Azure Schema Registry client library for Python

Azure Schema Registry is a schema repository service hosted by Azure Event Hubs, providing schema storage, versioning, and management. The registry is leveraged by encoders to reduce payload size while describing payload structure with schema identifiers rather than full schemas. This package provides:

1. A client library to register and retrieve schemas and their respective properties.

2. An JSON schema-based encoder capable of encoding and decoding payloads containing Schema Registry schema identifiers, corresponding to JSON schemas used for validation, and encoded content.

Source code | Package (PyPi) | Package (Conda) | API reference documentation | Samples | Changelog

Disclaimer

Azure SDK Python packages support for Python 2.7 has ended on 01 January 2022. For more information and questions, please refer to https://github.com/Azure/azure-sdk-for-python/issues/20691

Getting started

Install the package

Install the Azure Schema Registry client library for Python with pip:

pip install azure-schemaregistry==1.3.0b2

To use the built-in jsonschema validators with the JSON Schema Encoder, install jsonencoder extras:

pip install azure-schemaregistry[jsonencoder]==1.3.0b2

Prerequisites:

To use this package, you must have:

• Azure subscription - Create a free account
• Azure Schema Registry - Here is the quickstart guide to create a Schema Registry group using the Azure portal.
• Python 3.7 or later - Install Python

Authenticate the client

Interaction with Schema Registry starts with an instance of SchemaRegistryClient class. The client constructor takes an Azure Event Hubs fully qualified namespace and an Azure Active Directory credential:

• The fully qualified namespace of the Schema Registry instance should follow the format: <yournamespace>.servicebus.windows.net.

• An AAD credential that implements the TokenCredential protocol should be passed to the constructor. There are implementations of the TokenCredential protocol available in the azure-identity package. To use the credential types provided by azure-identity, please install the Azure Identity client library for Python with pip:

pip install azure-identity

• Additionally, to use the async API, you must first install an async transport, such as aiohttp:

pip install aiohttp

Create client using the azure-identity library:

from azure.schemaregistry import SchemaRegistryClient
from azure.identity import DefaultAzureCredential

credential = DefaultAzureCredential()
# Namespace should be similar to: '<your-eventhub-namespace>.servicebus.windows.net/'
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_FULLY_QUALIFIED_NAMESPACE']
schema_registry_client = SchemaRegistryClient(fully_qualified_namespace, credential)

Create JsonSchemaEncoder using the azure-schemaregistry library:

import os
from azure.schemaregistry import SchemaRegistryClient
from azure.schemaregistry.encoder.jsonencoder import JsonSchemaEncoder
from azure.identity import DefaultAzureCredential

credential = DefaultAzureCredential()
# Namespace should be similar to: '<your-eventhub-namespace>.servicebus.windows.net'
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ['SCHEMAREGISTRY_GROUP']
schema_registry_client = SchemaRegistryClient(fully_qualified_namespace, credential)
encoder = JsonSchemaEncoder(client=schema_registry_client, group_name=group_name)

Key concepts

• Schema: Schema is the organization or structure for data. More detailed information can be found here.

• Schema Group: A logical group of similar schemas based on business criteria, which can hold multiple versions of a schema. More detailed information can be found here.

• SchemaRegistryClient: SchemaRegistryClient provides the API for storing and retrieving schemas in schema registry.

• JsonSchemaEncoder: Provides API to encode content to and decode content from Binary Encoding, validate content against a JSON Schema, and cache schemas/schema IDs retrived from the registry using the SchemaRegistryClient locally.

• MessageType: Protocol defined under azure.schemaregistry that allows for JsonSchemaEncoder interoperability with certain Azure Messaging SDK message types. Support has been added to:

  • azure.eventhub.EventData for azure-eventhub>=5.9.0

MessageType

If a message type that follows the MessageType protocol is provided to the encoder, it will set the corresponding content and content type properties:

• content: Binary-encoded, JSON schema-validated payload (in general, format-specific payload)

• content type: a string of the format application/json;serialization=Json+<schema ID>, where:

  • application/json;serialization=Json is the format indicator
  • <schema ID> is the hexadecimal representation of GUID, same format and byte order as the string from the Schema Registry service.

If EventData is passed in as the message type, the following properties will be set on the EventData object:

• The body property will be set to the encoded content value.

• The content_type property will be set to the content type value.

If message type is not provided, and by default, the encoder will create the following dict: {"content": <encoded payload>, "content_type": 'application/json;serialization=Json+<schema ID>'}

Examples

The following sections provide several code snippets covering some of the most common Schema Registry and Json Schema Encoder tasks, including:

• Register a schema

• Get the schema by id

• Get the schema by version

• Get the id of a schema

• Encode

• Decode

• Event Hubs Send Integration

• Event Hubs Receive Integration

Register a schema

Use SchemaRegistryClient.register_schema method to register a schema.

import os

from azure.identity import DefaultAzureCredential
from azure.schemaregistry import SchemaRegistryClient

token_credential = DefaultAzureCredential()
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_AVRO_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ['SCHEMA_REGISTRY_GROUP']
name = "your-schema-name"
format = "Avro"
definition = """
{"namespace": "example.avro",
 "type": "record",
 "name": "User",
 "fields": [
     {"name": "name", "type": "string"},
     {"name": "favorite_number",  "type": ["int", "null"]},
     {"name": "favorite_color", "type": ["string", "null"]}
 ]
}
"""

schema_registry_client = SchemaRegistryClient(fully_qualified_namespace=fully_qualified_namespace, credential=token_credential)
with schema_registry_client:
    schema_properties = schema_registry_client.register_schema(group_name, name, definition, format)
    id = schema_properties.id

Get the schema by id

Get the schema definition and its properties by schema id.

import os

from azure.identity import DefaultAzureCredential
from azure.schemaregistry import SchemaRegistryClient

token_credential = DefaultAzureCredential()
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_AVRO_FULLY_QUALIFIED_NAMESPACE']
schema_id = 'your-schema-id'

schema_registry_client = SchemaRegistryClient(fully_qualified_namespace=fully_qualified_namespace, credential=token_credential)
with schema_registry_client:
    schema = schema_registry_client.get_schema(schema_id)
    definition = schema.definition
    properties = schema.properties

Get the schema by version

Get the schema definition and its properties by schema version.

import os

from azure.identity import DefaultAzureCredential
from azure.schemaregistry import SchemaRegistryClient

token_credential = DefaultAzureCredential()
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_AVRO_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ["SCHEMAREGISTRY_GROUP"]
name = "your-schema-name"
version = int("<your schema version>")

schema_registry_client = SchemaRegistryClient(fully_qualified_namespace=fully_qualified_namespace, credential=token_credential)
with schema_registry_client:
    schema = schema_registry_client.get_schema(group_name=group_name, name=name, version=version)
    definition = schema.definition
    properties = schema.properties

Get the id of a schema

Get the schema id of a schema by schema definition and its properties.

import os

from azure.identity import DefaultAzureCredential
from azure.schemaregistry import SchemaRegistryClient

token_credential = DefaultAzureCredential()
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_AVRO_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ['SCHEMA_REGISTRY_GROUP']
name = "your-schema-name"
format = "Avro"
definition = """
{"namespace": "example.avro",
 "type": "record",
 "name": "User",
 "fields": [
     {"name": "name", "type": "string"},
     {"name": "favorite_number",  "type": ["int", "null"]},
     {"name": "favorite_color", "type": ["string", "null"]}
 ]
}
"""

schema_registry_client = SchemaRegistryClient(fully_qualified_namespace=fully_qualified_namespace, credential=token_credential)
with schema_registry_client:
    schema_properties = schema_registry_client.register_schema(group_name, name, definition, format)
    id = schema_properties.id

Encode

Use the SchemaRegistryClient to pre-register the schema. Encode and validate the content with the JsonSchemaEncoder.

The encode method automatically retrieves the schema from the Schema Registry Service, validates against the content, and caches the schema locally.

import os
from azure.schemaregistry import SchemaRegistryClient, SchemaFormat
from azure.schemaregistry.encoder.jsonencoder import JsonSchemaEncoder, JsonSchemaDraftIdentifier
from azure.identity import DefaultAzureCredential
from azure.eventhub import EventData

token_credential = DefaultAzureCredential()
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_JSON_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ['SCHEMAREGISTRY_GROUP']
format = SchemaFormat.JSON

schema = {
    "$id": "https://example.com/person.schema.json",
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "title": "Person",
    "type": "object",
    "properties": {
        "name": {
            "type": "string",
            "description": "Person's name."
        },
        "favorite_color": {
            "type": "string",
            "description": "Favorite color."
        },
        "favorite_number": {
            "description": "Favorite number.",
            "type": "integer",
        }
    }
}
name = schema["title"]
definition = json.dumps(schema)

schema_registry_client = SchemaRegistryClient(fully_qualified_namespace, token_credential)
schema_properties = schema_registry_client.register_schema(group_name, name, definition, format)
schema_id = schema_properties.id

# group_name only needed if passing `schema` to encode
encoder = JsonSchemaEncoder(client=schema_registry_client, validate=JsonSchemaDraftIdentifier.DRAFT2020_12, group_name=group_name)

with encoder:
    dict_content = {"name": "Ben", "favorite_number": 7, "favorite_color": "red"}
    event_data = encoder.encode(dict_content, schema_id=schema_id, message_type=EventData)

    # OR

    message_content_dict = encoder.encode(dict_content, schema_id=schema_id)
    event_data = EventData.from_message_content(message_content_dict["content"], message_content_dict["content_type"])

    # OR

    dict_content = {"name": "Ben", "favorite_number": 7, "favorite_color": "red"}
    message_content = encoder.encode(dict_content, schema=definition)  # group_name required in constructor when `schema` is passed

Decode

Decode the content with the JsonSchemaEncoder.

The decode method automatically retrieves the schema from the Schema Registry Service, validates against the content, and caches the schema locally.

import os
from azure.schemaregistry import SchemaRegistryClient
from azure.schemaregistry.encoder.jsonencoder import JsonSchemaEncoder, JsonSchemaDraftIdentifier
from azure.identity import DefaultAzureCredential

token_credential = DefaultAzureCredential()
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ["SCHEMAREGISTRY_GROUP"]

schema_registry_client = SchemaRegistryClient(fully_qualified_namespace, token_credential)
encoder = JsonSchemaEncoder(client=schema_registry_client, validate=JsonSchemaDraftIdentifier.DRAFT2020_12)

with encoder:
    # event_data is an EventData object with encoded body
    decoded_content = encoder.decode(event_data)

    # OR 

    # content_dict is a TypedDict with encoded content and JSON content type 
    decoded_content = encoder.decode(content_dict)

Event Hubs Send Integration

Integration with Event Hubs to send an EventData object with body set to encoded content and corresponding content_type.

import os
from azure.eventhub import EventHubProducerClient, EventData
from azure.schemaregistry import SchemaRegistryClient
from azure.schemaregistry.encoder.jsonencoder import JsonSchemaEncoder, JsonSchemaDraftIdentifier
from azure.identity import DefaultAzureCredential

token_credential = DefaultAzureCredential()
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_JSON_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ['SCHEMAREGISTRY_GROUP']
eventhub_connection_str = os.environ['EVENT_HUB_CONN_STR']
eventhub_name = os.environ['EVENT_HUB_NAME']

schema_id = os.environ['PERSON_JSON_SCHEMA_ID']

schema_registry_client = SchemaRegistryClient(fully_qualified_namespace, token_credential)
json_schema_encoder = JsonSchemaEncoder(client=schema_registry_client, validate=JsonSchemaDraftIdentifier.DRAFT2020_12)

eventhub_producer = EventHubProducerClient.from_connection_string(
    conn_str=eventhub_connection_str,
    eventhub_name=eventhub_name
)

with eventhub_producer, json_schema_encoder:
    event_data_batch = eventhub_producer.create_batch()
    dict_content = {"name": "Bob", "favorite_number": 7, "favorite_color": "red"}
    event_data = json_schema_encoder.encode(dict_content, schema_id=schema_id, message_type=EventData)
    event_data_batch.add(event_data)
    eventhub_producer.send_batch(event_data_batch)

Event Hubs Receive Integration

Integration with Event Hubs to receive an EventData object and decode the encoded body value.

import os
from azure.eventhub import EventHubConsumerClient
from azure.schemaregistry import SchemaRegistryClient
from azure.schemaregistry.encoder.jsonencoder import JsonSchemaEncoder, JsonSchemaDraftIdentifier
from azure.identity import DefaultAzureCredential

token_credential = DefaultAzureCredential()
fully_qualified_namespace = os.environ['SCHEMAREGISTRY_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ['SCHEMAREGISTRY_GROUP']
eventhub_connection_str = os.environ['EVENT_HUB_CONN_STR']
eventhub_name = os.environ['EVENT_HUB_NAME']

schema_registry_client = SchemaRegistryClient(fully_qualified_namespace, token_credential)
json_schema_encoder = JsonSchemaEncoder(client=schema_registry_client, validate=JsonSchemaDraftIdentifier.DRAFT2020_12)

eventhub_consumer = EventHubConsumerClient.from_connection_string(
    conn_str=eventhub_connection_str,
    consumer_group='$Default',
    eventhub_name=eventhub_name,
)

def on_event(partition_context, event):
    decoded_content = json_schema_encoder.decode(event)

with eventhub_consumer, json_schema_encoder:
    eventhub_consumer.receive(on_event=on_event, starting_position="-1")

Troubleshooting

General

Schema Registry clients raise exceptions defined in Azure Core if errors are encountered when communicating with the Schema Registry service.

Errors when encoding and decoding related to invalid content/content types will be raised as azure.schemaregistry.encoder.jsonencoder.InvalidContentError, where __cause__ will possibly contain an underlying exception.

Logging

This library uses the standard logging library for logging. Basic information about HTTP sessions (URLs, headers, etc.) is logged at INFO level.

Detailed DEBUG level logging, including request/response bodies and unredacted headers, can be enabled on a client with the logging_enable argument:

import sys
import os
import logging
from azure.schemaregistry import SchemaRegistryClient
from azure.schemaregistry.encoder.jsonencoder import JsonSchemaEncoder, JsonSchemaDraftIdentifier
from azure.identity import DefaultAzureCredential

# Create a logger for the SDK
logger = logging.getLogger('azure.schemaregistry')
logger.setLevel(logging.DEBUG)

# Configure a console output
handler = logging.StreamHandler(stream=sys.stdout)
logger.addHandler(handler)

fully_qualified_namespace = os.environ['SCHEMAREGISTRY_FULLY_QUALIFIED_NAMESPACE']
group_name = os.environ['SCHEMAREGISTRY_GROUP']
credential = DefaultAzureCredential()
# This client will log detailed information about its HTTP sessions, at DEBUG level
schema_registry_client = SchemaRegistryClient(fully_qualified_namespace, credential, logging_enable=True)
encoder = JsonSchemaEncoder(client=schema_registry_client, validate=JsonSchemaDraftIdentifier.DRAFT2020_12)

Similarly, logging_enable can enable detailed logging for a single operation, even when it isn't enabled for the client:

schema_registry_client.get_schema(schema_id, logging_enable=True)

Next steps

More sample code

Please take a look at the samples directory for detailed examples of how to use this library to register and retrieve schema to/from Schema Registry.

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.microsoft.com.

When you submit a pull request, a CLA-bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., label, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.

Release History

1.3.0b3 (2023-11-09)

Features Added

• V2023_07_01 has been added to ApiVersion and set as the default api version.
  • Protobuf has been added to supported formats in SchemaFormat.

Other Changes

• Added support for Python 3.12.

1.3.0b2 (2023-08-09)

Features Added

The following features are experimental and may be removed:

• Sync and async JsonSchemaEncoder have been added under azure.schemaregistry.encoder.jsonencoder.
• InvalidContentError and JsonSchemaDraftIdentifier have been added under azure.schemaregistry.encoder.jsonencoder for use with the JsonSchemaEncoder.
• MessageType, MessageContent, SchemaContentValidate, SchemaEncoder have been added under azure.schemaregistry as protocols to define/for use with the JsonSchemaEncoder and future encoder implementations.

1.3.0b1 (2023-01-12)

Features Added

• V2022_10 has been added to ApiVersion and set as the default api version.
  • Json and Custom have been added to supported formats in SchemaFormat.
  • At the time of this release, only Draft 3 of JSON schemas is currently supported by the service.

Bugs Fixed

• Fixed a bug in sync/async register_schema and get_schema_properties that did not accept case insensitive strings as an argument to the format parameter.

Other Changes

• Added support for Python 3.11.

1.2.0 (2022-10-10)

This version and all future versions will require Python 3.7+, Python 3.6 is no longer supported.

Features Added

• group_name, name, and version have been added as optional parameters to the get_schema method on the sync and async SchemaRegistryClient.
• version has been added to SchemaProperties.

Other Changes

• Updated azure-core minimum dependency to 1.24.0.
• Added distributed tracing support for sync and async SchemaRegistryClient.

1.1.0 (2022-05-10)

This version and all future versions will require Python 3.6+. Python 2.7 is no longer supported.

Features Added

• group_name and name have been added as instance variables to SchemaProperties.

Other Changes

• Updated azure-core minimum dependency to 1.23.0.

1.0.0 (2021-11-10)

Note: This is the first stable release of our efforts to create a user-friendly and Pythonic client library for Azure Schema Registry.

Features Added

• SchemaRegistryClient is the top-level client class interacting with the Azure Schema Registry Service. It provides three methods:
  • register_schema: Store schema in the service by providing schema group name, schema name, schema definition, and schema format.
  • get_schema: Get schema definition and its properties by schema id.
  • get_schema_properties: Get schema properties by providing schema group name, schema name, schema definition, and schema format.
• SchemaProperties has the following instance variables: id and format:
  • The type of format has been changed from str to SchemaFormat.
• Schema has the following properties: properties and definition.
• SchemaFormat provides the schema format to be stored by the service. Currently, the only supported format is Avro.
• api_version has been added as a keyword arg to the sync and async SchemaRegistryClient constructors.

Breaking Changes

• version instance variable in SchemaProperties has been removed.
• schema_definition instance variable in Schema has been renamed definition.
• id parameter in get_schema method on sync and async SchemaRegistryClient has been renamed schema_id.
• schema_definition parameter in register_schema and get_schema_properties methods on sync and async SchemaRegistryClient has been renamed definition.
• serializer namespace has been removed from azure.schemaregistry.

1.0.0b3 (2021-10-05)

Breaking Changes

• get_schema_id method on sync and async SchemaRegistryClient has been renamed get_schema_properties.
• schema_id parameter in get_schema method on sync and async SchemaRegistryClient has been renamed id.
• register_schema and get_schema_properties methods on sync and async SchemaRegistryClient now take in the following parameters in the given order:
  • group_name, which has been renamed from schema_group
  • name, which has been renamed from schema_name
  • schema_definition, which has been renamed from schema_content
  • format, which has been renamed from serialization_type
• endpoint parameter in SchemaRegistryClient constructor has been renamed fully_qualified_namespace
• location instance variable in SchemaProperties has been removed.
• Schema and SchemaProperties no longer have positional parameters, as they will not be constructed by the user.

Other Changes

• Updated azure-core dependency to 1.19.0.
• Removed caching support of registered schemas so requests are sent to the service to register schemas, get schema properties, and get schemas.

1.0.0b2 (2021-08-17)

This version and all future versions will require Python 2.7 or Python 3.6+, Python 3.5 is no longer supported.

Features Added

• Support caching of registered schemas and send requests to the service only if the cache does not have the looked-up schema/schema ID.

1.0.0b1 (2020-09-09)

Version 1.0.0b1 is the first preview of our efforts to create a user-friendly and Pythonic client library for Azure Schema Registry.

New features

• SchemaRegistryClient is the top-level client class interacting with the Azure Schema Registry Service. It provides three methods:
  • register_schema: Store schema into the service.
  • get_schema: Get schema content and its properties by schema id.
  • get_schema_id: Get schema id and its properties by schema group, schema name, serialization type and schema content.