python yaml

大海

增强现实，深度学习，目标检测，位姿估计

个人学习总结，持续更新中…….

斜体代表个人的观点或想法。

YAML 入门教程 | 菜鸟教程

基本语法

大小写敏感
使用缩进表示层级关系
缩进不允许使用tab，只允许空格
缩进的空格数不重要，只要相同层级的元素左对齐即可
'#'表示注释

数据类型

YAML 支持以下几种数据类型：

对象：键值对的集合，又称为映射（mapping）/ 哈希（hashes） / 字典（dictionary）
数组：一组按次序排列的值，又称为序列（sequence） / 列表（list）
纯量（scalars）：单个的、不可再分的值

示例1 读取 .yml

gt.yml

0:
- cam_R_m2c: [0.6, 0.7, 0.2]
  obj_id: 19
- cam_R_m2c: [1.6, 0.7, 0.2]
  obj_id: 20
- cam_R_m2c: [3.6, 0.7, 0.2]
  obj_id: 17
- cam_R_m2c: [4.6, 0.7, 0.2]
  obj_id: 18

test.py

import yaml
gt = 'gt.yml'
with open(gt) as f:
    opt_dict = yaml.load(f, Loader=yaml.FullLoader)
print(opt_dict)
# {0: [{'cam_R_m2c': [0.6, 0.7, 0.2], 'obj_id': 19}, {'cam_R_m2c': [1.6, 0.7, 0.2], 'obj_id': 20}], 1: [{'cam_R_m2c': [3.6, 0.7, 0.2], 'obj_id': 17}, {'cam_R_m2c': [4.6, 0.7, 0.2], 'obj_id': 18}]}
print(type(opt_dict))
# <class 'dict'>
print(len(opt_dict))
# 2

示例2 读取 .yml

gt.yml

path: ../datasets/coco128  # dataset root dir
train: data/train.txt  # train images (relative to 'path') 128 images
val: data/val.txt  # val images (relative to 'path') 128 images
test:  # test images (optional)
# Classes
nc: 10  # number of classes
names: ['combustion_lining', 'fan', 'mixer']

test.py

import yaml
gt = 'gt.yml'
with open(gt) as f:
    opt_dict = yaml.load(f, Loader=yaml.FullLoader)
print(opt_dict)
# {'path': '../datasets/coco128', 'train': 'data/train.txt', 'val': 'data/val.txt', 'test': None, 'nc': 10, 'names': ['combustion_lining', 'fan', 'mixer']}
print(type(opt_dict))
# <class 'dict'>
print(len(opt_dict))
# 6

示例3 写入 .yml

test.py

import yaml
hyp1 = {'Name': 'Zara', 'Age': 7, 'score': [1, 2, 3]}
hyp2 = {0: [{'cam_R_m2c': [0.6, 0.7, 0.2], 'obj_id': 19}, {'cam_R_m2c': [1.6, 0.7, 0.2], 'obj_id': 20}],
        1: [{'cam_R_m2c': [3.6, 0.7, 0.2], 'obj_id': 17}, {'cam_R_m2c': [4.6, 0.7, 0.2], 'obj_id': 18}]}
gt = 'gt.yml'
with open(gt, 'w') as f:
    yaml.dump(hyp1, f, sort_keys=False)
with open(gt, 'a') as f:
    yaml.dump(hyp2, f, sort_keys=False)

gt.yml

Name: Zara
Age: 7
score:
- cam_R_m2c:
  - 0.6
  - 0.7
  - 0.2
  obj_id: 19
- cam_R_m2c:
  - 1.6
  - 0.7
  - 0.2
  obj_id: 20
- cam_R_m2c:
  - 3.6
  - 0.7
  - 0.2
  obj_id: 17
- cam_R_m2c:
  - 4.6
  - 0.7
  - 0.2
  obj_id: 18

关于numpy数据

yaml不能直接写入numpy数组，虽然写入不出错，但读取会出错。

import yaml
import numpy as np
a = np.array([1, 2, 3])
b = np.array([4, 5, 6])
hyp1 = {'a': a, 'b': b}
gt = 'gt_np.yml'
# 写入, 不出错
with open(gt, 'w') as f:
    yaml.dump(hyp1, f, sort_keys=False)
# 读取, 出错
with open(gt) as f:
    opt_dict = yaml.load(f, Loader=yaml.FullLoader)
print(opt_dict)
# yaml.constructor.ConstructorError: could not determine a constructor for the tag 'tag:yaml.org,2002:python/object/apply:numpy.core.multiarray._reconstruct'

需要通过 int() 转换一下。

import yaml
import numpy as np
a = np.array([1, 2, 3])
b = np.array([4, 5, 6])
a1 = [int(x) for x in a]
b1 = [int(x) for x in b]
hyp1 = {'a': a1, 'b': b1}
gt = 'gt_np.yml'
# 写入, 不出错
with open(gt, 'w') as f:
    yaml.dump(hyp1, f, sort_keys=False)