Harness FME Python Configs SDK

Overview

Harness FME Python Configs SDK is designed to serve standalone, versioned configuration objects in Python applications.

Compatibility

This SDK is compatible with Python 3.7 and higher.

Getting started

Below is a simple example that describes the instantiation and most basic usage of our SDK:

Run pip install splitio_configs_client

from splitio_configs import ConfigsClient, TimeoutException, Target

client = ConfigsClient('YOUR_SDK_KEY')
try:
    client.block_until_ready()
except TimeoutException:
    # SDK_READY_TIMED_OUT: evaluations will return the fallback config or a Config with variant "control".
    pass

# Without target: returns the default variant of the config
default_config = client.get_config('CONFIG_NAME')
print(default_config.variant)  # example: 'v1'
print(default_config.value)    # example: {'prompt': 'Hello world', 'temperature': 0.5}

# With target: evaluates targeting rules for the given key
target = Target('user-123')
config = client.get_config('CONFIG_NAME', target)
print(config.variant, config.value)

The SDK also support asyncio, see below example:

from splitio_configs import ConfigsClientAsync, TimeoutException, Target

async def main():
    client = ConfigsClientAsync('YOUR_SDK_KEY')
    try:
        await client.block_until_ready()
    except TimeoutException:
        # SDK_READY_TIMED_OUT: evaluations will return the fallback config or a Config with variant "control".
        pass

    # Without target: returns the default variant of the config
    default_config = await client.get_config('CONFIG_NAME')
    print(default_config.variant)  # example: 'v1'
    print(default_config.value)    # example: {'prompt': 'Hello world', 'temperature': 0.5}

    # With target: evaluates targeting rules for the given key
    target = Target('user-123')
    config = await client.get_config('CONFIG_NAME', target)
    print(config.variant, config.value)

asyncio.run(main())

AI Configs

SDK supoprt get_ai_config call to fetch the AI agents configurations, the ai data is parsed through a built-in parser, however, the SDK allows to pass a custom parser.

Example below using the built-in parser

from splitio_configs import ConfigsClient, TimeoutException, Target, LlmCallParser

client = ConfigsClient('YOUR_SDK_KEY')
try:
    client.block_until_ready()
except TimeoutException:
    # SDK_READY_TIMED_OUT: evaluations will return the fallback config or a Config with variant "control".
    pass

parser = LlmCallParser()

# With target: evaluates targeting rules for the given key
target = Target('user-123')
ai_config = client.get_ai_config('CONFIG_NAME', parser, target)
# config standard data
print(ai_config.variant, ai_config.value)
# ai config data
print(ai_config.parsed_value.provider)
print(ai_config.parsed_value.model)
print(ai_config.parsed_value.system_prompt)
print(ai_config.parsed_value.user_prompt)
print(ai_config.parsed_value.max_tokens)
print(ai_config.parsed_value.temperature)

Example below using a custom parser

from splitio_configs import ConfigsClient, TimeoutException, Target, Parse

class MyLlmCall(object):
    def __init__(self, model, temperature):
        self.model = model
        self.temperature = temperature

class MyParser(Parser):
    def parse(self, raw):
        return MyLlmCall(raw.get("model"), raw.get("temperature"))

client = ConfigsClient('YOUR_SDK_KEY')
try:
    client.block_until_ready()
except TimeoutException:
    # SDK_READY_TIMED_OUT: evaluations will return the fallback config or a Config with variant "control".
    pass

my_parser = MyParser()

# With target: evaluates targeting rules for the given key
target = Target('user-123')
ai_config = client.get_ai_config('CONFIG_NAME', my_parser, target)
# config standard data
print(ai_config.variant, ai_config.value)
# ai config data
print(ai_config.parsed_value.model)
print(ai_config.parsed_value.temperature)

Please refer to our official docs to learn about all the functionality provided by our SDK and the configuration options available for tailoring it to your current application setup.

Submitting issues

The Split team monitors all issues submitted to this issue tracker. We encourage you to use this issue tracker to submit any bug reports, feedback, and feature enhancements. We'll do our best to respond in a timely manner.

License

Licensed under the Apache License, Version 2.0. See: Apache License.