fix: Use configured LLM model in categorization instead of hardcoded gpt-4o-mini

[ron-42]

Description

This PR fixes a critical issue in OpenMemory's categorization function where the model was hardcoded to gpt-4o-mini, causing failures when users configured alternative LLM providers like SiliconFlow.

The Problem:

• The categorization function in openmemory/api/app/utils/categorization.py used a hardcoded gpt-4o-mini model
• Users configuring SiliconFlow or other OpenAI-compatible providers with different models (e.g., deepseek-ai/DeepSeek-R1) received "Model does not exist" errors
• The categorization logic was disconnected from user's LLM configuration

The Solution:

• Modified categorization to dynamically read LLM configuration from the database
• Added support for custom openai_base_url parameter (enables SiliconFlow .cn domain and other custom endpoints)
• Created SiliconFlowConfig class for proper type-safe configuration
• Updated LlmFactory to use SiliconFlowConfig instead of generic BaseLlmConfig
• Added comprehensive documentation for SiliconFlow usage with code examples

Dependencies:

• Existing openai Python package (no new dependencies)

Fixes #3576

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Refactor (does not change functionality, e.g. code style improvements, linting)
• Documentation update

How Has This Been Tested?

Manual Testing:

1. Configured OpenMemory with SiliconFlow provider:

   {
     "mem0": {
       "llm": {
         "provider": "openai",
         "config": {
           "model": "deepseek-ai/DeepSeek-R1",
           "openai_base_url": "https://api.siliconflow.cn/v1",
           "api_key": "sk-..."
         }
       }
     }
   }

[CLAassistant]

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.

---

Ronak Bhalgami seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}