Determine additional stores to support

[strawgate]

Let's determine what high value stores we should support next

[github-actions]

Claude finished @strawgate's task in 3m 46s —— View job

---

Recommended Additional Stores

Based on industry adoption, Python async library maturity, and gaps in the current store offerings, here are my prioritized recommendations:

Tier 1: High Value (Strong async support, widely used)

Store	Use Case	Python Async Library	Notes
SQLite	Local/embedded storage, testing, prototyping	aiosqlite (mature, well-maintained)	Extremely popular for local apps; complements RocksDB for different use cases
Couchbase	Document/KV hybrid, enterprise caching	acouchbase (official, well-supported)	Major NoSQL player; full asyncio support via official SDK
etcd	Configuration, service discovery, distributed coordination	aetcd, aetcd3	Popular in Kubernetes/cloud-native ecosystems

Tier 2: Medium Value (Good async support, specialized use cases)

Store	Use Case	Python Async Library	Notes
Consul	Service mesh KV, configuration	aioconsul, python-consul.aio	HashiCorp ecosystem integration
Apache Ignite	Distributed caching, compute grid	ignite-python-thin-client (has async examples)	Enterprise-grade distributed cache
Azure Cosmos DB	Multi-model cloud database	azure-cosmos.aio (official async)	Complements existing cloud stores (DynamoDB, Firestore)
Google Cloud Bigtable	Wide-column store for analytics	BigtableDataClientAsync (official async)	Google's equivalent to DynamoDB

Tier 3: Lower Priority (Limited async support or niche)

Store	Use Case	Python Async Library	Notes
Cassandra/ScyllaDB	Wide-column, high-throughput	scyllapy, acsylla	Wide-column model is a stretch for pure KV semantics
TiKV	Distributed KV (CNCF)	client-py (proof-of-concept stage)	Rust-based, async support exists but immature
Hazelcast	In-memory data grid	hazelcast-python-client (Future-based async)	Java-centric ecosystem
LevelDB	Embedded KV engine	plyvel (sync only, GIL-releasing)	Would need run_in_executor wrapper
LMDB	Memory-mapped KV	No mature async library	Similar to LevelDB situation
FoundationDB	ACID distributed KV	asyncio-foundationdb (third-party)	Powerful but niche adoption

My Recommendation

Start with SQLite, Couchbase, and etcd. These three stores cover:

1. SQLite - Local/embedded use cases (development, testing, single-user apps). The aiosqlite library is mature and well-maintained. This fills a gap for users who want a simple, zero-config persistent store without running a server.

2. Couchbase - Enterprise NoSQL with official async support. It's a major player alongside MongoDB and has a well-documented async API.

3. etcd - Cloud-native configuration storage. Very popular in Kubernetes environments. Multiple async clients exist (aetcd, aetcd3).

For cloud completeness, Azure Cosmos DB and Google Cloud Bigtable would round out the major cloud provider offerings (you already have AWS DynamoDB/S3 and GCP Firestore).

Findings

Current Coverage Analysis

Well covered:

• AWS: DynamoDB, S3
• GCP: Firestore
• In-memory: Memory, Memcached
• Redis-compatible: Redis, Valkey
• Document stores: MongoDB, Elasticsearch, OpenSearch
• Secret management: Vault, Keyring
• Embedded: RocksDB, DuckDB, Disk

Gaps identified:

• Embedded SQL: No SQLite support despite being ubiquitous
• Cloud providers: Missing Azure Cosmos DB, Google Cloud Bigtable
• Service discovery: No etcd or Consul
• Enterprise NoSQL: No Couchbase (major competitor to MongoDB)

Python Async Library Maturity

Library	Maturity	Maintenance
aiosqlite	⭐⭐⭐ Excellent	Active (Dec 2025 release)
acouchbase	⭐⭐⭐ Excellent	Official Couchbase SDK
azure-cosmos.aio	⭐⭐⭐ Excellent	Official Microsoft SDK
google-cloud-bigtable async	⭐⭐⭐ Excellent	Official Google SDK
aetcd/aetcd3	⭐⭐ Good	Community maintained
aioconsul	⭐⭐ Good	Community maintained
scyllapy	⭐⭐ Good	Active development
tikv client-py	⭐ Early	Proof-of-concept

Detailed Action Plan

For SQLite Store Implementation

1. Add aiosqlite as optional dependency: pip install py-key-value-aio[sqlite]
2. Create src/key_value/aio/stores/sqlite/ with:
   • store.py - Main SQLiteStore implementation
   • __init__.py - Public exports
3. Store schema: Single table with key TEXT PRIMARY KEY, collection TEXT, value TEXT, created_at REAL, expires_at REAL
4. Implement TTL via expires_at column with cleanup on access or periodic culling
5. Add tests in tests/stores/sqlite/

For Couchbase Store Implementation

1. Add couchbase as optional dependency: pip install py-key-value-aio[couchbase]
2. Use acouchbase.cluster.AsyncCluster for connections
3. Map collections to Couchbase scopes/collections or use document key prefixes
4. TTL maps directly to Couchbase's document expiration
5. Handle CouchbaseException appropriately

For etcd Store Implementation

1. Add aetcd or aetcd3 as optional dependency
2. Use etcd's prefix-based key organization for collections
3. Consider lease-based TTL for automatic expiration
4. Handle etcd's revision-based consistency model

Related Items

Type	Link	Relevance
External	aiosqlite	Python asyncio SQLite library
External	Couchbase Python SDK Async	Official async documentation
External	aetcd documentation	etcd async client
External	Azure Cosmos DB async preview	Official async SDK announcement
External	DB-Engines Key-Value Ranking	Industry popularity metrics

Related Files

File	Relevance
src/key_value/aio/stores/	Directory for new store implementations
src/key_value/aio/stores/base.py	Base store class to extend
tests/stores/base.py	Test mixin patterns to follow
pyproject.toml	Add optional dependencies here

---

Why is Claude responding? | Type @claude to interact further

Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.

[strawgate]

@claude make a PR for cosmosdb please

[strawgate]

@claude make a PR for couch base please

[github-actions]

Claude finished @strawgate's task in 7m 18s —— View job • claude/issue-292-20260203-0843 • Create PR ➔

---

Implemented Azure Cosmos DB Store

• Gather context and understand existing store patterns
• Research Azure Cosmos DB async Python SDK
• Implement CosmosDB store
• Add tests for CosmosDB store
• Update pyproject.toml with optional dependency
• Run linting and type checking
• Create PR

Summary

I've implemented the Azure Cosmos DB store with full support for the AsyncKeyValue protocol. The implementation includes:

New Files:

• src/key_value/aio/stores/cosmosdb/store.py - Main store implementation
• src/key_value/aio/stores/cosmosdb/__init__.py - Public exports
• tests/stores/cosmosdb/test_cosmosdb.py - Test suite
• tests/stores/cosmosdb/__init__.py - Test package init

Features:

• Full CRUD operations (get, put, delete with single and batch variants)
• TTL support via Cosmos DB's native TTL feature
• Collection isolation using partition keys
• Collection sanitization strategy (CosmosDBV1CollectionSanitizationStrategy)
• Support for both client injection and URL/credential configuration
• Automatic database and container creation (configurable via auto_create)
• Proper context manager support with client lifecycle management

Usage:

from key_value.aio.stores.cosmosdb import CosmosDBStore

# Using URL and credential
async with CosmosDBStore(
    url="https://your-account.documents.azure.com:443/",
    credential="your-key",
    database_name="my-database",
    container_name="my-container",
) as store:
    await store.put(key="my-key", value={"data": "value"}, ttl=3600)
    value = await store.get(key="my-key")

Installation:

pip install py-key-value-aio[cosmosdb]

---

Why is Claude responding? | Type @claude to interact further

Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.
|

[github-actions]

Claude finished @strawgate's task in 9m 35s —— View job • claude/issue-292-20260203-0851 • Create PR ➔

---

Couchbase Store Implementation Complete

• Gather repository context and understand existing patterns
• Create Couchbase store implementation
• Add tests for Couchbase store
• Update pyproject.toml with optional dependency
• Run linting and type checking
• Commit and push changes
• Create PR

Summary

I've implemented a new CouchbaseStore backend using the official acouchbase async SDK.

Files created:

• src/key_value/aio/stores/couchbase/store.py - Main store implementation
• src/key_value/aio/stores/couchbase/__init__.py - Public exports
• tests/stores/couchbase/test_couchbase.py - Test suite
• tests/stores/couchbase/__init__.py - Test package init

Features:

• Full AsyncKeyValue protocol implementation
• TTL support via Couchbase's built-in document expiration
• CouchbaseV1CollectionSanitizationStrategy for safe collection names
• Support for both connection string and pre-configured cluster
• Proper resource management with async context manager

Install with: pip install py-key-value-aio[couchbase]

--- | Branch: claude/issue-292-20260203-0851

---

Why is Claude responding? | Type @claude to interact further

Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.