feat: add x-exa-integration header and modernize ExaTools

[tgonzalezc5]

Summary

Small follow-up to PR #7099 to align ExaTools with current Exa conventions and enable API-usage attribution.

• Set the x-exa-integration header on the Exa client ("agno"). This lets Exa attribute API traffic to this integration; without it, usage from Agno is indistinguishable from raw exa_py calls. Wrapped in a best-effort try/except so upstream client changes cannot break init.
• Surface highlights as a content mode alongside text and summary. The Exa API returns any combination of these; the parser now cascades through each.
• Add include_text, exclude_text, and user_location filters to search_exa / get_contents / find_similar. These are standard Exa search parameters that were not previously exposed.
• Fix the type docstring — the current ExaTools docstring describes type as an article/blog/video content filter, but per the Exa docs type is the search mode (auto / neural / fast / deep / deep-lite / deep-reasoning / instant). Updated to match.

Example

from agno.tools.exa import ExaTools

tools = ExaTools(
    highlights=True,
    include_text=["quarterly revenue"],
    user_location="US",
)
# API calls from here now include `x-exa-integration: agno`

Type of change

• Bug fix
• New feature
• Breaking change
• Improvement
• Model update
• Other:

Files changed

• libs/agno/agno/tools/exa.py — header, highlights, new filters, docstring
• libs/agno/tests/unit/tools/test_exa.py — tests for the above

Checklist

• Code complies with style guidelines
• Ran format/validation scripts (./scripts/format.sh and ./scripts/validate.sh)
• Self-review completed
• Documentation updated (comments, docstrings)
• Examples and guides: Relevant cookbook examples have been included or updated (if applicable)
• Tested in clean environment
• Tests added/updated (if applicable)

Duplicate and AI-Generated PR Check

• I have searched existing open pull requests and confirmed that no other PR already addresses this issue
• If a similar PR exists, I have explained below why this PR is a better approach
• Check if this PR was entirely AI-generated (by Copilot, Claude Code, Cursor, etc.)

Test plan

• pytest libs/agno/tests/unit/tools/test_exa.py — 21 passed (5 new)
• ruff format + ruff check on modified files — clean
• mypy libs/agno/agno/tools/exa.py — no new errors introduced