redis-enterprise Python bindings

Python bindings for the Redis Enterprise REST API, built with PyO3 on top of the redis-enterprise Rust crate.

Scope and philosophy

Goals

• Convenience, not parity. The Python layer exposes the most common read and inspection operations — cluster info, database listing, node status, user listing — plus raw HTTP access for anything else.
• Sync and async. Every named method ships as both a blocking _sync variant and an async/await coroutine so it fits naturally in scripts, notebooks, or async applications.
• Plain Python dicts. All responses are deserialized to Python dict/list (via serde_json). No custom model classes to learn; use SimpleNamespace, pydantic, or dataclasses as you prefer.
• Zero required event loop. The _sync variants drive a dedicated Tokio runtime internally, so they work from plain Python scripts or Jupyter notebooks without any asyncio setup.

Non-goals

• Full API parity with the Rust crate. Write operations (database create/update/ delete), CRDB management, module uploads, and so on are outside the initial scope.
• Typed response models. Raw dicts keep the binding thin and avoid a maintenance burden on the Python side.
• Re-exporting every Rust builder knob. For advanced configuration use the Rust crate directly.

Installation

pip install redis-enterprise

Requires Python ≥ 3.9. Wheels are published for Linux (x86_64, aarch64), macOS (x86_64, arm64), and Windows (x86_64) via the publish-python CI workflow.

Quick start

from redis_enterprise import EnterpriseClient

# Explicit configuration
client = EnterpriseClient(
    base_url="https://cluster.example.com:9443",
    username="admin@redis.local",
    password="secret",
    insecure=True,       # skip TLS verification — dev/lab only
    timeout_secs=30,     # optional; omit to use the default
)

# Or let the client read credentials from environment variables
client = EnterpriseClient.from_env()

# ── Synchronous ──────────────────────────────────────────────
cluster = client.cluster_info_sync()
print(cluster["name"])

dbs = client.databases_sync()
for db in dbs:
    print(db["name"], db["uid"])

# ── Async ────────────────────────────────────────────────────
import asyncio

async def main():
    cluster = await client.cluster_info()
    print(cluster["name"])

asyncio.run(main())

API reference

All methods return plain Python dict or list[dict].

Constructor

Cluster

Databases

Nodes

Users

Raw HTTP access

Use the raw pass-through methods for any endpoint not yet covered by a named method. Paths are relative to the cluster base URL (e.g. "/v1/bdbs").

# GET
roles = client.get_sync("/v1/roles")

# POST — body is a plain Python dict
new_role = client.post_sync(
    "/v1/roles",
    {"name": "ops-reader", "management": "db_viewer"},
)

# DELETE
client.delete_sync(f"/v1/roles/{role_id}")

# Async equivalents
roles = await client.get("/v1/roles")
new_role = await client.post(
    "/v1/roles",
    {"name": "ops-reader", "management": "db_viewer"},
)
await client.delete(f"/v1/roles/{role_id}")

Error handling

from redis_enterprise import EnterpriseClient, RedisEnterpriseError

try:
    db = client.database_sync(9999)
except ValueError:
    print("Database not found")
except RedisEnterpriseError as e:
    print(f"API error: {e}")
except ConnectionError as e:
    print(f"Could not reach cluster: {e}")

RedisEnterpriseError is a custom exception class exported by the module and is a subclass of Exception.

Planned expansion — first tranche

The following operations are targeted for the next minor release:

Mutation operations (database create/update/delete, user management) are under consideration for a subsequent tranche but are not committed.

Testing strategy

Unit tests live in python/tests/ and use pytest + pytest-asyncio. Integration tests require a live Redis Enterprise cluster; see the live-validation runbook for standing one up with Docker Compose.

# Install dev extras
pip install -e "python/[dev]"

# Unit tests (no cluster required)
pytest python/tests/unit/

# Integration tests (cluster required — see docs/live-validation.md)
pytest python/tests/integration/

Release and packaging

The publish-python GitHub Actions workflow fires on version tags (v*). It uses maturin to compile the PyO3 extension module for each supported platform and architecture, then publishes the wheels and an sdist to PyPI.

The Python package version is derived from python/Cargo.toml, which is kept in sync with the root crate version.

Environment variables

These variables are read by EnterpriseClient.from_env():