fix(mcp): reasoning.effort 'minimal' incompatible with gpt-5.4-mini

[daniel-rudaev]

Summary

The MCP server factory (factories.py:136) hardcodes reasoning='minimal' when creating OpenAIClient for reasoning models. However, newer OpenAI models (gpt-5.4-mini, gpt-5.4, etc.) reject 'minimal' — they only accept 'none', 'low', 'medium', 'high', and 'xhigh'.

This causes all entity extraction to fail with a 400 error when using any gpt-5.4 series model:

Error code: 400 - {'error': {'message': "Unsupported value: 'minimal' is not supported 
with the 'gpt-5.4-mini' model.", 'param': 'reasoning.effort'}}

Related: #902

Changes

• mcp_server/src/services/factories.py — change reasoning='minimal' to reasoning='low'
• mcp_server/docker/Dockerfile.standalone — patch graphiti-core's DEFAULT_REASONING constant from 'minimal' to 'low' (defense in depth for the same value in the pip package), recompile bytecode after patching

Test plan

• Verify gpt-5.4-mini entity extraction works without 400 errors
• Verify gpt-5-mini (default model) still works with reasoning='low'
• Verify non-reasoning models (gpt-4o-mini etc.) still pass reasoning=None

🤖 Generated with Claude Code

[danielchalef]

Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

---

I have read the CLA Document and I hereby sign the CLA

---

_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}

[daniel-rudaev]

I have read the CLA Document and I hereby sign the CLA