[wenh06]

Summary

• tutorial-first-experiment: compare FedAvg / FedProx / SCAFFOLD at two sampling ratios on synthetic data (six runs into a pandas table), then dissect what num_communications counts — messages, not rounds, with the discriminating check (3 rounds full sampling = 6 rounds half sampling = 180) and the per-direction decomposition. A closing section runs FedPD next to FedAvg and shows the saved budget comes from skipped uplinks only; it explains the expected RuntimeWarnings on both server and client side and the 0 < p < 1 constraint. All shown code is embedded via literalinclude from the runnable test-files/tutorial_first_experiment.py.
• tutorial-cli-matrix: run an experiment matrix from one YAML file; the new example-configs/fedsynthetic-demo.yml expands into four synthetic-data runs. The page documents the interpolation syntax and the per-arm tag that keeps same-minute runs from overwriting each other's logs (the demo initially hit exactly that — only the last run's json survived until the tags were added).
• index.rst: new Tutorials toctree section; quickstart next-steps now point at the tutorials.
• test/test_tutorial_scripts.py: behavioral smoke tests extended — the comparison-table communications invariants (90 / 180), the FedPD coda (fewer communications than FedAvg with at least one real uplink, guarding the degenerate never-uploads case), and the CLI matrix expansion (distinct combinations, interpolated values, one json+txt pair per arm, budget asserted from each run's config.json).
• zh_CN translations for the new and updated pages via the gettext pipeline.

Validation

• pytest test/test_tutorial_scripts.py test/test_hook_contract.py test/test_messages.py: 19 passed (~19 s for the five behavioral smoke tests, real runs included).
• English docs build: zero warnings. zh_CN docs build: zero new warnings vs master.
• The T1 script's FedAvg ratio-0.5 run reproduces the CLI matrix's ratio-0.5 run bit-for-bit in final accuracy (0.8114), confirming the tutorial narrative about the fixed aggregation order.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.70%. Comparing base (11c19de) to head (787ed71).

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #58      +/-   ##
==========================================
+ Coverage   82.32%   83.70%   +1.38%     
==========================================
  Files          55       55              
  Lines        5640     5640              
==========================================
+ Hits         4643     4721      +78     
+ Misses        997      919      -78     

see 6 files with indirect coverage changes

🚀 New features to boost your workflow:

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