Python: Emit partial result for magentic pattern when retrieving final result, if available

[moonbox3]

Motivation and Context

When using the magentic orchestration pattern, if the max_round_count is hit, the final result shows as:

Final result:
Max round count reached.

The interim messages do show as part of the agent_response_callback; however, not everyone may have that configured. We should return more meaningful results, even if partial. This PR updates to try and get a partial result if it exists. When one calls:

value = await orchestration_result.get()

for the step5_magentic.py sample, with max_round_count=1 they should receive the partial result:

Final result:
Based on the available data, here is a comparison of the estimated training and inference energy consumption for ResNet-50, BERT-base, and GPT-2, along with the associated CO₂ emissions when training on an Azure Standard_NC6s_v3 VM for 24 hours.

**Model Architectures and Datasets:**

- **ResNet-50**: Image classification model trained on ImageNet.
- **BERT-base**: Text classification model fine-tuned on the GLUE benchmark.
- **GPT-2**: Text generation model trained on WebText.

... <rest omitted for brevity> ...

Description

Return partial results for magentic orchestration if they exist.

• Closes Python: Bug: max_round_count ends loop without returning results—unlike AutoGen behavior #12625

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the SK Contribution Guidelines and the pre-submission formatting script raises no violations
• All unit tests pass, and I have added new tests where possible
• I didn't break anyone 😄

[markwallace-microsoft]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
agents/orchestration
magentic.py	292	102	65%	122–124, 256, 379, 447, 450, 457, 461, 463, 490–495, 497, 502, 504, 510, 512, 517–518, 520–521, 527, 529–530, 533–534, 540, 548–549, 558, 561–562, 564–567, 570–571, 575–578, 580–581, 583, 585–591, 594–595, 602, 611–613, 615, 617, 625–626, 628, 633, 637–638, 640, 642–643, 647–648, 650, 653, 657–659, 662, 666–667, 673–674, 676, 678, 692–693, 697–698, 700, 702, 704, 706, 715–719, 786–787
TOTAL	26428	4514	82%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
3639	22 💤	0 ❌	0 🔥	1m 51s ⏱️