我正在尝试编写一个配置变量引擎，该引擎接受一个YAML文件（包含AWS配置变量）作为输入并转换为JSON，以便可以将其上传到HTTP k / v API（例如Consul）。我难倒的一个特性将允许开发人员在后续密钥中“包含”密钥集（用下划线标识，在最终有效负载中省略）。示例如下：
# Region
us-east-1:
# Any key preceded by an underscore (_) is considered a "tag group" and will not be uploaded to Consul KV unless explicitly included.
_taggroup1:
key1: value1
key2: value2
key3: value3
_taggroup2:
key4: value1
key5: value2
key6: value3
dev:
_include: us-east-1/_taggroup1
qa:
_include:
- us-east-1/_taggroup1
- us-east-1/_taggroup2
key6: baz
prod:
_include:
- us-east-1/_taggroup1
- us-east-1/_taggroup2
us-west-1:
_taggroup1:
key1: value1
key2: value2
key3: value3
_taggroup2:
key4: value1
key5: value2
key6: value3
dev:
_include:
- us-west-1/_taggroup1
qa:
_include:
- us-west-1/_taggroup1
- us-west-1/_taggroup2
key2: foo
prod:
_include:
- us-west-1/_taggroup1
- us-west-1/_taggroup2
key4: foo
key5: bar
key1: undef
us-west-1a:
qa:
_include: us-west-1/qa
prod:
_include: us-west-1/prod
us-west-1b:
_include: us-west-1/us-west-1a
正如您所看到的，我正在尝试构建一个配置文件，允许开发人员对变量进行分组，并在需要时随后包含/覆盖它们。
到目前为止，我为此实验编写的代码本质上是您的标准递归函数，其中添加了特定于此应用程序的内容：
# parse_input is a separate function that converts a YAML stream into
# an OrderedDict
original_dict = parse_input(stream1)
def print_dict(input_dict):
new_dict = collections.OrderedDict()
for key, value in input_dict.iteritems():
if key.startswith('_'):
if key == '_include':
if isinstance(value, list):
for item in value:
x = dpath.util.get(original_dict, item)
for k, v in x.iteritems():
new_dict[k] = v
else:
x = dpath.util.get(original_dict, value)
for k, v in x.iteritems():
new_dict[k] = v
else:
continue
continue
elif isinstance(value, dict):
new_dict[key] = print_dict(value)
else:
new_dict[key] = value
return new_dict
到目前为止我所取得的成果是这样的：
{
"us-east-1": {
"dev": {
"key1": "value1",
"key2": "value2",
"key3": "value3"
},
"qa": {
"key1": "value1",
"key2": "value2",
"key3": "value3",
"key4": "value1",
"key5": "value2",
"key6": "baz"
},
"prod": {
"key1": "value1",
"key2": "value2",
"key3": "value3",
"key4": "value1",
"key5": "value2",
"key6": "value3"
}
},
"us-west-1": {
"dev": {
"key1": "value1",
"key2": "value2",
"key3": "value3"
},
"qa": {
"key1": "value1",
"key2": "foo",
"key3": "value3",
"key4": "value1",
"key5": "value2",
"key6": "value3"
},
"prod": {
"key1": "undef",
"key2": "value2",
"key3": "value3",
"key4": "foo",
"key5": "bar",
"key6": "value3"
},
"us-west-1a": {
"qa": {
"_include": [
"us-west-1/_taggroup1",
"us-west-1/_taggroup2"
],
"key2": "foo"
},
"prod": {
"_include": [
"us-west-1/_taggroup1",
"us-west-1/_taggroup2"
],
"key4": "foo",
"key5": "bar",
"key1": "undef"
}
},
"us-west-1b": {
"qa": {
"_include": "us-west-1/qa"
},
"prod": {
"_include": "us-west-1/prod"
}
}
}
}
正如你所看到的，我似乎已经到了一半。我的问题是，在我最初的实验中，我通过在引用包含集时引用函数中的original_dict变量来获得有利结果（使用dpath返回键）。这很快就会变成一个问题，因为函数会更深入地解析（例如，在这种情况下是特定于AZ的变量），因为我不知道如何动态更新原始字典中的键，或者以其他方式跟踪更改，因此函数将注入带有_include键的密钥集，无法重新评估它们。
如何根据引用原始字典消除，而是动态跟踪更改，以便在树中更深入地正确评估_include个键？
答案 0 :(得分：1)
我认为此代码可以解决您遇到的问题。关键更改是使用print_dict的结果递归到dpath。我还崩溃了一些冗余的代码。
<强>代码：
import yaml
import collections
import json
import dpath
with open('data.yml', 'rb') as f:
original_dict = yaml.load(f)
def print_dict(input_dict):
new_dict = collections.OrderedDict()
for key, value in input_dict.iteritems():
if key.startswith('_'):
if key == '_include':
if not isinstance(value, list):
value = [value]
for item in value:
x = print_dict(dpath.util.get(original_dict, item))
for k, v in x.iteritems():
new_dict[k] = v
elif isinstance(value, dict):
new_dict[key] = print_dict(value)
else:
new_dict[key] = value
return new_dict
print(json.dumps(print_dict(original_dict), indent=2))
<强>输出：
{
"us-east-1": {
"qa": {
"key3": "value3",
"key2": "value2",
"key1": "value1",
"key6": "baz",
"key5": "value2",
"key4": "value1"
},
"prod": {
"key3": "value3",
"key2": "value2",
"key1": "value1",
"key6": "value3",
"key5": "value2",
"key4": "value1"
},
"dev": {
"key3": "value3",
"key2": "value2",
"key1": "value1"
}
},
"us-west-1": {
"qa": {
"key2": "value2",
"key3": "value3",
"key1": "value1",
"key6": "value3",
"key5": "value2",
"key4": "value1"
},
"us-west-1b": {
"qa": {
"key2": "value2",
"key3": "value3",
"key1": "value1",
"key6": "value3",
"key5": "value2",
"key4": "value1"
},
"prod": {
"key1": "value1",
"key3": "value3",
"key2": "value2",
"key6": "value3",
"key5": "bar",
"key4": "foo"
}
},
"prod": {
"key1": "value1",
"key3": "value3",
"key2": "value2",
"key6": "value3",
"key5": "bar",
"key4": "foo"
},
"us-west-1a": {
"qa": {
"key2": "value2",
"key3": "value3",
"key1": "value1",
"key6": "value3",
"key5": "value2",
"key4": "value1"
},
"prod": {
"key1": "value1",
"key3": "value3",
"key2": "value2",
"key6": "value3",
"key5": "bar",
"key4": "foo"
}
},
"dev": {
"key3": "value3",
"key2": "value2",
"key1": "value1"
}
}
}