feat: add Tavily search engine support to core search utility

[ParakhJaggi]

Summary

• Added Tavily as a fifth search engine option in research_web.py, alongside existing duckduckgo, bing, searxng, and serper backends
• All existing providers remain unchanged (additive migration)
• Tavily API key can be passed directly or read from TAVILY_API_KEY environment variable

Files changed

• scrapegraphai/utils/research_web.py — Added tavily to valid engines, tavily_api_key field to SearchConfig, tavily_api_key parameter to search_on_web(), new _search_tavily() helper function, and dispatch branch
• pyproject.toml — Added tavily-python>=0.5.0 to dependencies

Dependency changes

• Added tavily-python>=0.5.0 to pyproject.toml core dependencies

Environment variable changes

• Added TAVILY_API_KEY — required when using search_engine="tavily" (can also be passed as tavily_api_key parameter)

Notes for reviewers

• The implementation follows the same pattern as the existing _search_serper() helper (API key validation, SDK call, URL extraction)
• Falls back to TAVILY_API_KEY env var if no explicit key is provided
• Includes ImportError handling for missing tavily-python package

🤖 Generated with Claude Code

Automated Review

• Passed after 1 attempt(s)
• Final review: The Tavily migration is correct, complete, and additive with no regressions. SDK usage patterns are valid, the dispatch logic follows the existing structure, and both pyproject.toml and research_web.py are updated consistently. Two minor issues noted but neither blocks approval.