fix(agents): stop internal Agent Config subclassing from warning on plain imports

[chelsealong]

Link to Issue or Description of Change

• Closes: [BUG]: importing LlmAgent emits BaseAgentConfig deprecation warning from ADK internals #6968

Problem:

Importing the public LlmAgent, SequentialAgent, LoopAgent, or
ParallelAgent classes emits a BaseAgentConfig DeprecationWarning, even
when the application never imports or uses the deprecated YAML Agent Config
feature. Test suites that treat deprecations as errors fail during an
ordinary ADK import:

$ python -W error::DeprecationWarning -c 'from google.adk.agents import LlmAgent, SequentialAgent'
...
DeprecationWarning: BaseAgentConfig is deprecated and will be removed in future versions. ...

The root cause: ADK's own internal config classes (LlmAgentConfig,
SequentialAgentConfig, LoopAgentConfig, ParallelAgentConfig) subclass
the deprecated BaseAgentConfig. typing_extensions.deprecated warns on
any subclassing of a deprecated class, so simply defining these built-in
subclasses at import time — before any application code touches Agent
Config — triggers the warning.

Solution:

Scope a warnings.catch_warnings() suppression around just the internal
import of each *_agent_config module in llm_agent.py, sequential_agent.py,
loop_agent.py, and parallel_agent.py. This only silences the warning at
the exact point ADK defines its own internal subclasses; it does not touch
BaseAgentConfig's deprecation machinery itself, so applications that
actually use the deprecated Agent Config APIs (importing a config class
directly, instantiating BaseAgentConfig, or subclassing it themselves)
still get warned as before.

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
• All unit tests pass locally.

Added tests/unittests/agents/test_import_deprecation_warnings.py, which
runs imports in a fresh subprocess with -W error::DeprecationWarning (so
module caching across tests can't hide the warning), per the reproduction
approach the issue itself suggested.

Confirmed the new test fails without the fix (checked out the pre-fix source
files against the new test):

FAILED tests/unittests/agents/test_import_deprecation_warnings.py::test_importing_runtime_agents_does_not_warn_about_agent_config
...
DeprecationWarning: BaseAgentConfig is deprecated and will be removed in future versions. ...

And passes with the fix:

tests/unittests/agents/test_import_deprecation_warnings.py::test_importing_runtime_agents_does_not_warn_about_agent_config PASSED
tests/unittests/agents/test_import_deprecation_warnings.py::test_directly_importing_llm_agent_config_still_warns PASSED
2 passed in ...

Manually verified genuine Agent Config usage still warns:

$ python -W error::DeprecationWarning -c "from google.adk.agents import LlmAgentConfig"
DeprecationWarning: BaseAgentConfig is deprecated and will be removed in future versions. ...

$ python -W error::DeprecationWarning -c "
from google.adk.agents import LlmAgent
from google.adk.agents.base_agent_config import BaseAgentConfig
BaseAgentConfig(name='x')
"
DeprecationWarning: BaseAgentConfig is deprecated and will be removed in future versions. ...

Ran the full unit test suite:

$ python -m pytest tests/unittests -q
13534 passed, 81 skipped, 26 xfailed, 2 xpassed, 2942 warnings, 24 subtests passed in 449.62s

Formatted and linted with pyink/isort per CONTRIBUTING.md; pylint
score on the touched files is unchanged from main (9.20 vs 9.21, same
pre-existing warnings, none introduced by this change).

Manual End-to-End (E2E) Tests:

Not applicable — this is an import-time warning fix with no behavioral
change to runtime agent execution.

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• 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 manually tested my changes end-to-end. (N/A, import-time-only fix; see testing plan)
• Any dependent changes have been merged and published in downstream modules. (N/A)

AI assistance disclosure

This change was written with the assistance of an AI coding agent (Claude),
with the diff reviewed and tests verified by the submitter before opening
this PR.