feat: add OpenAI Responses routing for chat

[afurm]

What this does

Adds OpenAI Responses API support behind the existing RubyLLM.chat interface.

This keeps Chat Completions as the default path in :auto, then routes native OpenAI requests to Responses when the request requires it, including native Responses tools like web_search, deep-research models, Responses-only params, and GPT-5 tool calls with reasoning enabled.

The change also adds:

• config.openai_api_mode = :auto | :chat_completions | :responses
• per-request override via with_params(openai_api_mode: :responses)
• Responses payload rendering for messages, files, schemas, reasoning, Ruby tools, and native OpenAI tools
• Responses parsing for output text, tool calls, reasoning summaries, and usage
• semantic Responses streaming event handling
• docs for routing and native tool usage

Closes #213.

Type of change

• Bug fix
• New feature
• Breaking change
• Documentation
• Performance improvement

Scope check

• I read the Contributing Guide
• This aligns with RubyLLM's focus on LLM communication
• This isn't application-specific logic that belongs in user code
• This benefits most users, not just my specific use case

Required for new features

• I opened an issue before writing code and received maintainer approval
• Linked issue: OpenAI Responses API support #213

Quality check

• I ran overcommit --install and all hooks pass
• I tested my changes thoroughly
  • For provider changes: Re-recorded VCR cassettes with bundle exec rake vcr:record[provider_name]
  • All tests pass: bundle exec rspec
• I updated documentation if needed
• I didn't modify auto-generated files manually (models.json, aliases.json)

Validation run:

• bundle exec rspec spec/ruby_llm/providers
• focused OpenAI Responses/config/streaming specs
• bundle exec rubocop --ignore-unrecognized-cops ...
• git diff --check

VCR cassettes were not re-recorded in this branch.

AI-generated code

• I used AI tools to help write this code
• I have reviewed and understand all generated code (required if above is checked)

API changes

• Breaking change
• New public methods/classes
• Changed method signatures
• No API changes

Adds a new OpenAI configuration option, openai_api_mode, and a provider-specific per-request routing param with the same name.

[codecov]

Codecov Report

❌ Patch coverage is 90.29536% with 23 lines in your changes missing coverage. Please review.
✅ Project coverage is 87.25%. Comparing base (4942d6c) to head (a84517d).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
lib/ruby_llm/providers/openai/responses.rb	86.51%	12 Missing ⚠️
lib/ruby_llm/providers/openai.rb	92.85%	5 Missing ⚠️
lib/ruby_llm/providers/openai/streaming.rb	92.85%	3 Missing ⚠️
lib/ruby_llm/providers/openai/tools.rb	88.88%	2 Missing ⚠️
lib/ruby_llm/stream_accumulator.rb	83.33%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #770      +/-   ##
==========================================
+ Coverage   87.05%   87.25%   +0.19%     
==========================================
  Files         119      120       +1     
  Lines        5594     5820     +226     
  Branches     1407     1475      +68     
==========================================
+ Hits         4870     5078     +208     
- Misses        724      742      +18     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.