Python SDK

Status: In development

The Python SDK provides a high-level API for managing infrastructure with Senren.

Installation

pip install senren

Quick Start

from senren import SenrenClient, Database

# Connect to control plane
client = SenrenClient(endpoint="control.senren.dev")

# Define infrastructure
client.apply_state(
    databases=[
        Database(
            name="user-cache",
            type="redis-cluster",
            memory_mb=8192,
            regions=["aws:us-east-1", "gcp:us-central1"],
        )
    ]
)

# Query status
status = client.get_status("user-cache")
print(f"Ready: {status.ready}")

API Reference

SenrenClient

Constructor:

SenrenClient(
    endpoint: str,              # Control plane gRPC endpoint
    timeout_seconds: int = 30,  # Request timeout
)

Methods:

apply_state()

Apply desired infrastructure state.

client.apply_state(
    databases: list[Database] = [],
    # feature_stores: list[FeatureStore] = [],  # Coming soon
    # search_indexes: list[SearchIndex] = [],   # Coming soon
) -> ApplyResponse

Returns: - ApplyResponse with applied_version field

get_status()

Get status of a specific resource.

client.get_status(name: str) -> ResourceStatus

Returns: - ResourceStatus with fields: - ready: bool - Whether resource is ready - host: str - Connection host - port: int - Connection port - error: str | None - Error message if not ready

Database

Constructor:

Database(
    name: str,                  # Unique name
    type: str,                  # "redis" | "redis-cluster" | "dragonfly"
    memory_mb: int,             # Memory allocation in MB
    regions: list[str],         # List of region identifiers
    namespace: str = "default", # Kubernetes namespace
)

Region identifiers: - AWS: aws:us-east-1, aws:eu-west-1, etc. - GCP: gcp:us-central1, gcp:europe-west1, etc. - Multi-cluster: aws:us-east-1:prod, aws:us-east-1:shadow

ResourceStatus

Fields:

@dataclass
class ResourceStatus:
    ready: bool         # Is the resource ready?
    host: str          # Connection host
    port: int          # Connection port
    error: str | None  # Error message if not ready

Examples

Multi-region Deployment

from senren import SenrenClient, Database

client = SenrenClient(endpoint="control.senren.dev")

# Deploy to 3 regions
client.apply_state(
    databases=[
        Database(
            name="global-cache",
            type="redis-cluster",
            memory_mb=16384,
            regions=[
                "aws:us-east-1",
                "aws:eu-west-1",
                "gcp:asia-east1",
            ],
        )
    ]
)

Updating Configuration

# Increase memory
client.apply_state(
    databases=[
        Database(
            name="global-cache",
            type="redis-cluster",
            memory_mb=32768,  # Doubled
            regions=[
                "aws:us-east-1",
                "aws:eu-west-1",
                "gcp:asia-east1",
            ],
        )
    ]
)

Removing Resources

# Remove all databases by sending empty state
client.apply_state(databases=[])

Checking Status

import time

status = client.get_status("global-cache")

if status.ready:
    print(f"Database ready at {status.host}:{status.port}")
else:
    print(f"Database not ready: {status.error}")
    # Poll until ready
    while not status.ready:
        time.sleep(1)
        status = client.get_status("global-cache")

Error Handling

from senren import SenrenClient, SenrenError

try:
    client.apply_state(databases=[...])
except SenrenError as e:
    print(f"Failed to apply state: {e}")

Coming Soon

The following APIs are planned:

Feature Stores (Q2 2025)

from senren import FeatureStore

store = FeatureStore(
    name="user-features",
    type="online",
    backend="redis-cluster",
    schema={...},
    regions=[...],
)

Search Indexes (Q3 2025)

from senren import SearchIndex

index = SearchIndex(
    name="product-search",
    type="elasticsearch",
    memory_gb=16,
    regions=[...],
)

Model Serving (Q4 2025)

from senren import ModelServer

server = ModelServer(
    name="recommendation-model",
    type="torchserve",
    model_url="s3://...",
    regions=[...],
)

gRPC API

The Python SDK is a wrapper around the gRPC API. For advanced use cases or other languages, you can use the gRPC API directly (documentation coming soon).