[codex] Clarify TensorFlow compatibility docs

[shenweichen]

Summary

• Document the current TensorFlow installation model: install a platform-appropriate TensorFlow build first, then install DeepCTR.
• Clarify Python, h5py, and NumPy compatibility guidance for modern Python environments.
• Replace example and FAQ imports from private tensorflow.python.keras APIs with public tensorflow.keras APIs.
• Update TFRecord examples to use tf.io.FixedLenFeature.

Why

Several open issues report installation failures or model serialization problems caused by outdated dependency guidance and mixed Keras imports. The docs now steer users toward public TensorFlow APIs and make the TensorFlow/NumPy/h5py relationship explicit.

Validation

• git diff --check
• python -m compileall -q examples
• python -m sphinx -b html docs/source /tmp/deepctr-docs was attempted locally, but the local environment is missing sphinx_rtd_theme; this branch includes that package in docs/requirements.readthedocs.txt for ReadTheDocs builds.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.11%. Comparing base (694d6d9) to head (1cfd859).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #550   +/-   ##
=======================================
  Coverage   78.11%   78.11%           
=======================================
  Files          61       61           
  Lines        3304     3304           
=======================================
  Hits         2581     2581           
  Misses        723      723           

Flag	Coverage Δ
pytest	78.11% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.