Add async support using separate files (non-breaking)

[johanzander]

Summary

• Adds async API classes as separate files alongside existing sync code (OpenAI SDK pattern)
• Sync code is completely untouched — requests remains the sync HTTP library
• Async classes use httpx as an optional dependency (pip install growattServer[async])
• Lazy imports via __getattr__ so httpx is never required unless async classes are used

New files

File	Class	Async copy of
growattServer/async_base_api.py	AsyncGrowattApi	GrowattApi
growattServer/open_api_v1/async_open_api_v1.py	AsyncOpenApiV1	OpenApiV1
growattServer/open_api_v1/devices/async_min.py	AsyncMin	Min
growattServer/open_api_v1/devices/async_sph.py	AsyncSph	Sph

Usage

from growattServer import AsyncOpenApiV1

async with AsyncOpenApiV1(token="your_token") as api:
    plants = await api.plant_list()

Trade-offs vs #150

	This PR (separate files)	#150 (shared base + httpx)
Breaking change	No	Yes (requests to httpx)
Code duplication	Yes (~4 files duplicated)	No
Complexity	Low (straightforward copies)	Higher (coroutine passthrough)
Sync HTTP library	requests (unchanged)	httpx

This is an alternative to #150 for discussion on #149. The main advantage is that it is non-breaking — existing consumers (including Home Assistant) continue to work without any changes.

Test plan

• Verify sync imports still work without httpx installed
• Verify async imports work with httpx installed
• Verify AsyncGrowattApi and AsyncOpenApiV1 instantiate correctly
• Verify get_device() returns AsyncMin/AsyncSph correctly
• Integration test with real Growatt API (manual)

🤖 Generated with Claude Code