Taris Run Full Tests | Skills Pool
Taris Run Full Tests Run the full taris test suite across all categories: offline unit tests (telegram, screen_loader, LLM), voice regression (local source-inspection and Pi), Web UI Playwright, and smoke tests. Reports pass/fail per category.
Test Categories
Cat Suite Location Command Pi required? F Telegram offline src/tests/telegram/pytest ... -qNo G Screen DSL loader src/tests/screen_loader/pytest ... -qNo H LLM providers src/tests/llm/pytest ... -qNo A Voice regression (source inspection T01-T41) src/tests/test_voice_regression.pypython3 ...No (for source tests) A Voice regression (hardware T01-T16) Pi target plink ... python3 ...Yes (Pi) B Web UI Playwright
npx skillvault add stas-ka/stas-ka-sintaris-pl-github-skills-taris-run-full-tests-skill-md
星標 0
更新時間 2026年3月28日
職業 src/tests/ui/
pytest --base-url ... --browser chromium
E Smoke manual or Playwright smoke varies Yes
Quick Local Run (No Pi, ~5s) Runs categories F + G + H and source-inspection voice tests (T17–T41):
cd /home/stas/projects/sintaris-pl
DEVICE_VARIANT=openclaw PYTHONPATH=src python3 -m pytest \
src/tests/telegram/ src/tests/screen_loader/ src/tests/llm/ \
-q --tb=short
Then source-inspection voice regression:
DEVICE_VARIANT=openclaw PYTHONPATH=src python3 src/tests/test_voice_regression.py \
--test t_confidence_strip t_tts_escape t_i18n_string_coverage t_lang_routing \
t_bot_name_injection t_profile_resilience t_note_edit_append_replace \
t_calendar_tts_call_signature t_calendar_console_classifier \
t_db_voice_opts_roundtrip t_db_migration_idempotent t_rag_lr_products \
t_web_link_code_roundtrip t_system_chat_clean_output t_openclaw_stt_routing \
t_openclaw_ollama_provider t_web_stt_provider_routing t_pipeline_logger \
t_dual_stt_providers t_voice_debug_mode t_stt_language_routing_fw \
t_stt_fallback_chain t_tts_multilang t_voice_llm_routing \
t_voice_system_mode_routing_guard t_voice_lang_stt_lang_priority
Full Voice Regression (Pi target required)
On PI2 (engineering target) — run before committing # TariStation2 (OpenClaw local):
DEVICE_VARIANT=openclaw PYTHONPATH=src python3 src/tests/test_voice_regression.py
# Or on Taris PI2 (OpenClawPI2):
plink -pw "$HOSTPWD" -batch stas@OpenClawPI2 \
"python3 /home/stas/.taris/tests/test_voice_regression.py"
On PI1 (production) — run only after PI2 passes and code is on master plink -pw "$HOSTPWD" -batch stas@OpenClawPI \
"python3 /home/stas/.taris/tests/test_voice_regression.py"
Web UI Playwright (requires taris-web running) # Target: TariStation2 (local web server at :8080)
DEVICE_VARIANT=openclaw PYTHONPATH=src \
python3 -m pytest src/tests/ui/test_ui.py -v \
--base-url http://localhost:8080 --browser chromium
# Target: OpenClawPI2
python3 -m pytest src/tests/ui/test_ui.py -v \
--base-url https://openclawpi2:8080 --browser chromium
Pass / Fail Rules Result Action All PASS Proceed to commit / deploy Any FAIL Fix before committing — never deploy a failing test WARN (>30% slower than baseline) Investigate; if intentional run --set-baseline SKIP OK for optional features (Whisper, VAD, Ollama not running)
Decision Table — Which Categories to Run Changed files Run categories src/features/bot_voice.py, src/core/bot_config.pyA + F + H src/ui/*.py, src/screens/*.yamlG + F src/core/bot_llm.pyH src/telegram_menu_bot.py, src/telegram/F src/bot_web.py, src/web/B src/strings.jsonT13 (i18n) + F Any Quick local run (F + G + H) first, then voice regression
Reporting After a full run, paste summary table in commit message:
Tests: 40 F-pass | 64 G-pass | 18 H-pass | T01-T41 A-pass (src-inspect)
Baseline for voice hardware tests lives on Pi: ~/.taris/tests/voice/results/baseline.json
Reset after re-image: python3 test_voice_regression.py --set-baseline
02
Quick Local Run (No Pi, 5s)