Add live telemetry dashboard and batch-simulation automation - #16
Open
DanielFourman wants to merge 8 commits into
Open
Add live telemetry dashboard and batch-simulation automation#16DanielFourman wants to merge 8 commits into
DanielFourman wants to merge 8 commits into
Conversation
Adds a telemetry package (observer + sinks) wired into the live game
loop, new /telemetry, /simulate, /simulate/{job_id} server routes, and
a simulate.py CLI for running batch games between two drivers without
the websocket/UI rate throttle.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Coins are worthless individually; collecting exactly config.coins_for_jackpot (5) awards a one-time config.score_jackpot (50) bonus. Coins collected beyond that grant no further bonus. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Splits the blended hits/pickups counters into per-obstacle success/fail/miss data the redesigned telemetry dashboard needs: wall_hits, water_hits, crack_hits, coin_misses, and a finally-wired-up misses (penguin miss was tracked in the field but never incremented). Also exposes a `running` flag and `total_finished` round count on GET/POST /telemetry(/clear), and attaches each player's full end-of-game state to game results so "Recent games" can show a real per-driver stat breakdown instead of just final score. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds duration_seconds to the game-end result (both the live loop and the headless simulate.py path), timed from round start to round end. Meaningful mainly for the headless/automation path, where rounds run back-to-back as fast as the drivers respond rather than on a fixed wall-clock cadence. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Coins are gone from the game, matching upstream RedHat-Israel/rose-game-engine: no COIN obstacle, no coins_for_jackpot/score_jackpot config, no coin/coin_misses player counters, no coin scoring branch. Kept everything else added since (wall_hits/water_hits/crack_hits/misses fix, running flag, duration_seconds, players-in-result) since none of that is coin-specific. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Clear history previously left the round count monotonically increasing forever (by design, since result_count() backs the live-batch "has another round finished" poll and must never go backwards or reset mid-job). Split that: result_count() stays untouched across a clear, but a new _round_offset captures total_finished at clear time so the number shown to users (via snapshot()'s total_finished) restarts at 1 for the next game, without disturbing the internal counter. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
_run_live_batch() drives batch games through the shared game_loop (the same reset/running state a human's Reset+Run would set), not through simulate.py's headless runner. The test never started that loop, so telemetry.result_count() never advanced and the job polled forever until timing out. Start logic.game_loop() as a background task for the duration of the test (mirroring what server.run() does in production) and cancel it afterward, and give the poll loop enough budget to clear the mandatory inter-round pause.
2 tasks
Member
|
Thank you for the pull request 👍 Can you fix the tests ? |
state is a dict that's mutated in place (state["key"] = ...), never reassigned, so global was never required.
Author
|
Thanks for the review! Tests/lint: Fixed — UI change: Yes, this pairs with a UI change — the telemetry dashboard and automation pages are added in rose-game-web-ui#17: RedHat-Israel/rose-game-web-ui#17 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Player:wall_hits,water_hits,crack_hitsadded alongside the existing success counters, andmissesis now actually incremented on a missed penguin (previously declared but dead) — this is what lets the telemetry UI show truesuccess/totalratios instead of always 100%.COIN,coins_for_jackpot,score_jackpot, and its scoring branch).LiveSink(telemetry): finished-round results now include full per-player state andduration_seconds; a round-numbering fix lets "Clear history" reset the displayed round count without disturbing the internal counter batch-polling relies on.rose/engine/simulate.py(run_single_game/run_batch/_aggregate) plus a root-levelsimulate.pyCLI (python simulate.py --drivers A B --games 50 --output stats.json), documented in the README./simulate(POST) and/simulate/{job_id}(GET) endpoints: start a batch job and poll it._run_live_batch()drives the games through the same shared game loop a human's Reset+Run would use, so a batch run plays out live on the main game screen.rose/telemetry/observer.py: theTelemetryObserver/NullObserver/CompositeObserverinterface both the live dashboard and automation are built on.Dockerfile'sCMDreferencedrose/main.py, which hasn't existed for 2 years — corrected tomain.py.Test plan
pytest— 75/75 passing (fixed one pre-existing hang:test_simulate_job_runs_in_background_and_completesnever started the background game loop the/simulateendpoint depends on, so it polled forever; now starts/cancelslogic.game_loop()for the test's duration, matching whatserver.run()does in production)flake8/black --check— no new issues (pre-existing warnings confirmed present in the pre-change baseline too)radon cc— no complexity regression/simulatewith two real driver processes, polled to completion, verified aggregated results; confirmed/telemetryand the dashboard both serve real per-obstacle ratios during live play🤖 Generated with Claude Code