py-jwt-verifier

Python JWT Verfier - Verifies the signature of a digitally signed JWT.

Realease notes

Version	Release notes
0.7.0	Added support for Salesforce. Minor fixes. Latest stable version. Re-named to py_jwt_verifier.
0.6.0	Added support for Google, Microsoft and Auth0.
0.5.0	MAJOR Release. Production stable. Added cache control.
0.4.0-beta	Security fix. Certificate Chain is mandatory for Okta Custom URL Domain.
0.3.0-beta	Minor release. Added support for Okta Custom URL Domain.
0.2.0-beta	Minor release. Added support for AWS Cognito JWT.
0.1.0-beta	Minor release. Increased configurability.
0.0.1-beta	First package release.

PyPy: https://pypi.org/project/py-jwt-verifier/

Supported IdPs

• Okta
• AWS Cognito
• Google
• Microsoft
• Auth0
• Salesforce

---

Supported Signing Algorithms

• RS256

---

Disclaimer

This library is provded as is. None of the listed IdPs will provide support for issues related with the present library. I am the sole maintainer of it.

---

Process Chain

1. Once the class is instantiated the following checks are performed:

   • JWT Format.
   • JWT Expiration time.
   • JWT Claims if given when the class was instantiated.
   • Is Algorithm Supported.

2. After the above checks are done, it will verify the token signature with the apropriate signing algorithm based on the "alg" header claim. If the signature is valid, it will return None. Else, it will raise a exception.

   • If a check fails at any given step, the exception PyJwtException will be raised.
   • The /.well-knwon/openid-configuration endoint will be compiled based on the 'iss' claim.
   • The /keys endpoint will be determined from the output of the /.well-knwon/openid-configuration endpoint
   • The response from /keys will be cached (requests_cache) for subsequent calls.

---

Installation

pip install py-jwt-verifier

---

Usage Examples

from py_jwt_verifier import PyJwtVerifier, PyJwtException
jwt = access_token / id_token
try:
    PyJwtVerifier(jwt)
except PyJwtException as e:
    print(f"Exception caught. Error: {e}")

• If auto_verify is set to False the class will not perform the signature validation. To check the signature the verify() method needs to be used. By default, the method will return None. In order to return the decoded jwt data (header + payload) True has to be passed. Example:

from py_jwt_verifier import PyJwtVerifier, PyJwtException
jwt = access_token / id_token
validator = PyJwtVerifier(jwt, auto_verify=False)
try:
    payload = validator.verify(True)
    print(payload)
except PyJwtException as e:
    print(f"Exception caught. Error: {e}")

Custom Claim Validation:

from py_jwt_verifier import PyJwtVerifier, PyJwtException
jwt = access_token / id_token
validator = PyJwtVerifier(jwt, auto_verify=False, custom_claim_name="custom_claim_value")
try:
    payload = validator.verify(True)
    print(payload)
except PyJwtException as e:
    print(f"Exception caught. Error: {e}")

Cache Control:

• redis

from redis import StrictRedis
from py_jwt_verifier import PyJwtVerifier, PyJwtException


redis = StrictRedis(host="localhost", port=6390)

jwt = access_token / id_token
validator = PyJwtVerifier(jwt, auto_verify=False, cache_store="redis", cache_store_connection=redis)
try:
    payload = validator.verify(True)
    print(payload)
except PyJwtException as e:
    print(f"Exception caught. Error: {e}")

• pymongo

from pymongo import MongoClient
from py_jwt_verifier import PyJwtVerifier, PyJwtException


mongo = MongoClient("localhost", 27017)

jwt = access_token / id_token
validator = PyJwtVerifier(jwt, auto_verify=False, cache_store="mongo", cache_store_connection=mongo)
try:
    payload = validator.verify(True)
    print(payload)
except PyJwtException as e:
    print(f"Exception caught. Error: {e}")

---

Class Attributes

• The class PyJwtVerifier currently accepts:

Attribute	Required	Default value
jwt	Yes	None
cid - OIDC Client ID	No	None
aud - Audience	No	None
iss - Issuer	No	None
auto_verify	No	True
check_expiry	No	True
cache_enabled	No	True
cache_lifetime	No	1 day
cache_store	No	sqlite
cache_store_connection	No	None

• The class method verify() currently accepts:

Attribute	Required	Default value
get_payload	No	False

Note

The reason why this class returns None or exception is to provide more flexibility. Not everyone needs to return the decoded payload of the jwt. It is recommended to use it within try: except blocks.

---

Cache Control

This library relies on the requests and requests_cache libraries. The caching control mechanism was implemented around these libraries.

Cache Control attributes explained

• cache_enabled - Attribute type: Boolean - Accepted Values: True / False. Determines if the cache control mechanism will be used. If set to False, the response from the /.well-knwon/openid-configuration and /keys endpoints will never be cached.
• cache_lifetime - Attribute type: Integer - Accepted Values: 1 - 30. Represents the number of days the cache will be stored and used. The maximum value can not be higher than 30 days or less than 1 day.
• cache_store - Attribute type: String - Accepted Values: "memory", "sqlite", "mongo", "redis". Determines how and where the requests_cache library will store the responses and connect to the caching store. When there is no cache_store_connection provided, it will rely on the defaults for mongo and redis.
• cache_store_connection - Attribute type: DB instance object - Should be used only when the cache_store is set to "mongo" or "redis". This is necessary for production environments so you can specify the host, port, database, user and password to use to connect to the respective caching database selected.

Note

When using redis or mongo as caching database solutions, the appropriate python connector libraries will be required (pymongo / redis). For additional information in regards of how requests_cache works, please review their docs: https://requests-cache.readthedocs.io/en/latest/

---

SUGGESTIONS?

Please feel free to email me at adrian.lazar95@outlook.com or lzr.adrian95@gmail.com. I am opened to improvement / suggestions and critics.