granular-configuration-language 2.1.0

This general purpose configuration utility library allows your code use YAML as a configuration language for internal and external parties, allowing configuration to be crafted from multiple sources and merged just before use, with added YAML Tags that run lazily for added functionality.

granular-configuration-language – A General Purpose Configuration Utility Library

⚠️ This library is meant for trusted configuration files. ⚠️

How to get started?

See Documentation - Getting Started.

How to install?

From PyPi:

pip install granular-configuration-language

Why does this exist?

This library exists to allow your code use YAML as a configuration language for internal and external parties, allowing configuration to be crafted from multiple sources and merged just before use, with added YAML Tags that run lazily for added functionality, and plugin support for creating custom YAML Tags.

Some use cases:

• You are writing a library to help connect to some databases. You want users to easily changes settings and defined databases by name.
  • Conceptual Example:
    • Library Code:

      # database_util/config/config.py
      CONFIG = LazyLoadConfiguration(
          Path(__file___).parent / "config.yaml",
          "./database-util-config.yaml",
          "~/configs/database-util-config.yaml",
          base_path="database-util",
          env_location_var_name="ORG_COMMON_CONFIG_LOCATIONS",
      )
      

    • Library configuration:

      # database_util/config/config.yaml
      database-util:
        common_settings:
          use_decimal: true
          encryption_type: secure
        databases: {} # Empty mapping, for users define
      

    • User application configuration:

      # ~/configs/database-util-config.yaml
      database-util:
        common_settings:
          use_decimal: false
        databases:
          datebase1:
            location: http://somewhere
            user: !Mask ${DB_USERNAME}
            password: !Mask ${DB_PASSWORD}
      

• You are deploying an application that has multiple deployment types with specific settings.
  • Conceptual Example:
    • Library Code:

      # app/config/config.py
      CONFIG = LazyLoadConfiguration(
          Path(__file___).parent / "config.yaml",
          "./database-util-config.yaml",
          base_path="app",
      )
      

    • Base configuration:

      # app/config/config.yaml
      app:
        log_as: really cool app name
        log_to: nowhere
      

    • AWS Lambda deploy:

      # ./database-util-config.yaml
      app:
        log_to: std_out
      

    • Server deploy:

      # ./database-util-config.yaml
      app:
        log_to: !Sub file://var/log/${$.app.log_as}.log
      

• You are writing a pytest plugin that create test data using named fixtures configured by the user.
  • Conceptual Examples:
    • Library Code:

      # fixture_gen/config/config.py
      CONFIG = LazyLoadConfiguration(
          Path(__file___).parent / "fixture_config.yaml",
          *Path().rglob("fixture_config.yaml"),
          base_path="fixture-gen",
      ).config
      #
      for name, spec in CONFIG.fixtures:
          generate_fixture(name, spec)
      

    • Library configuration:

      # fixture_gen/config/fixture_config.yaml
      fixture-gen:
        fixtures: {} # Empty mapping, for users define
      

    • User application configuration:

      # fixture_config.yaml
      fixture-gen:
        fixtures:
          fixture1:
            api: does something
      

Why the long name?

• It's "granular" because you can specify settings across multiple files at a fine granularity for overriding values.
• It is meant for trusted "configuration" files.
• Including "language" makes it clear that this is not the source of configuration, but a library for processing generic configuration files.
  • Feedback was that "granular-configuration" sounded like it was the source for configuration.
  • "Format" sounded weirder than "language".
  • Including "YAML" sounded like this was trying to be more than YAML, rather than just using YAML.