Add Deepseek Reasoning example

[ivanleomk]

This PR adds an example snippet of how to use deepseek's reasoner model with instructor

---

Important

Adds a deepseek reasoner model example to documentation, updates GitHub action version, and modifies a test URL.

• Documentation:
  • Adds example for using deepseek-reasoner model with instructor in deepseek.md.
  • Demonstrates structured output and reasoning trace extraction.
• CI/CD:
  • Updates upload-artifact action to v4 in ruff.yml.
• Tests:
  • Updates audio_url in test_multimodal.py to a new URL.

^{This description was created by for 8d91399. It will automatically update as commits are pushed.}

[ellipsis-dev]

👍 Looks good to me! Reviewed everything up to 5c5dadf in 25 seconds

More details

• Looked at 54 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 6 drafted comments based on config settings.

1. docs/integrations/deepseek.md:269

• Draft comment:
  The new 'Reasoning Models' section is clear and well-documented. Ensure that using 'instructor.Mode.MD_JSON' is highlighted as a requirement for proper reasoning output.
• Reason this comment was not posted:
  Confidence changes required: 0%
  None

2. docs/integrations/deepseek.md:292

• Draft comment:
  The use of 'create_with_completion' is clear; verify that its returned tuple (completion, raw_completion) is documented in the API docs for consistency.
• Reason this comment was not posted:
  Confidence changes required: 33%
  None

3. docs/integrations/deepseek.md:278

• Draft comment:
  The 'from rich import print' is used in this example. Confirm that highlighting colored or formatted output is intentional, else consider using the standard print for consistency.
• Reason this comment was not posted:
  Confidence changes required: 0%
  None

4. docs/integrations/deepseek.md:271

• Draft comment:
  Good example demonstrating the MD_JSON mode with the deepseek-reasoner model. As a suggestion, consider adding a brief note or error handling if the raw_completion structure doesn’t include choices or the 'reasoning_content' field, to improve robustness in edge cases.
• Reason this comment was not posted:
  Confidence changes required: 33%
  None

5. docs/integrations/deepseek.md:292

• Draft comment:
  The use of create_with_completion() is clear and useful for obtaining both structured output and raw reasoning. It might be useful to add a short comment explaining that this method returns a tuple (parsed output, raw completion) and that it requires updated support in the client.
• Reason this comment was not posted:
  Confidence changes required: 33%
  None

6. docs/integrations/deepseek.md:302

• Draft comment:
  Directly accessing raw_completion.choices[0].message.reasoning_content assumes that the API response always includes at least one choice with a message containing 'reasoning_content'. Consider a safer access pattern or a try/except block to handle cases where the response might not match this structure.
• Reason this comment was not posted:
  Confidence changes required: 33%
  None

Workflow ID: wflow_4y9sWRrdHo7MxPeH

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[cloudflare-workers-and-pages]

Deploying instructor-py with    Cloudflare Pages

Latest commit:	8d91399
Status:	✅  Deploy successful!
Preview URL:	https://d6c47fed.instructor-py.pages.dev
Branch Preview URL:	https://add-deepseek-reasoning.instructor-py.pages.dev

View logs

[ellipsis-dev]

👍 Looks good to me! Incremental review on 09bd013 in 15 seconds

More details

• Looked at 13 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 2 drafted comments based on config settings.

1. .github/workflows/ruff.yml:34

• Draft comment:
  Verify that using upload-artifact@v4 is fully compatible with our workflow. Update docs if needed.
• Reason this comment was not posted:
  Comment did not seem useful.

2. .github/workflows/ruff.yml:34

• Draft comment:
  Upgrading to actions/upload-artifact@v4 is good, but please confirm that its inputs (name and path) remain compatible with v4.
• Reason this comment was not posted:
  Confidence changes required: 33%
  None

Workflow ID: wflow_CNATawLbZ6KLL3wa

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

👍 Looks good to me! Incremental review on 8d91399 in 26 seconds

More details

• Looked at 13 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 2 drafted comments based on config settings.

1. tests/llm/test_openai/test_multimodal.py:11

• Draft comment:
  Updated audio URL domain. Confirm the new URL is valid and reliable.
• Reason this comment was not posted:
  Confidence changes required: 0%
  None

2. tests/llm/test_openai/test_multimodal.py:11

• Draft comment:
  Updated audio_url from 'www2.cs.uic.edu' to 'www.cs.uic.edu'. Confirm that this URL is reachable and returns the expected audio file.
• Reason this comment was not posted:
  Confidence changes required: 0%
  None

Workflow ID: wflow_JWQiVToeoc6qnS99

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.