Description

This PR fixes the OpenAI-compatible VLM backend so the thinking flag is propagated to DashScope-compatible endpoints via extra_body.enable_thinking.

The change is scoped to DashScope detection only, so official OpenAI and Azure requests keep their existing behavior.

Related Issue

Fixes #923

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Refactoring (no functional changes)
• Performance improvement
• Test update

Changes Made

• Detect DashScope-compatible OpenAI backends from the configured model prefix and api_base host.
• Pass extra_body={"enable_thinking": <bool>} through all four OpenAI VLM completion paths.
• Add regression tests covering DashScope text/async vision calls and non-DashScope guard rails for official OpenAI and Azure.

Testing

• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• I have tested this on the following platforms:
  • Linux
  • macOS
  • Windows

Validated with:

• python3 -m py_compile openviking/models/vlm/backends/openai_vlm.py tests/unit/test_extra_headers_vlm.py
• Isolated behavior verification for DashScope sync text, DashScope async vision, official OpenAI, and Azure.

Checklist

• My code follows the project's coding style
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• Any dependent changes have been merged and published

Screenshots (if applicable)

N/A