python-config-parser lets you create runtime configuration objects using json or yaml files.
- Declarative configurations without using .ini files
- Access using OOP or subscriptable, which means that you can iterate the config object items
- Runtime validation using schema
- Automatic environment variables interpolation
- Automatic parser selecting by config file extension
Use pip to install it.
pip install python-config-parserBy default, the config file will look for the following config files in the .config directory: config.json, config.yaml, config.yml.
You can also pass a config directory of your preference (assuming your current directory).
You may or may not use schema validation. If you want to use it, it will validate the whole config object before returning it.
If you choose not to use it, it won't validate the config object before returning it, and it may generate runtime access inconsistencies.
How to use schema
from schema import Use, And SCHEMA_CONFIG = { 'core': { 'logging': { 'format': And(Use(str), lambda string: len(string) > 0), 'date_fmt': And(Use(str), lambda string: len(string) > 0), 'random_env_variable': str }, 'allowed_clients': [{ 'ip': str, # <- Here you can use regex to validate the ip format 'timeout': int } ] } }The config.yml file
core: random_env_variable: ${RANDOM_ENV_VARIABLE} logging: format: "[%(asctime)s][%(levelname)s]: %(message)s" date_fmt: "%d-%b-%y %H:%M:%S" allowed_clients: - ip: 192.168.0.10 timeout: 60 - ip: 192.168.0.11 timeout: 100A json config file would be something like:
{ "core": { "random_env_variable": "${RANDOM_ENV_VARIABLE}", "logging": { "format": "[%(asctime)s][%(levelname)s]: %(message)s", "date_fmt": "%d-%b-%y %H:%M:%S" }, "allowed_clients": [ { "ip": "192.168.0.10", "timeout": 60 }, { "ip": "192.168.0.11", "timeout": 100 } ] } }The instance of Config Class:
from pyconfigparser import configparser, ConfigError import logging try: config = configparser.get_config(SCHEMA_CONFIG) # <- Here I'm using that SCHEMA_CONFIG we had declared, and the dir file default value is being used except ConfigError as e: print(e) exit() # to access your config you need just: fmt = config.core.logging.format # look this, at this point I'm already using the config variable date_fmt = config['core']['logging']['date_fmt'] # here subscriptable access logging.getLogger(__name__) logging.basicConfig( format=fmt, datefmt=date_fmt, level=logging.INFO ) # the list of object example: for client in config.core.allowed_clients: print(client.ip) print(client.timeout) # The config object's parts which is not a list can also be itered but, it'll give you the attribute's names # So you can access the values by subscriptale access for logging_section_attr_key in config.core.logging: print(config.core.logging[logging_section_attr_key]) # Accessing the environment variable already resolved print(config.random_env_variable)Since you've already created the first Config's instance this instance will be cached inside Config class, so after this first creation you can just invoke Config.get_config()
from pyconfigparser import configparser config = configparser.get_config() # At this point you already have the configuration properties in your config objectYou can also disable the action to cache the instance config
from pyconfigparser import configparser configparser.hold_an_instance = FalseFork https://github.com/BrunoSilvaAndrade/python-config-parser/ , create commit and pull request to develop.