Status (master branch)

This is an opinionated Python2/Python3 thin layer over the Python configparser library to deal with:

• configuration variants (PROD, DEV...) expressed as alternative keys in ini file
• jinja2 as (optional) interpolation method for configuration values

For Python2, we use Python3 configparser backport.

With this example:

[group1]
debug=0
debug[DEV]=1

We have two default configuration variant for the debug key: the DEV one and the default one.

If you don't do anything special, the debug value will be 0 (standard default value).

But if you initialize the library with DEV as configuration name, the debug value will be 1.

Now, if you use PROD as configuration name, as there is no debug[PROD] line/variant, the debug value will fallback to standard value: 0 (in this example).

Still with the same example:

[group1]
debug=0
debug[DEV]=1

What about if we use DEV_JOHN_MONDAY as configuration name when initializing the library?

As there is no debug[DEV_JOHN_MONDAY] line, one might think that the retained value would be the default one: 0.

In fact, the retrained value will be 1! Why? Because _ (underscore) has a special meaning in configuration names. This is a kind of inheritance mark.

So DEV_JOHN_MONDAY means as a configuration name:

• use DEV_JOHN_MONDAY if there is a variant with this exact name
• (else) use DEV_JOHN (first level of inheritance) if there is a variant with this name: DEV_JOHN
• (else) use DEV (second level of inheritance) if there is a variant with this name: DEV
• (else) use standard/default value

So with this example:

[group1]
debug=0
debug[DEV]=1
debug[DEV_JOHN]=2
debug[DEV_PETER]=3
debug[DEV_JOHN_MONDAY]=4
debug[DEV_JOHN_TUESDAY]=5
debug[QA]=6

We get this table:

By default, we use jinja2 as interpolation method for configuration values. The Jinja2 context is initialized with environment variables.

So with this example:

[group1]
key=This is a Jinja2 test value: {{HOME}}

You will get the {{HOME}} placeholder replaced by the corresponding environment variable value.

Missing variables will be replaced by the empty string (without errors).

You can define your own Jinja2 context by adding interpolation=opinionated_configparser.Jinja2Interpolation(jinja2_context) to the OpinionatedConfigParser constructor call.

opinionated_configparser is just a thin layer over Python configparser library. So you use it exactly in the same way.

Just an example:

from opinionated_configparser import OpinionatedConfigParser


TEST_DICT = {
    "section1": {
        "key1": "value1",
        "key1[foo]": "value2",
        "key1[foo_bar]": "value3",
        "key2": "value4"
    },
    "section2": {
        "key3": "value5"
    }
}


parser = OpinionatedConfigParser(configuration_name="foo")
parser.read_dict(TEST_DICT)

# will output: value2
print(parser.get("section1", "key1"))

# [...]
# use the parser object exactly as configparser.ConfigParser one
# [...]

Just pass interpolation=None keyword argument in OpinionatedConfigParser constructor.

Or use interpolation=configparser.BasicInterpolation() to get the default interpolation method of configparser API.

Just pass interpolation=opinionated_configparser.Jinja2Interpolation(jinja2_context, **jinja2_template_kwargs) keyword argument in OpinionatedConfigParser constructor.

See CONTRIBUTING.md file.

See CODE_OF_CONDUCT.md file.

(If you are officially paid to work on MetWork Framework, please contact us to add your company logo here!)