feat: add tool for new conversions schema and spec

[matt-landers]

Description

This PR implements the new Conversions Reporting functionality for the Google Analytics MCP server, allowing users to query conversion-specific metrics, ad performance, and attribution data using the Google Analytics Data API.

Key Changes

• Library Upgrade: Updated google-analytics-data to version 0.22.0 in pyproject.toml to support the conversion_spec field in RunReportRequest.
• New Tool Implementation: Created analytics_mcp/tools/reporting/conversions.py containing the run_conversions_report function.
  • Utilizes the v1alpha client's run_report method.
  • Explicitly maps the conversion_spec parameter (including conversion_actions and attribution_model).
• Enhanced LLM Guidance:
  • Added a "WHEN TO USE" section to the tool's docstring to help LLM agents distinguish between standard reports and conversion reports.
  • Explicitly listed all 12 allowed camelCase dimensions and 12 allowed camelCase metrics directly in the tool description to minimize errors during tool calling.
  • Provided clear examples for conversion_spec structure and attribution models.
• Server Registration: Registered the new tool in analytics_mcp/coordinator.py and marked mandatory fields (property_id, date_ranges, dimensions, metrics, conversion_spec) in the MCP
  schema.