Add DaoXE OpenAI-compatible AsyncOpenAI example (Python) - #560

Summary

Adds examples/python/daoxe-openai-compat showing AsyncZep agent memory with AsyncOpenAI pointed at DaoXE:

from openai import AsyncOpenAI

client = AsyncOpenAI(
    api_key=os.environ["DAOXE_API_KEY"],
    base_url="https://daoxe.com/v1",
)

Why this fits the examples tree

• Mirrors existing OpenAI-style Zep samples (AsyncOpenAI + thread memory / context injection).
• DaoXE is a multi-model, multi-protocol API gateway. This PR documents the OpenAI Chat Completions path only (same migration shape as other OpenAI-compatible backends). DaoXE also supports OpenAI Responses and Anthropic Messages / Claude protocol — those are out of scope for this folder.
• Model IDs are account-scoped via DAOXE_MODEL (no hardcoded public model list).
• Availability: DaoXE is not offered in mainland China; README notes regional guidance without hardcoding alternate hosts.

Files

I maintain DaoXE. Related snippets: https://github.com/seven7763/DaoXE-AI

Test plan

• cd examples/python/daoxe-openai-compat && python -m venv .venv && source .venv/bin/activate && pip install -r requirements.txt
• cp .env.example .env and set ZEP_API_KEY, DAOXE_API_KEY, DAOXE_MODEL (exact account catalog ID)
• python daoxe_zep_memory.py
• Confirm streaming assistant reply references seeded Lisbon / walking-tour context when memory is available
• Confirm no secrets committed (env-only keys/models)