Python: Fix response_format crash on background polling with empty text

[eavanvalkenburg]

Motivation and Context

When using response_format with a Pydantic model or dict schema together with background=True, the SDK crashes during polling because _parse_structured_response_value is called with an empty string while the response is still in-progress, causing model_validate_json("") to raise ValidationError and json.loads("") to raise JSONDecodeError.

Fixes #5145

Description

The root cause is that _parse_structured_response_value unconditionally attempts to parse the text even when it is empty, which happens on every poll iteration before the background response completes. The fix adds an early return of None when text is falsy, before reaching either the Pydantic or dict-schema parsing branches. This allows .value to safely return None on in-progress polls and the fully parsed result once the response is complete. Regression tests cover both the Pydantic and mapping schema paths as well as the ChatResponse and AgentResponse wrappers.

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the Contribution Guidelines
• All unit tests pass, and I have added new tests where possible
• Is this a breaking change? If yes, add "[BREAKING]" prefix to the title of the PR.

---

Note: PR autogenerated by eavanvalkenburg's agent

[eavanvalkenburg]

Automated Code Review

Reviewers: 4 | Confidence: 98% | Result: All clear

Reviewed: Correctness, Security Reliability, Test Coverage, Design Approach

---

Automated review by eavanvalkenburg's agents

[Copilot]

Pull request overview

Fixes a crash in the Python core SDK when polling background Responses API runs with response_format enabled, where structured parsing was attempted against an empty in-progress .text.

Changes:

• Add an early-return guard in _parse_structured_response_value when text is empty.
• Add regression tests covering empty-text behavior for both Pydantic and mapping schema formats, including ChatResponse and AgentResponse wrappers.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
python/packages/core/agent_framework/_types.py	Prevents structured parsing from running on empty in-progress text during polling by returning None early.
python/packages/core/tests/core/test_types.py	Adds regression coverage ensuring empty text with response_format returns None (and does not raise) across parsing paths and response wrappers.

[moonbox3]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
packages/core/agent_framework
_types.py	1082	90	91%	58, 67–68, 122, 127, 146, 148, 152, 156, 158, 160, 162, 180, 184, 210, 232, 237, 242, 246, 276, 687–688, 847–848, 1233, 1305, 1340, 1360, 1370, 1422, 1554–1556, 1738, 1841–1846, 1871, 1959, 1967–1969, 1974, 2065, 2077, 2100, 2355, 2379, 2474, 2723, 2929, 2998, 3009, 3011–3015, 3017, 3020–3028, 3038, 3242–3244, 3247–3249, 3253, 3258, 3262, 3346–3348, 3377, 3454–3458
TOTAL	27059	3188	88%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
5377	20 💤	0 ❌	0 🔥	1m 24s ⏱️