Repository navigation
Experiment 002: Receipt Rave — the sub has a receipt - #2
Merged
Merged
Conversation
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Sorry @EmergentMonk, you've used your own review budget of 250,000 diff characters for the last 7 days.
You can request another review in 2 days and 1 hour by commenting @sourcery-ai review. Upgrade to get a review now.
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
Reviewer's GuideIntroduces Receipt Rave, a dependency-free deterministic five-voice tracker with browser ABX listening trials, SHA-256 replay receipts, Node verification, comprehensive golden-vector tests, CI coverage, and documentation linking it alongside Experiment 001. Sequence diagram for deterministic render and receipt verificationsequenceDiagram
participant User
participant App as Browser or CLI
participant Core as core.mjs
participant Hash as WebCrypto SHA256
User->>App: forge or render config
App->>Core: config(input)
Core-->>App: normalized settings
App->>Core: render(config)
Core-->>App: reference PCM and score
App->>Core: render(config, true)
Core-->>App: shifted PCM and score
App->>Core: makeReceipt(config)
Core->>Hash: sha256(score, PCM, WAV bytes)
Hash-->>Core: hashes
Core-->>App: receipt
User->>App: verify receipt
App->>Core: verifyReceipt(receipt)
Core->>Core: regenerate expected receipt
Core-->>App: verified or mismatches
Sequence diagram for the eight-trial ABX listening gamesequenceDiagram
actor Listener
participant UI as Browser UI
participant Audio as Web Audio
participant Session as ABX session
Listener->>UI: Start blind session
UI->>Session: nextTrial()
loop 8 trials
Listener->>UI: Play A, B, and X
UI->>Audio: play(which, label)
Audio-->>UI: Playback started
UI->>Session: Record audition
Listener->>UI: Guess X is A or B
UI->>Session: trial(bit, choice)
Session-->>UI: Record answer
UI->>Session: nextTrial()
end
UI->>Session: finish()
Session-->>Listener: Reveal score and export listening record
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
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
1 participant
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.
RANDO's second context-guided experiment is Receipt Rave: a playable industrial micro-tracker with five synthetic voices, deterministic PCM16 audio, 432 Hz A-minor tuning, exact lead-delay variants, an eight-trial ABX listening game, and SHA-256 receipts that replay in the browser or Node.
The model inferred the build from recurring music, tracker timing, CPU, and evidence themes. Experiment 001's code and record remain available. The experiment record distinguishes this software intervention from uncollected human preference and listening results.
Validation: local and GitHub Actions
node --testpass 14/14 tests. Coverage includes the original three tests, frozen score/PCM/WAV hashes, identical controls, nonzero delay changes, four seeds at every tempo, receipt mutations, and damaged external WAV rejection. FFprobe recognizes the default export as mono PCM16 at 24 kHz.After merge, the test-gated Pages deployment succeeded. The hosted browser produced the same default PCM hashes as Node. Playback started, the one-character tamper was rejected, ABX settings and answers were gated, and a scripted eight-trial session completed and unlocked export. These scripted answers are UI checks, not human listening evidence. Browser download capture and receipt file-import automation were inconclusive due to browser-tool timeouts; the CLI render/verify flow and shared byte encoders are tested.
The previously configured Pages URL returned 404. The existing Experiments workflow now deploys the static repository after successful main-branch replay checks. Live artifact: https://qsolkcb.github.io/RANDO/experiment-002/
No package install, build step, external samples, or model calls are required by the artifact. Method, replay commands, limits, and comedic release notes are in
experiment-002/.