chore(pipelines): standardize pytest and testing patterns across Python solution guides

[iht]

Description

Harmonizes the Python testing methodology, developer dependencies, CI test execution, and documentation across all solution guides in the repository.

Key Changes

1. Requirements Standardization:

   • Standardized requirements-dev.txt across all 5 Python pipelines using non-breaking version ranges:
     • yapf>=0.43,<1.0
     • pylint>=4.0,<5.0
     • pytest>=9.0,<10.0
     • setuptools>=78
   • Updated pipelines/anomaly_detection, pipelines/iot_analytics, and pipelines/ml_ai_python.
   • Added standard requirements-dev.txt to pipelines/cdp and pipelines/marketing_intelligence.

2. Baseline Unit Tests for ml_ai_python:

   • Added pipelines/ml_ai_python/tests/test_pipeline.py with 7 unit tests covering turn tag prompt formatting, output parsing (PredictionResult), choices formatting, and MyPipelineOptions parsing with default model preset google/gemma-4-E2B-it.

3. CI Test Execution (pull_request.yml):

   • Updated python-build job to invoke PYTHONPATH=. pipenv run pytest tests/ -v.
   • Preserved lightweight standard library unittest in minimal container compatibility checks (anomaly-training-check and anomaly-serving-check) to keep deployment containers lean.

4. Renovate Noise Reduction (renovate.json):

   • Added python-dev-tools package group for pytest, pylint, and yapf so that future dependency updates are bundled across all pipelines.

5. Documentation & Skills:

   • Updated AGENTS.md and .agents/skills/dataflow-pipeline-dev/SKILL.md to document running unit tests with pytest tests/ -v.
   • Standardized testing commands across all Python pipeline README.md files.

Verification

• Tested unit test execution across pipelines (ml_ai_python: 7/7 passed, anomaly_detection: 28/28 passed, marketing_intelligence: 8/8 passed).
• Ran yapf --diff -r (0 diffs).
• Ran pylint (10.00/10 score).
• Verified python setup.py sdist succeeded in ml_ai_python.
• Validated renovate.json with jq.

[iht]

Should not we use also pyttest for the training tests?

[iht]

Yes, absolutely! Updated pipelines/anomaly_detection/README.md and use_cases/Anomaly_Detection.md to pytest training/tests/ -v for full consistency.

[iht]

Should not we use also pyttest for the serving tests?

[iht]

Yes! Added fastapi>=0.115,<1 and httpx>=0.28,<1 to pipelines/anomaly_detection/requirements-dev.txt so pytest serving/tests/ -v can execute locally without requiring the Docker container, and updated pipelines/anomaly_detection/README.md and use_cases/Anomaly_Detection.md to pytest serving/tests/ -v.