birch 0.0.18

Utilities for pandas.

Simple hierarchical configuration for Python packages.

from birch import Birch
cfg = Birch('mypackage')
# read using a single API both the MYPACKAGE_SERVER_HOSTNAME environment variable
# and ~/.mypackage/cfg.json containing {'server': {'port': 55}}
connect(cfg['SERVER__HOSTNAME'], cfg['server']['port'])

Contents

• 1 Installation

• 2 Features

• 3 Use

  • 3.1 Basic use

  • 3.2 The get and mget methods

  • 3.3 Hierarchical configuration

  • 3.4 Resolution order

  • 3.5 Reloading configuration

• 4 Configuring birch

  • 4.1 Configuration directories

  • 4.2 File formats

• 5 Contributing

  • 5.1 Installing for development

  • 5.2 Running the tests

  • 5.3 Adding documentation

• 6 Credits

1 Installation

pip install birch

2 Features

• Supported formats: JSON, YAML.

• Pure python.

• Supports Python 3.5+.

• Supported and fully tested on Linux, OS X and Windows.

• XDG Base Directory Specification support.

3 Use

3.1 Basic use

birch provides an easy way to read simple hierarchical configurations for your Python package or application from both environment variables and configuration files.

birch uses namespaces to manage configuration values. The access to each namespace is done via a Birch object initialized with that namespace. Though written with a specific use case in mind, where a single package uses a single namespace to manage its configuration, any number of namespaces can be used in a single context. For example:

from birch import Birch
zubat_cfg = Birch('zubat')
golbat_cfg = Birch('golbat')

Each namespace encompasses all values set by either environment variables starting with <uppercase_namespace>_, or defined within cfg files (of a supported format) located in a set of pre-configured directories; this set defaults to the ~/.config/<namespace> (as par the XDG Base Directory Specification) and the ~/.<namespace> directories.

For example, the zubat namespace encompasses environment variables such as ZUBAT_HOSTNAME and ZUBAT__PORT, and all mappings in one of the files ~/.config/.zubat/cfg.json or ~/.zubat/cfg.json (if such a file exists).

Once defined in such a way, the Birch object can be used to access the values of mappings of both types (with or without the namespace suffix; casing is also ignored). For example:

>>> os.environ['ZUBAT_SERVER_HOST'] = 'www.zubat.com'
>>> os.environ['ZUBAT_SERVER_PORT'] = '87'
>>> from birch import Birch
>>> zubat_cfg = Birch('zubat')
>>>> zubat_cfg['ZUBAT_SERVER_HOST']
'www.zubat.com'
>>> zubat_cfg['SERVER_PORT']
'87'
>>> zubat_cfg['server_port']
'87'

3.2 The get and mget methods

Birch objects expose two methods that allow more nuanced retreival of configuration items:

The mget method allows the caller to supply a caster callable, through-which any found return value will be passed:

>>> os.environ['ZUBAT__PORT'] = '555'
>>> zubat_cfg = Birch('zubat')
>>> zubat_cfg.mget('port', int)
555

The get method additionally allows you to supply a default value, which is returned if no matching configuration entry is found:

>>> import os; os.environ['ZUBAT__PORT'] = '555'
>>> zubat_cfg = Birch('zubat')
>>> zubat_cfg.get('port', default=8888, caster=int)
555
>>> zubat_cfg.get('host', default='defhost')  # Default value is returned
'defhost'
>>> zubat_cfg.get('host')  # No error is thrown, None is returned

If no default value is provided, None is returned. To still have a KeyError raised in this case use throw=True in the function call: .. code-block:: python

>>> import os; os.environ['ZUBAT__PORT'] = '555'
>>> zubat_cfg = Birch('zubat')
>>> zubat_cfg.get('host', throw=True)  # An error is thrown
Traceback (most recent call last):
  ...
KeyError: zubat: No configuration value for HOST.

3.3 Hierarchical configuration

birch supports a simple hierarchy between configuration mappings. Hierarchy is either expressed explicitly in configuration files as nested object/entries (in the case of json and YAML files), or using __ (two underscore characters) in the configuration key - both in configuration files and environment variables. Thus, the ZUBAT__SERVER__PORT environment variable is equivalent to both {'server': {'port': 55}} and {'server__PORT': 55} mappings given in a ~/.zubat/cfg.json file, for example. Casing is ignored on all levels.

As such, hierarchical mappings can be accessed either using __ to indicate a hierarchical path, or using dict-like item access:

>>> os.environ['ZUBAT__SERVER__HOST'] = 'www.zubat.com'
>>> from birch import Birch
>>> zubat_cfg = Birch('zubat')
>>>> zubat_cfg['SERVER__HOST']
'www.zubat.com'
>>>> zubat_cfg['server']['HOST']
'www.zubat.com'
>>>> zubat_cfg['SERVER']['host']
'www.zubat.com'

Note that this is also true for non-hierarchical configuration file mappings, so {'server__port': 55}, even when given in this form in a configuration file, can be accessed using both zubat_cfg['SERVER__PORT'] and zubat_cfg['SERVER']['PORT'] (casing is still ignored on all levels).

3.4 Resolution order

A namespace is always loaded with matching environment variables after the configuration file has been loaded, and corresponding mappings will thus override their file-originating counterparts; e.g. the ZUBAT__SERVER__PORT environment variable will overwrite the value of the mapping {'server': {'port': 55}} given in a ~/.zubat/cfg.json file.

The lookup order of different files, while deterministic, is undefined and not part of the API. Thus, even with the load_all option set (see the Configuring birch section), cfg files with different file extensions can not be relied upon to provide private-vs-shared configuration functionality, or other such configuration modes.

3.5 Reloading configuration

Configuration values can be reloaded from all sources - both configuration files and environment variables - by calling the reload method:

>>> os.environ['ZUBAT__SERVER__HOST'] = 'www.zubat.com'
>>> from birch import Birch
>>> zubat_cfg = Birch('zubat')
>>>> zubat_cfg['SERVER__HOST']
'www.zubat.com'
>>> os.environ['ZUBAT__SERVER__HOST'] = 'New.value!'
>>> zubat_cfg.reload()
>>>> zubat_cfg['server']['HOST']
'New.value!'

You can set automatic configuration reload on every value inspection by setting auto_reload=True when initializing the Birch object:

>>> os.environ['ZUBAT__SERVER__HOST'] = 'www.zubat.com'
>>> from birch import Birch
>>> zubat_cfg = Birch('zubat', auto_reload=True)
>>>> zubat_cfg['SERVER__HOST']
'www.zubat.com'
>>> os.environ['ZUBAT__SERVER__HOST'] = 'New.value!'
>>>> zubat_cfg['server']['HOST']
'New.value!'

4 Configuring birch

4.1 Configuration directories

By default birch looks for files only in the ~/.config/<namespace> and ~/.<namespace> directories. You can set a different set of directories to read by populating the directories constructor parameter with a different directory path, or a list of paths.

Similarly, be default birch reads into the configuration tree only the first compliant file encountered during a lookup in all pre-configured directories; to instead load hierarchical configurations from all such files instead, the load_all constructor parameter can be set to True. Again, load order is undefined, and thus so is the resulting hierarchical configuration.

4.2 File formats

By default, birch will only try to read cfg.json files. To dictate a different set of supported formats, populate the supported_formats constructor parameter with the desired formats.

For example, Birch('zubat', supported_formats=['json', 'yaml']) will read both cfg.json and cfg.yaml files, while Birch('golbat', supported_formats='yaml') will ony read cfg.yaml (and cfg.yml) files.

Currently supported formats are:

• JSON - Looks for cfg.json files.

• YAML - Looks for cfg.yaml and cfg.yml files.

5 Contributing

Package author and current maintainer is Shay Palachy ([e-mail]); You are more than welcome to approach him for help. Contributions are very welcomed.

5.1 Installing for development

Clone:

git clone [e-mail]:shaypal5/birch.git

Install in development mode, including test dependencies:

cd birch
pip install -e '.[test]'

5.2 Running the tests

To run the tests use:

cd birch
pytest

5.3 Adding documentation

The project is documented using the numpy docstring conventions, which were chosen as they are perhaps the most widely-spread conventions that are both supported by common tools such as Sphinx and result in human-readable docstrings. When documenting code you add to this project, follow these conventions.

Additionally, if you update this README.rst file, use python setup.py checkdocs to validate it compiles.

6 Credits

Created by Shay Palachy ([e-mail]).