fix: wire AgentGraph to YAML config

[bhavyakeerthi3]

Description

This PR addresses the long-standing issue where AgentGraph hardcoded the LLM and embedding models, ignoring any settings in config.yml.
First reported in #108, this completes the YAML configurability work by extending it to the most critical parameters.

Closes #108, Closes #148

Changes

• Pydantic Model: Added llm and embedding fields to the Config class in src/util/config_yml/__init__.py.
• Schema & Defaults: Updated .config.schema.yaml and config_default.yml to support these fields with safe defaults (gpt-4o-mini, text-embedding-3-large).
• Graph Wiring: Updated the AgentGraph constructor to accept and use these configurations instead of hardcoded strings.
• Entry Point: Updated chat-chainlit.py to correctly pass the YAML-loaded configurations to the agent during startup.

Impact

The chatbot is now fully configurable. Users can switch to Claude, Gemini, or local Ollama models via config.yml without touching Python source code — making the project significantly more accessible to contributors and deployers.

Tests Added

Added test_config_loads_llm_and_embedding to tests/test_config.py:

def test_config_loads_llm_and_embedding(mock_config_file):
    config = Config.from_yaml(mock_config_file)
    assert config.llm == "gpt-4o-mini"
    assert config.embedding == "text-embedding-3-large"

Verification

Verified that AgentGraph correctly receives and uses parameters from config_default.yml on startup.

Files Affected

• src/util/config_yml/__init__.py
• config_default.yml
• .config.schema.yaml
• src/agent/graph.py
• bin/chat-chainlit.py
• tests/test_config.py

Relation to Other PRs

Completes configurability work started in #112 by AaryanCode69, extending it to cover LLM and embedding model parameters which were not included in that PR.