Skip to main content

Environment interface with (optionally encrypted) .env with hashicorp vault support

Project description

Maintenance PyPI version PyPI downloads Python versions License: MIT

ENV EXtended

envex is a dotenv .env aware environment variable handler with typing features and HashiCorp Vault support.

Overview

This module offers a convenient interface for managing application environments and configurations while adhering to the 12-factor app methodology. It avoids having any environment-specific variables and security-sensitive data within application code.

An Env instance delivers a lot of functionality by providing a type-smart front-end to os.environ, providing a superset of os.environ functionality.

Installation

pip install envex

envex supports Python 3.12 through 3.14.

Features

In addition to the standard os.environ functionality with automatic type conversion, envex provides the following features:

Supports AES-256 encrypted environment files

When enabled with the decrypt=True argument and provided with the decryption password envex searches for .env.enc files first but falls back to .env. Using encrypted environment files avoids using plain text files on the filesystem that contain sensitive information. The provided envcrypt utility conveniently allows conversion between encrypted and non-encrypted formats.

Vault support

Alternatively, envex provides seamless integration with HashiCorp Vault. This reduces the need to store plaintext secrets on the filesystem and provides a more secure approach for managing secrets. HashiCorp Vault functionality is optional, and is activated automatically when the hvac module is installed into the active virtual environment, and where connection and authentication to Vault succeed. Values fetched from Vault are cached in the SecretsManager instance during its lifetime to reduce repeated API reads. Envex does not automatically expire this cache or invalidate it when values are changed outside the instance; create a new SecretsManager or call get_secrets() to refresh values from Vault.

Extended variables

envex provides many features not available in other dotenv handlers (python-dotenv, etc.) including recursive expansion of "template" style variables, supporting don't-repeat-yourself (DRY) patterns.

envex provides multiple ways to fetch environment variables:

from envex import env

assert env['HOME'] == '/Users/username'

env['TESTING'] = 'This is a test'
assert env.get('TESTING') == 'This is a test'
assert env['TESTING'] == 'This is a test'
assert env('TESTING') == 'This is a test'

It provides the ability to set variable defaults without overriding, much the same as a standard python dict:

import os

assert os.environ['TESTING'] == 'This is a test'

assert env.get('UNSET_VAR') is None
env.set('UNSET_VAR', 'this is now set')
assert env.get('UNSET_VAR') is not None
env.setdefault('UNSET_VAR', 'and this is a default value but only if not set')
assert env.get('UNSET_VAR') == 'this is now set'
env.set('UNSET_VAR', None)
assert env.get('UNSET_VAR') is None
del env['UNSET_VAR']
assert env.get('UNSET_VAR') is None

Note that there is a subtle difference between

  • env.get(<variable>, default=<default value>) and
  • env(<variable>, default=<default value>). If the variable is initially unset, the former simply returns the default value or None, but the second also sets the value to the default value if one was provided in the environment, unless it was already set.

An Env instance can also read a .env (the default name) file and update the application environment accordingly. It can read this either when created with Env(readenv=True) or directly by using the method read_env(). If provided, the readenv=True parameter enable reading environment files according to the search_path provided (the current directory by default) and the parents=True parameter extends the search to parent directories should the initial target not be found.

To override the default name of the environment file, use the DOTENV environment variable.

Variables in environment files will not overwrite existing environment variables by default. overwrite=True must be used to change this behaviour.

Env can also be passed one or more BytesIO or String IO objects as positional parameters from which to read environment variables are read as though they were files. IO objects passed in this way differ only in that by default variables evaluated from their content overwrites existing variables as though overwrite=True was used. To change the default behaviour, explicitly use overwrite=False.

Other kwargs that can be passed to Env when created:

  • environ (env): pass the environment to update, default is os.environ, passing an empty dict will create a new env
  • readenv (bool): search for and read .env files (default is False)
  • env_file (str): name of the env file, os.environ["DOTENV"] if set, or .env is the default
  • search_path (str or list): a single path or list of paths to search for the env files search_path may also be passed as a colon-separated list (or semicolon on Windows) of directories to search.
  • parents (bool): search (or not) parents of dirs in the search_path
  • overwrite (bool): overwrite already set values read from .env, default is to only set if not currently set
  • update (bool): push loaded values to os.environ if true (default), otherwise changes internally only. The "export" prefix in a .env file is accepted for dotenv syntax compatibility, but it does not force publication to os.environ when update=False.
  • working_dirs (bool): add CWD for the current process and PWD of source .env file
  • exception: (optional) Exception class to raise on error (default is KeyError)
  • errors: bool whether to raise error on missing env_file (default is False)
  • decrypt: bool whether to support decryption of encrypted env files (default is False)
  • password: str the password, environment variable or file/path to use for decryption (see below)
  • kwargs: (keyword args, optional) additional environment variables to add/override, after recognized loader and Vault options are applied

In addition, Env supports a few HashiCorp Vault configuration parameters as well:

  • url: (str, optional) vault url, default is $VAULT_ADDR
  • token: (str, optional) vault token, default is $VAULT_TOKEN or content of ~/.vault-token
  • cert: (str or tuple, optional) path to a combined client certificate/key PEM file, or a (cert, key) tuple of separate PEM file paths
  • verify: (bool | str | None, optional) whether to verify the server certificate, a CA bundle file or CA directory path, or None to derive verification from environment. Non-None values take precedence over environment variables. When None, VAULT_SKIP_VERIFY=true disables verification; otherwise VAULT_CACERT, VAULT_CAPATH, then the default trust store are used (default is None).
  • base_path: (optional) logical secrets base path, or "environment" for secrets (default is None). This is used to prefix the logical path to the secret, i.e. f"{base_path}/key".
  • mount_point: (optional) Vault secrets engine mount point (default is secret/; values may be provided with or without slashes and are normalized internally).

Environment values override Vault values by default. Set ENVEX_SOURCE=vault to let Vault values override local environment values.

Some type-smart functions act as an alternative to Env.get and having to parse the result:

from envex import env

env['AN_INTEGER_VALUE'] = 2875083
assert env.get('AN_INTEGER_VALUE') == '2875083'
assert env.int('AN_INTEGER_VALUE') == 2875083
assert env('AN_INTEGER_VALUE', type=int) == 2875083

env['A_TRUE_VALUE'] = True
assert env.get('A_TRUE_VALUE') == 'True'
assert env.bool('A_TRUE_VALUE') is True
assert env('A_TRUE_VALUE', type=bool) is True

env['A_FALSE_VALUE'] = 0
assert env.get('A_FALSE_VALUE') == '0'
assert env.int('A_FALSE_VALUE') == 0
assert env.bool('A_FALSE_VALUE') is False
assert env('A_FALSE_VALUE', type=bool) is False

env['A_FLOAT_VALUE'] = 287.5083
assert env.get('A_FLOAT_VALUE') == '287.5083'
assert env.float('A_FLOAT_VALUE') == 287.5083
assert env('A_FLOAT_VALUE', type=float) == 287.5083

env['A_LIST_VALUE'] = '1,"two",3,"four"'
assert env.get('A_LIST_VALUE') == '1,"two",3,"four"'
assert env.list('A_LIST_VALUE') == ['1', 'two', '3', 'four']
assert env('A_LIST_VALUE', type=list) == ['1', 'two', '3', 'four']

Integer and boolean parsing is strict: missing or empty integer values parse as 0, but malformed non-empty integer strings raise ValueError.

Environment variables are always stored as strings. This is enforced by the underlying os.environ, but also true of any provided environment, which must use the MutableMapping[str, str] contract.

Encrypted Environment Files

To enhance security of environment files that exist on the filesystem envex supports the creation and use of authenticated AES-256-GCM encrypted files. Version 5.0.0 changes the encrypted file format from the legacy AES-CBC format to AES-GCM because the old format did not authenticate ciphertext and was vulnerable to padding-oracle style attacks if decryption failures were exposed. The new format authenticates encrypted data before returning plaintext and is not subject to that vulnerability.

Encrypted .env files are named as .env.enc by default (strictly, ${DOTENV:-.env}.enc), to distinguish them from the unencrypted version, but this is only by convention; both to distinguish the files visually, and to prevent other dot-env readers from using them.

If the feature is enabled and a passphrase is provided when the environment file is read, envex determines automatically if it contains encrypted data. If the .enc version of the environment file does not exist, the .env file - encrypted or not - is used as a fallback, but will otherwise be ignored.

Version 5 Encrypted File Compatibility

This is a breaking encrypted-file format change:

  • Files encrypted by envex 5.0.0 and later cannot be decrypted by envex 4.x or older.
  • Envex 5.0.0 can still decrypt legacy AES-CBC files with explicit legacy mode so existing files can be migrated.
  • Any file re-encrypted with envex 5.0.0 will require envex 5.0.0 or later to decrypt.
  • Normal decryption rejects legacy AES-CBC files by default so a modified AES-GCM file cannot be downgraded into the unauthenticated legacy path.

To migrate an existing encrypted file, decrypt it with a version that can read the current file, then re-encrypt it with envex 5.0.0 or later:

envcrypt --decrypt --legacy --password "$PASSPHRASE" .env.enc .env
envcrypt --encrypt --password "$PASSPHRASE" .env .env.enc

If the passphrase is stored in an environment variable or file, use --environ NAME or --file PATH instead of --password. After migration, verify the new file can be decrypted by envex 5.0.0 or later and remove any temporary plaintext file:

envcrypt --decrypt --password "$PASSPHRASE" .env.enc .env.verify
rm .env .env.verify

The envcrypt CLI utility supports the encryption and decryption of environment files.

usage: envcrypt.py [-h] [-P PASSWORD | -E ENVIRON | -F FILE] [-e | -d] [--legacy] [-r] [-v] input [output]

envcrypt: File encrypt/decrypt (authenticated AES-256-GCM)

positional arguments:
input                    File to encrypt or decrypt
output                   Output file (optional) (default: None)

options:
-h, --help               show this help message and exit

-P, --password PASSWORD  Use given password (default: None)
-E, --environ ENVIRON    Read password from provided environment variable (default: None)
-F, --file FILE          Read password from a given file (default: None)

-e, --encrypt            Encrypt input (default: False)
-d, --decrypt            Decrypt input (default: False)
--legacy                 Allow decrypting legacy AES-CBC files for migration (default: False)

-r, --rm                 Remove input file after successful conversion (default: False)
-v, --verbose            Increase output verbosity (default: False)

Either --encrypt or --decrypt must be provided

A passphrase is required, one of --password, --environ, or --file must also be given. If the passphrase is not provided, the utility will prompt for it stdin is available and is a terminal.

After an encryption or decryption operation, the input file is retained by default, but can be removed using the --rm option.

Specifying the output filename is optional, and if not given the utility will append .enc to the input filename for encryption, or remove .enc for decryption. The --rm option will remove the input file on success.

Decrypted values are loaded with the same rules as plaintext dotenv values. By default, load_env() and Env(readenv=True) publish loaded values to os.environ; pass update=False or use an isolated environ mapping to keep loaded decrypted values out of the process environment. The "export" prefix in a decrypted environment file does not override update=False. However, the passphrase must be available in order for the environment to be read, and therefore the security of the encrypted file is only as strong as the security of that passphrase.

Three options are available when using the Env class to read encrypted environment files. A value passed to the password parameter can be a string which is by default the plain text passphrase. If it is prefixed by $ then it reads the passphrase via the named environment variable, or if it is prefixed by @ it is read read from a file.

Benefits of Encrypted Environment Files

While slightly less convenient (having to manually encrypt and decrypt environment files), the benefits of using encrypted environment files are:

  • Prevents sensitive data leakage from plaintext .env files.
  • Ideal for use in shared or distributed systems.
  • Mitigates risks associated with misconfigured access permissions.
  • Provides an additional layer of security for sensitive data.

Also, encrypted .env files are not available to other .env aware software.

Vault

In addition to handling of the os environment and .env files - encrypted or plain text - envex supports selectively fetching secrets from HashiCorp Vault using the kv.v2 engine. This provides a secure secrets store, and completely avoids exposing secrets in plain text on the filesystem and in particular in published docker images. It also prevents storing secrets in the operating system’s environment, which can be inspected by external processes.

This document does not cover how to set up and configure a Vault server, but you can find useful resources on the following websites:

  • hashicorp.com for the developer documentation and detailed information and tutorials on setting up and hosting your own vault server, and
  • vaultproject.io for information about HashiCorp's managed cloud offering.

To access the Vault server, you need a token with a role that has read permission for the path to the secrets. A read-only profile is the strongly recommended policy for tokens used at runtime by the application.

This library provides a utility called env2hvac that can import (create or update) a typical .env file into vault. The utility stores one KV document per application and environment, using the logical path / - for example, myapp/prod - with .env keys stored inside that document. The utility requires that the token has a role with create or update permission for the target secrets path on the vault server. By default, the utility uses the kv.v2 secrets engine mounted at secret/. A key such as DB_PASSWORD from myapp/prod is stored inside the KV document at logical path myapp/prod.

Environment Variables

The SecretsManager and Vault client leverage environment variables for their configuration. This ensures a degree of transparency as it allows the client to use them but mitigates the need for the client code to be aware of their presence. A summary of these variables is in the following table:

Variable Description
VAULT_ADDR The URL of the vault server
VAULT_TOKEN The vault token to use for authentication
VAULT_CACERT The path to the CA certificate to use for TLS verification
VAULT_CAPATH The path to a directory of CA certificates to use for TLS verification
VAULT_CLIENT_CERT The path to the client certificate PEM file, or a combined certificate/key PEM file for TLS client auth
VAULT_CLIENT_KEY The optional path to a separate client key PEM file for TLS client auth

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

envex-5.0.0.tar.gz (32.0 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

envex-5.0.0-py3-none-any.whl (38.4 kB view details)

Uploaded Python 3

File details

Details for the file envex-5.0.0.tar.gz.

File metadata

  • Download URL: envex-5.0.0.tar.gz
  • Upload date:
  • Size: 32.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.16 {"installer":{"name":"uv","version":"0.11.16","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for envex-5.0.0.tar.gz
Algorithm Hash digest
SHA256 1a1144c6c112d58a4c21e35f23375c12b1ce25c8ddd4660a6990066dadf2576b
MD5 57e79083a1c83f259787986f43d3baca
BLAKE2b-256 c2d0856dbb783d10c343c03dfe717cddd8f3f87934e651c8ea0d77e36f16bd04

See more details on using hashes here.

File details

Details for the file envex-5.0.0-py3-none-any.whl.

File metadata

  • Download URL: envex-5.0.0-py3-none-any.whl
  • Upload date:
  • Size: 38.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.16 {"installer":{"name":"uv","version":"0.11.16","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for envex-5.0.0-py3-none-any.whl
Algorithm Hash digest
SHA256 af04b9b03445fb16118d1277458b2ce936a24df5079bcfd9bdc984d69d34dd99
MD5 721ea9b6db85c96ed09859a84c20d446
BLAKE2b-256 85b05bbc112840e632ba44142d8436afb23183b86822c9ef5c013d78fe9cd773

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page