deployramp 0.1.7

DeployRamp SDK - Feature flag management for Python

DeployRamp Python SDK

Python SDK for DeployRamp — AI-native feature flag management with gradual rollouts, real-time updates, and automatic error-monitored rollbacks.

Installation

pip install deployramp

Quick Start

import deployramp

deployramp.init(deployramp.SdkConfig(
    public_token="drp_pub_your_token",
    traits={"plan": "pro", "region": "us-east"},
))

# Evaluate a feature flag
if deployramp.flag("new-checkout-flow"):
    process_new_checkout()
else:
    process_old_checkout()

# Report errors — DeployRamp uses these to auto-roll back bad deploys
try:
    process_checkout()
except Exception as e:
    deployramp.report(e, flag_name="new-checkout-flow")

deployramp.close()

Trait-Based Targeting

import deployramp

deployramp.init(deployramp.SdkConfig(public_token="drp_pub_your_token"))

# Update traits after login
deployramp.set_traits({"plan": "enterprise", "cohort": "beta"})

# Override traits for a single evaluation
enabled = deployramp.flag("beta-feature", trait_overrides={"cohort": "alpha"})

Measure Performance

Track how each flag branch performs across your user population:

import deployramp

result = deployramp.measure(
    "fast-algorithm",
    enabled_fn=lambda: new_algorithm(data),
    disabled_fn=lambda: old_algorithm(data),
)

API Reference

init(config: SdkConfig) -> None

Initialize the SDK. Fetches flags and opens a WebSocket for real-time updates.

flag(name: str, trait_overrides: dict | None = None) -> bool

Evaluate a feature flag. Returns False if the SDK is not initialized or the flag is unknown/disabled.

set_traits(traits: dict[str, str]) -> None

Replace the current traits used for flag evaluation.

measure(name: str, enabled_fn, disabled_fn, trait_overrides: dict | None = None)

Run the appropriate branch and report execution time to DeployRamp.

report(error: Exception | str, flag_name: str | None = None, trait_overrides: dict | None = None) -> None

Report an error to DeployRamp for automatic rollback monitoring.

close() -> None

Flush pending evaluations and close connections.

SdkConfig

@dataclass
class SdkConfig:
    public_token: str                      # Required. Your public token from deployramp.com
    base_url: str = "https://flags.deployramp.com"
    traits: dict[str, str] = field(default_factory=dict)

Requirements

• Python 3.10+
• websockets >= 12.0

Links

• deployramp.com
• GitHub
• PyPI

License

MIT