Refactor email agent, add more tools, cleanup code

[gautamp8]

This PR adds new tools like Wikipedia api, refactor and improve existing ones(search tools with fallback). It cleans up the code in email agent initialisation. Reduces logging verbosity.

[Copilot]

Pull Request Overview

Refactors the email agent initialization to use a unified fallback search tool, adds Wikipedia search support, reduces logging verbosity, and cleans up related tooling code.

• Adds a new SearchWithFallbackTool and removes the old FallbackWebSearchTool.
• Introduces Wikipedia API dependency and tool.
• Encapsulates search and research tool setup into helper methods and converts many info logs to debug.

Reviewed Changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
pyproject.toml	Added wikipedia-api dependency.
mxtoai/tools/search_with_fallback_tool.py	New generic fallback search tool combining multiple engines.
mxtoai/tools/fallback_search_tool.py	Removed deprecated fallback search implementation.
mxtoai/tools/init.py	Replaced export of old tool with SearchWithFallbackTool.
mxtoai/routed_litellm_model.py	Cleaned up comments and streamlined logging.
mxtoai/agents/email_agent.py	Refactored tool initialization into _initialize_* helpers, added Wikipedia and web search tools, and reduced log verbosity.

Comments suppressed due to low confidence (2)

mxtoai/tools/search_with_fallback_tool.py:14

• The tool name web_search duplicates the name used by individual WebSearchTool instances, which may confuse consumers. Consider renaming this to search_with_fallback or similar to differentiate it.

name = "web_search"  # Consistent name for the agent to use

mxtoai/tools/search_with_fallback_tool.py:8

• The new SearchWithFallbackTool class lacks accompanying unit tests. Consider adding tests for success and failure scenarios of both primary and fallback search engines.

class SearchWithFallbackTool(Tool):