fix: support Anthropic tool_use and tool_result in token counter

[orolega]

Title

Fix token counting for Anthropic tool interactions

Relevant issues

Fixes #14236

Pre-Submission checklist

Please complete all items before asking a LiteLLM maintainer to review your PR

• I have Added testing in the tests/litellm/ directory, Adding at least 1 test is a hard requirement - see details
• I have added a screenshot of my new test passing locally
• My PR passes all unit tests on make test-unit
• My PR's scope is as isolated as possible, it only solves 1 specific problem

Type

🐛 Bug Fix
✅ Test

Changes

Adds support for counting tokens in Anthropic's tool_use and tool_result
content blocks. Previously, these content types caused "Invalid content
type" errors in the router's pre-call checks, preventing proper context
window filtering for Claude Code requests with tool interactions.

• Add tool_use content type handler in _count_content_list() to count tokens from tool name and input parameters
• Add tool_result content type handler supporting both string and nested list content formats
• Add comprehensive test coverage for tool_use, tool_result, and combined tool conversation scenarios
• Fix router pre-call checks to properly filter deployments by context window when tool interactions are present

[vercel]

@orolega is attempting to deploy a commit to the CLERKIEAI Team on Vercel.

A member of the Team first needs to authorize it.

[krrishdholakia]

can we have this in a separate helper function, and ideally use a typed dict, so we know what to expect here? we might already have one for this in types/llms/anthropic.py