fontFace = cv2.FONT_HERSHEY_COMPLEX
         image = cv2.putText(image, str(text),
                             # (int(cx), int(cy) + 10),
                             (10, 10),
@@ -2379,7 +2447,7 @@ def draw_yaw_pitch_roll_in_left_axis(image,
     cv2.arrowedLine(image, (int(cx), int(cy)), (int(x2), int(y2)), color_pitch_y, 2, tipLength=0.2)
     cv2.arrowedLine(image, (int(cx), int(cy)), (int(x3), int(y3)), color_roll_z, 2, tipLength=0.2)
     if vis:
-        fontFace = cv2.FONT_HERSHEY_SIMPLEX
+        fontFace = cv2.FONT_HERSHEY_COMPLEX
         image = cv2.putText(image, str(text),
                             # (int(cx), int(cy) + 10),
                             (10, 10),
@@ -2715,6 +2783,34 @@ def draw_image_contours(image, contours: List[np.ndarray], texts=[], color=(), a
     return image
 
 
+def draw_image_contours_labels_texts(image, contours: List[np.ndarray], labels=[], texts=[], color=(), alpha=0.5,
+                                     thickness=1, fontScale=0.8, drawType="ch", colors=None):
+    """
+    参考：draw_image_mask_color
+    :param image:
+    :param contours: List[np.ndarray],每个列表是一个轮廓(num_points,1,2)
+    :param texts:轮廓文本
+    :param color:绘制轮廓的颜色
+    :param alpha:绘制颜色的透明度 0: 轮廓填充完全透明，1: 轮廓填充完全不透明
+    :param thickness:轮廓线宽
+    :return:
+    """
+    colors = [(0, 0, 0)] + colors if colors else color_table
+    for i in range(0, len(contours)):
+        c = color if color else color_hash.get_color(labels[i])  # 相同类别相同颜色
+        t = str(texts[i]) if texts else ""
+        p = np.asarray(contours[i], dtype=np.int32)
+        b = (min(p[:, 0]), min(p[:, 1]), max(p[:, 0]), max(p[:, 1]))
+        if len(p.shape) == 2: p = [p]
+        image[:] = cv2.drawContours(image, p, contourIdx=-1, color=c, thickness=thickness)
+        bgimg = image.copy()
+        bgimg = cv2.fillPoly(bgimg, p, color=c)
+        image[:] = cv2.addWeighted(src1=image, alpha=1 - alpha, src2=bgimg, beta=alpha, gamma=0)
+        if t: image[:] = draw_text(image, point=(b[0], b[1]), color=c, text=t, thickness=thickness,
+                                   fontScale=fontScale, drawType=drawType)
+    return image
+
+
 draw_image_mask_color = color_utils.draw_image_mask_color
 
 

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/pybaseutils/metrics/class_report.py

@@ -85,7 +85,7 @@ def get_confusion_matrix(true_labels, pred_labels, target_names=None, filename=N
 
 
 def get_metrics_report(true_labels, pred_labels, target_names=None, labels=None,
-                       output_dict=False, matrix=False, plot=False):
+                       output_dict=False, output_csv="", matrix=False, plot=False):
     """
     Usage
     -------
@@ -108,6 +108,7 @@ def get_metrics_report(true_labels, pred_labels, target_names=None, labels=None,
     :param target_names:
     :param labels:
     :param output_dict:
+    :param output_csv: 'output/data.csv'
     :param matrix: 是否绘制混淆矩阵
     :return:
     """
@@ -132,6 +133,10 @@ def get_metrics_report(true_labels, pred_labels, target_names=None, labels=None,
         result["accuracy"] = {'precision': p, 'recall': r, 'f1-score': result["accuracy"], 'support': support}
         output = pandas_utils.dict2df(result)
         output = output.round(4)  # 保留4位小数
+        if output_csv:
+            output_csv = output_csv.replace(".csv", "-f1s{:3.4f}.csv".format(result["accuracy"]["f1-score"]))
+            output.to_csv(output_csv, index=True)
+        # TODO: 若发现表格中英文字体显示没有对齐，则需要安装wcwidth库: pip install --upgrade tabulate wcwidth
         output = output.to_markdown()
     if matrix:
         conf_matrix = get_confusion_matrix(true_labels, pred_labels, target_names=target_names,
@@ -176,10 +181,10 @@ def create_file_path(filename):
 if __name__ == "__main__":
     # true_labels = [0, 1, 2, 3, 3, 1, 1]  # Y
     # pred_labels = [1, 1, 2, 2, 2, 1, 0]  # X
-    true_labels = ["A", "B", "A", "B", "C", "D"]
-    pred_labels = ["A", "D", "C", "B", "C", "A"]
+    true_labels = ["A手", "B机器", "A机器人", "B", "C", "D"]
+    pred_labels = ["A", "D", "C", "B", "C", "A你好"]
     target_names = None
-    confuse_file = "./confuse.csv"
+    output_excel = "./result.csv"
     result = get_metrics_report(true_labels, pred_labels, target_names=target_names, output_dict=False,
-                                matrix=True, plot=False)
+                                output_csv=output_excel, plot=False)
     print(result)

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/pybaseutils.egg-info/PKG-INFO

@@ -1,13 +1,19 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: pybaseutils
-Version: 2.4.34
+Version: 2.4.35
 Summary: pybaseutils
 Home-page: https://github.com/PanJinquan/base-utils
 Author: PanJinquan
 Author-email: 390737991@qq.com
 License: MIT
-Platform: UNKNOWN
 License-File: LICENCE
+Dynamic: author
+Dynamic: author-email
+Dynamic: description
+Dynamic: home-page
+Dynamic: license
+Dynamic: license-file
+Dynamic: summary
 
 base-utils
 ==========
@@ -142,5 +148,3 @@ base_utils(C++)
        DEBUG_IMSHOW("image", image);
        return 0;
    }
-
-

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/setup.py

@@ -9,15 +9,34 @@
 """
 import os
 import pybaseutils
-import pypandoc
 from setuptools import setup, find_packages
 
 root = os.path.dirname(__file__)
-long_description = pypandoc.convert_file('README.md', 'rst')
+
+
+def get_description():
+    """
+    安全读取 README 作为 long_description。
+    优先尝试用 pypandoc 转换为 rst；若不可用则回退到原始 markdown。
+    """
+    mdfile = os.path.join(root, 'README.md')
+    try:  # 优先尝试 pypandoc 转换
+        import pypandoc
+        return pypandoc.convert_file(mdfile, 'rst')
+    except (ImportError, OSError):
+        pass
+    try:
+        # 回退：直接读取原始 markdown
+        with open(mdfile, 'r', encoding='utf-8') as f:
+            return f.read()
+    except FileNotFoundError:
+        return ''
+
+
 setup(name='pybaseutils',
       version=pybaseutils.__version__,
       description='pybaseutils',
-      long_description=long_description,
+      long_description=get_description(),
       url='https://github.com/PanJinquan/base-utils',
       author='PanJinquan',
       author_email='390737991@qq.com',

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/test_py/camera/demo.py

@@ -81,7 +81,8 @@ class CameraDemo():
 
     def capture(self, title="camera", save_video="./camera.avi"):
         format = '%Y-%m-%dT%H:%M:%S'  # 海康摄像头默认格式
-        video_url = f"rtsp://{self.username}:{self.password}@{self.ip}/Streaming/Channels/1"
+        # video_url = f"rtsp://{self.username}:{self.password}@{self.ip}/Streaming/Channels/1"
+        video_url = f"rtsp://{self.username}:{self.password}@{self.ip}:554"
         print(f"正在连接摄像头: {video_url}")
         # 创建视频流连接
         cap = cv2.VideoCapture(video_url)
@@ -140,8 +141,9 @@ class CameraDemo():
 
 if __name__ == "__main__":
     # 摄像头信息，按'q'键退出程序
-    ip = "192.168.2.35"
+    #  rtsp://admin:zhgp@2024@192.168.28.61
+    ip = "192.168.28.61"
     username = "admin"
-    password = "C2332416"
+    password = "zhgp@2024"
     c = CameraDemo(ip, username, password)
     c.capture()

pybaseutils-2.4.35/test_py/demo3.py

@@ -0,0 +1,21 @@
+# -*- coding: utf-8 -*-
+"""
+# --------------------------------------------------------
+# @Author : Pan
+# @E-mail :
+# @Date   : 2025-07-08 14:10:15
+# @Brief  :
+# --------------------------------------------------------
+"""
+import cv2
+import numpy as np
+from pybaseutils import file_utils, image_utils
+
+import cv2
+
+data="ABCDabcd你好我是一名学生1234567890"
+fontScale = 1.0
+thickness = 1
+for text in data:
+    text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX	, fontScale, abs(thickness))
+    print(f"fontScale={fontScale}, thickness={thickness}, text={text}, text_wh={text_wh}, baseline={baseline}")

pybaseutils-2.4.35/test_py/demo4.py

@@ -0,0 +1,30 @@
+# -*- coding: utf-8 -*-
+"""
+# --------------------------------------------------------
+# @Author : Pan
+# @E-mail :
+# @Date   : 2025-07-08 14:10:15
+# @Brief  : 转换labelme标注数据为voc格式
+# --------------------------------------------------------
+"""
+import os
+import numpy as np
+import cv2
+from pybaseutils.converter import convert_labelme2voc
+from pybaseutils import time_utils, image_utils, file_utils, json_utils
+fontScale=2.0
+alpha = 0.
+if __name__ == "__main__":
+    image_file = "../data/test_image/grid1.png"
+
+    for i in range(100):
+        image = cv2.imread(image_file)
+        boxes = [[0, i, 500, 300]]
+        # texts = ["ABC"] * len(boxes)
+        texts = ["ABC你是一名学生"] * len(boxes)
+        print(boxes)
+        image = image_utils.draw_image_boxes_texts(image, boxes, texts, thickness=1, fontScale=fontScale, alpha=alpha,drawType="simple")
+        image_utils.show_image("ch", image)
+        image = cv2.imread(image_file)
+        image = image_utils.draw_image_boxes_texts(image, boxes, texts, thickness=1, fontScale=fontScale, alpha=alpha,drawType="en")
+        image_utils.show_image("en", image)

{pybaseutils-2.4.34 → pybaseutils-2.4.35}/test_py/demo_gif_video.py

@@ -39,7 +39,9 @@ def android_gif():
 
 def python_gif():
     video_file = "/home/PKing/nasdata/Project/LLM/MLLM-Factory/docs/chat-demo.mp4"
-    video_utils.video2gif(video_file, interval=10, func=resize_fun, fps=4, time=(0.5, 24), use_pil=True, vis=True)
+    video_file = "/media/PKing/dev1/project/AI-Labelme/docs/video-track-label.mp4"
+    # video_utils.video2gif(video_file, interval=10, func=resize_fun, fps=4, time=(0.5, 24), use_pil=True, vis=True)
+    video_utils.video2gif(video_file, interval=8, func=resize_fun, fps=4,  use_pil=True, vis=True)
 
 
 def image_gif():