如何根据python的键值将YAML解析为多个compose.yaml

Question

我正在解析YAML并将其分解为多个不同的YAML文件。我使用PyYAML的构造函数来实现它，但效果很差。

这是我项目的一部分，我需要根据我收到的yaml文件中的键值解析并拆分成多个不同的yaml文件。

我收到的yaml文件看起来像这样

testname: testname
testall:
    test1:
        name: name1
        location: 0
    test2: 
        name: name2
        location: 2
    test3: 
        name: name3
        location: 0
    test4: 
        name: name4
        location: 2
    ...
locations:
    - 0
    - 2
    - ...  

我想解析它并按设备拆分如下：

# location0.yaml
testname:test
tests:
    test1:
        name:test1
        location:0
    test3: 
        name: test3
        location: 0

# location2.yaml
testname:test
tests:
    test2:
        name:test2
        location:0
    test4: 
        name: test4
        location: 0

如何解析上面的表格？

Answer 1

虽然您可以使用PyYAML执行此操作，但您必须将自己限制为YAML 1.1。对于这种读 - 修改 - 写，你应该使用ruamel.yaml（免责声明：我是该包的作者）。这不仅支持YAML 1.2，它还会保留任何注释，标签和锚名称，以防它们出现在源代码中，并且如果需要，可以保留标量，文字和折叠样式等的引号。

另请注意，您的输出是无效的YAML，您不能将多行普通（即未引用）标量作为（块样式）映射的关键。你必须写：

"testname:test
tests":

但我认为你的意思是成为根级映射的两个键：

testname: test
tests:

---

假设你的输入是在input.yaml：

testname: testname
testall:
    test1:
        name: name1    # this is just the first name
        location: 0
    test2: 
        name: "name2"  # quotes added for demo purposes
        location: 2
    test3: 
        name: name3    # as this has the same location as name1 
        location: 0    # these should be together
    test4: 
        name: name4    # and this one goes with name2
        location: 2
locations:
    - 0
    - 2

你可以做：

import sys
from pathlib import Path
import ruamel.yaml

in_file = Path('input.yaml')


yaml = ruamel.yaml.YAML()
yaml.indent(mapping=4, sequence=6, offset=4)  # this matches your input
yaml.preserve_quotes = True
data = yaml.load(in_file)

for loc in data['locations']:
    out_name = f'location{loc}.yaml'
    tests = {}
    d = ruamel.yaml.comments.CommentedMap(dict(testname="test", tests=tests))
    d.yaml_set_start_comment(out_name)
    testall = data['testall']
    for test in testall:
        if loc == testall[test]['location']:
           tests[test] = testall[test]
           tests[test]['location'] = 0
    # since you set location to zero and this affects data, make sure to remove 
    # the items. This will prevent things from going wrong in case the
    # locations sequence does have zero, but not as its first number
    for key in tests:
         del testall[key]
    yaml.dump(d, Path(out_name))

给location0.yaml：

# location0.yaml
testname: test
tests:
    test1:
        name: name1    # this is just the first name
        location: 0
    test3:
        name: name3    # as this has the same location as name1 
        location: 0    # these should be together

和location2.yaml：

# location2.yaml
testname: test
tests:
    test2:
        name: "name2"  # quotes added for demo purposes
        location: 0
    test4:
        name: name4    # and this one goes with name2
        location: 0