Maailmantutkija

Purpose: Systematic testing of all revised components and features Status: Ready for testing Test Levels:

• levels/comprehensive-test-level.json - All components integrated
• levels/edge-cases-test-level.json - Edge cases and parameter limits
• levels/inference-chain-test-level.json - Complex Bayesian inference chains

Phase 1: Component-Specific Tests

✓ Observation Component (5×5 Grid)

Test Level: edge-cases-test-level.json (Component ID 5)

• Basic Grid Layout
  • Place 1-25 balls, verify they arrange in 5×5 grid
  • Check spacing is correct (0.16 units between balls)
  • Verify ball scale (0.15) is applied correctly
• Overflow Handling (>25 balls)
  • Test with 30+ balls (edge-cases-test-level spawns 50 balls through duplicator)
  • Verify overflow behavior (stacking behind or warning)
  • Check console for any errors
• Visual Rendering
  • Balls should be small (ballScale=0.15)
  • No overlapping
  • Grid should be centered in component

Expected Issues: None Actual Issues: [Document any issues found]

✓ Duplicator Component (Multi-Output)

Test Levels:

• comprehensive-test-level.json (ID 3: 3-way split)
• edge-cases-test-level.json (ID 3: max 10 copies, ID 19: 4-way split)

Test 1: Multi-Output Distribution

• 3-Output Pattern (comprehensive-test, ID 3)
  • Pattern: right:3, down:2, up:1 (6 total)
  • Verify exactly 3 balls go right, 2 down, 1 up
  • Check distribution is deterministic (not random)
• 4-Output Pattern (edge-cases-test, ID 19)
  • Pattern: right:2, down:2, left:1, up:1 (6 total)
  • Verify all 4 outputs work simultaneously
  • Check each observation gets correct count
• Max Copies (edge-cases-test, ID 3)
  • 10 copies to single output
  • Verify all 10 copies are created
  • Check performance with many copies

Test 2: Visual Rendering

• Input arrow drawn from correct side (opposite to first output)
• Gold channels drawn for each unique output side
• Each output labeled with count (e.g., “3” next to right arrow)
• Center shows total: “×6” or “×10”
• Rendering works correctly in all 4 rotations

Test 3: Bayesian Tracking

• Basic Tracking (inference-chain-test, ID 3)
  • Duplicator has plex glass
  • Verify ballToSack mapping correct for all copies
  • Check console for any tracking warnings
• Multi-Path Tracking
  • Copies going to different paths
  • Verify dependencies tracked correctly
  • All copies should have same sackId

Test 4: Rotation

• Rotate duplicator with R key
• Verify all output sides rotate correctly
• Check connections update automatically
• Pattern remains deterministic after rotation

Expected Issues: None (Task 4 deferred - might see defensive warnings) Actual Issues: [Document any issues found]

✓ Shuffler Component (Multi-Output + Retention)

Test Levels:

• comprehensive-test-level.json (ID 7: dual output + retention)
• edge-cases-test-level.json (ID 8: max retention, ID 26: minimal config)
• inference-chain-test-level.json (ID 5: in complex chain)

Test 1: Single Output (No Retention)

• Minimal Config (edge-cases-test, ID 26)
  • 1 input, minBufferSize=1, output all to right
  • Verify balls pass through immediately
  • Check single-ball buffering works

Test 2: Dual Output (No Retention)

• Create test with pattern: [{side: 'right', count: 1}, {side: 'down', count: 1}]
  • minBufferSize=2
  • Verify deterministic split (not random)
  • First ball after shuffle → right, second → down

Test 3: Single Output with Retention

• Max Retention (edge-cases-test, ID 8)
  • minBufferSize=10, pattern: output 5 right, retain 5
  • Cycle 1: Wait for 10 balls → shuffle → output 5 → retain 5
  • Cycle 2: 5 retained + 5 new = 10 → shuffle → output 5 → retain 5
  • Verify retained balls persist across cycles
  • Check console logs show cycle numbers

Test 4: Dual Output with Retention

• Complex Pattern (comprehensive-test, ID 7)
  • minBufferSize=5, pattern: right:2, down:1, retain:2
  • Cycle 1: 5 balls → shuffle → 2 right, 1 down, 2 retained
  • Cycle 2: 2 retained + 3 new = 5 → repeat
  • Verify retention + multi-output works together

Test 5: Visual Rendering

• Gold channels for each output side
• Dashed gray rectangle when retention enabled
• Buffer count display: “2/5” (current/minBufferSize)
• Rendering correct in all rotations

Test 6: Bayesian Inference with Retention

• Cross-Cycle Dependencies (inference-chain-test, ID 5)
  • Shuffler has retention, receives balls from 2 sacks
  • Balls from different cycles should be grouped together
  • Check inference treats retained + new balls as single group
  • Verify KL-divergence calculated correctly
  • Console: Look for “shuffle” notifications with all balls

Test 7: Rotation

• Rotate shuffler with R key
• Verify outputPattern sides rotate (retained entries unchanged)
• Check connections update for all output sides
• Verify retained entries don’t get rotated

Expected Issues: Line 293 warning about >3 balls (not a bug, just warning) Actual Issues: [Document any issues found]

✓ Filter Component (Configurable Outputs)

Test Levels:

• comprehensive-test-level.json (ID 11)
• edge-cases-test-level.json (ID 14: all match scenario)

Test 1: Custom Color Filtering

• Red Filter (comprehensive-test, ID 11)
  • targetColor=’red’
  • Verify only red balls go to match output
  • Non-red balls go to non-match output
• Green Filter (edge-cases-test, ID 14)
  • All balls are green, targetColor=’green’
  • Verify all balls go to match output (left)
  • Non-match output should be empty

Test 2: Output Side Configuration

• Standard (matchOutputSide=’left’, nonMatchOutputSide=’right’)
  • Verify match balls go left
  • Non-match balls go right
• Swapped (matchOutputSide=’right’, nonMatchOutputSide=’left’)
  • Create test with swapped sides
  • Verify routing reverses correctly

Test 3: Capacity and Buffering

• capacity=3: Wait for 3 balls, then output
• capacity=10: Test with larger buffer (edge-cases-test, ID 14)
• Verify balls buffer before outputting

Test 4: Visual Rendering

• Target color indicator (small circle) shows correct color
• Buffer status: “2/3” display
• Rendering correct in all 4 directions
• Plex glass overlay if enabled

Test 5: Rotation

• Rotate filter with R key
• Verify direction parameter updates
• Check connections update correctly

Expected Issues: None Actual Issues: [Document any issues found]

✓ Merger Component

Test Levels:

• comprehensive-test-level.json (ID 14)
• inference-chain-test-level.json (ID 23: merges 2 sources)

Test 1: Two-Source Merge

• Basic Merge (comprehensive-test, ID 14)
  • Receives from shuffler (down) and filter (left)
  • Balls from both sources should merge into single stream
  • Verify no balls lost
• Four-Source Merge (inference-chain-test, ID 23)
  • Receives from 2 conveyors (from 2 different sacks)
  • All balls should merge correctly
  • Check ordering is reasonable (FIFO-ish)

Test 2: Bayesian Tracking

• Balls from different sources maintain separate sackIds
• Verify dependencies don’t incorrectly merge
• Check inference handles mixed-source balls correctly

Test 3: Rotation

• Rotate merger with R key
• Verify connections update
• Check all input sides still work

Expected Issues: None Actual Issues: [Document any issues found]

✓ Conveyor-Turn Component

Test Levels: Both comprehensive-test and edge-cases-test have multiple turns

Test 1: All 8 Turn Types

Test each turn type in isolation:

• right-to-down
• right-to-up
• left-to-down
• left-to-up
• down-to-right
• down-to-left
• up-to-right
• up-to-left

Test 2: Direction Validation (Task 3)

• Balls entering from correct side should pass through
• Check console for warnings if ball enters from wrong side
• Verify trajectory curves correctly

Test 3: Rotation

• Rotate turn with R key
• Verify cycles through all 8 types
• Check connections update

Expected Issues: None Actual Issues: [Document any issues found]

Phase 2: Integration Tests

✓ Complete Chain Test

Test Level: comprehensive-test-level.json

Full Flow Testing

• Path 1: Sack → Duplicator → Observation (top path, 3 copies)
  • Verify 3 balls reach top observation
• Path 2: Sack → Duplicator → Shuffler → Filter → Observations
  • 2 balls to shuffler from duplicator
  • Shuffler receives from 2 inputs (duplicator + turn)
  • Shuffler outputs 2 right + 1 down, retains 2
  • Filter separates by color
  • Verify ball counts at both filter outputs
• Path 3: Duplicator → Merger ← Blue Sack → Final Observation
  • 1 ball from duplicator’s down output
  • Balls from blue sack
  • Merger combines both sources
  • Verify mixed balls at final observation
• Path 4: Duplicator up → Black Pit
  • 1 ball disappears into pit
  • No errors in console

Cross-Component Interactions

• Duplicator feeds multiple paths simultaneously
• Shuffler receives from multiple sources
• Filter receives shuffled balls
• Merger combines balls from different paths
• No balls stuck or lost
• No trajectory errors

Expected Issues: None Actual Issues: [Document any issues found]

✓ Edge Cases Test

Test Level: edge-cases-test-level.json (50 balls spawned)

Overflow Scenarios

• Observation overflow (30+ balls to 25-capacity grid)
  • Verify graceful handling
  • Check for console warnings/errors
• Duplicator max copies (10 copies)
  • All copies created successfully
  • No performance issues
• Shuffler max retention (retain 5, output 5)
  • Multiple cycles work correctly
  • Retained balls don’t accumulate incorrectly
• Filter max capacity (capacity=10)
  • Large buffer works correctly
  • Outputs all balls eventually

Minimal Configurations

• Shuffler with 1 input, minBufferSize=1
  • Acts like pass-through when minimal
  • No errors

Empty Scenarios

• Filter where all balls match target color
  • Non-match observation remains empty
  • No errors for empty output

Expected Issues: Observation overflow behavior may vary Actual Issues: [Document any issues found]

✓ Bayesian Inference Test

Test Level: inference-chain-test-level.json

Hypothesis Space

• 4 sacks with 2 alternatives each = 16 total hypotheses
• Betting interface shows all 16 hypotheses
• Initial probabilities: 1/16 each (≈6.25%)

Duplication Tracking

• Duplicator (ID 3, has plex glass)
  • No observations inside duplicator
  • Copies maintain sackId association
  • All 4 copies traced to original sack
  • Check console for tracking logs

Shuffler Retention Dependencies

• Shuffler (ID 5, retains 2 of 4 balls)
  • Balls shuffled together create dependencies
  • Retained balls shuffle with new arrivals next cycle
  • Cross-cycle dependencies: Retained ball from cycle N shuffles with new ball in cycle N+1
  • Verify both balls treated as same group for inference
  • Check console: “shuffle” events show all balls being shuffled

Filter Color Routing

• Filter (ID 11, has plex glass, targets red)
  • No observations inside filter
  • Red balls → right observation (ID 12)
  • Non-red balls → left observation (ID 13)
  • Verify dependencies tracked through filtering

Multi-Source Merger

• Merger (ID 23, combines 2 sacks)
  • Balls from sacks 26 and 29 merge
  • Both sources maintain separate sack identities
  • No incorrect dependency grouping

Arm Sampling

• Arm (ID 19, sampleRate=0.5)
  • Approximately 50% of balls sampled
  • Observations in ID 30 trigger inference updates
  • Check betting interface updates

Inference Accuracy

• Make bet on hypothesis space
• Run simulation completely
• Observe balls entering observation points
• Verify posterior probabilities update correctly
• Colors observed should eliminate incompatible hypotheses
• Final KL-divergence score calculated
• Check score reflects bet accuracy

Console Verification

• No “Missing sackId” warnings (Task 4 might show some)
• Shuffle events logged with ball counts
• Inference updates logged when balls observed
• Dependency groups shown in console

Expected Issues:

• Line 293 warning about >3 balls in shuffler (not a bug)
• Task 4: Possible defensive warnings about missing sackId (investigation needed)

Actual Issues: [Document any issues found]

Phase 3: Editor Tests

✓ Parameter Configuration UI

Duplicator Properties Panel

• Number of copies input (2-10 range)
• Output pattern editor button opens JSON prompt
• Pattern display shows formatted distribution
• Validation catches invalid patterns:
  • Invalid sides (not right/down/left/up)
  • Negative counts
  • Non-integer counts
• Auto-update connections after pattern change
• Changes saved correctly

Shuffler Properties Panel

• minBufferSize input (1-10 range)
• outputSide dropdown (backward compatibility)
• Output pattern editor button opens JSON prompt
• Pattern display shows output + retention
• Validation catches invalid patterns:
  • Invalid sides
  • Invalid retain entries
  • Total count mismatch with minBufferSize (offers to update)
• Auto-update connections after pattern change
• Changes saved correctly

Filter Properties Panel

• targetColor selector/input
• matchOutputSide dropdown (left/right)
• nonMatchOutputSide dropdown (left/right)
• capacity input
• outputDelay input
• Plex glass checkbox
• Changes applied immediately

General UI

• All property panels update when selecting components
• Changes render immediately on canvas
• Properties persist after save/load
• Undo/redo works with parameter changes

Expected Issues: None Actual Issues: [Document any issues found]

✓ Rotation System

Component Rotation (R Key)

• Duplicator: All sides in outputPattern rotate
• Shuffler: Output sides rotate, retained entries unchanged
• Filter: Direction rotates correctly
• Merger: Direction rotates correctly
• Conveyor: Direction rotates correctly
• Conveyor-turn: Cycles through all 8 types correctly
• Splitter: Direction rotates correctly

Pending Component Rotation (R Key Before Placement)

• Preview shows rotated orientation
• All components rotate in preview mode
• Placement uses rotated orientation
• Auto-connections created correctly for rotated component

Auto-Connection Updates

• Connections update immediately after rotation
• Multi-output components (duplicator, shuffler) create all connections
• Old connections removed correctly
• New connections added correctly
• No orphaned connections
• No duplicate connections

Complex Rotation Tests

• Rotate component in middle of long chain
  • Connections before and after update correctly
• Rotate multi-output component with 3+ connections
  • All connections remap correctly
• Rotate shuffler with retention
  • Output connections update, retention unaffected
• Performance with 20+ components
  • No lag when rotating
  • All connections update correctly

Expected Issues: None Actual Issues: [Document any issues found]

✓ Save/Load System

Level Saving

• Create level with all new parameters:
  • Duplicator with multi-output pattern
  • Shuffler with retention pattern
  • Filter with custom colors and sides
  • Observation with custom ballScale
• Save level to localStorage
• Verify success message

Level Loading

• Close editor (or refresh page)
• Open level from localStorage
• Verify all components loaded correctly:
  • Duplicator outputPattern preserved
  • Shuffler outputPattern and minBufferSize preserved
  • Filter custom parameters preserved
  • Observation ballScale preserved
• Verify connections preserved correctly
• Test simulation works with loaded level

Backward Compatibility

• Load old levels (if any exist)
• Components with old parameter format should still work
• Shuffler: outputSide converts to outputPattern
• Duplicator: direction converts to outputPattern
• No errors in console

JSON Export/Import

• Export level to JSON
• Verify JSON structure matches expected format
• Import JSON level
• Verify level works correctly after import

Expected Issues: None (backward compatibility built in) Actual Issues: [Document any issues found]

Phase 4: Performance & Polish

Performance Tests

• Large Level (20+ components)
  • Create complex level with many components
  • No lag during simulation
  • Canvas rendering smooth (60fps)
  • Auto-connection updates fast (<100ms)
• Many Balls (50+ balls)
  • Spawn 50+ balls
  • Simulation runs smoothly
  • No memory leaks
  • Trajectory calculations performant
• Long Chains (10+ component chain)
  • Ball passes through 10+ components
  • Trajectories smooth
  • Bayesian tracking works correctly
  • No accumulated errors

Console Cleanliness

• Run each test level completely
• Check browser console for:
  • Red errors (none expected except Task 4 investigation)
  • Yellow warnings (document any unexpected ones)
  • Excessive logging (reduce if too verbose)

Visual Polish

• All components render correctly
• Multi-output indicators clear and readable
• Retention indicators visible but not distracting
• Ball colors distinct
• Canvas grid lines clear
• Component icons recognizable
• Observation grids neat and organized

User Experience

• Component placement feels responsive
• Selection feedback clear
• Properties panel easy to use
• Pattern editors provide good feedback
• Error messages helpful
• Rotation feels intuitive
• Connections visually clear

Expected Issues: None Actual Issues: [Document any issues found]

Summary Checklist

Critical Tests (Must Pass)

• Duplicator multi-output distribution deterministic
• Shuffler retention works across cycles
• Filter routes by color correctly
• Observation handles 25+ balls
• Rotation updates all parameters correctly
• Auto-connections work for multi-output components
• Bayesian inference tracks through complex chains
• Save/load preserves all new parameters

Important Tests (Should Pass)

• All edge cases handle gracefully
• No console errors during normal operation
• Performance acceptable with large levels
• UI provides good feedback
• All property panels functional

Nice-to-Have Tests (Optional)

• Extreme edge cases (100+ balls, 50+ components)
• Backward compatibility with old levels
• Visual polish and aesthetics
• Advanced Bayesian inference scenarios

Issues Log

Known Issues (from TODO)

1. Task 4: Duplicator ballToSack tracking may have defensive warnings (needs proper fix)
2. Line 293 Warning: Shuffler warns about >3 balls (not a bug, just informational)

New Issues Found

[Document any new issues discovered during testing]

Format:

Issue #: [Brief title]
Severity: [Critical/High/Medium/Low]
Component: [Which component(s)]
Reproduction: [Steps to reproduce]
Expected: [What should happen]
Actual: [What actually happens]
Console Output: [Any error messages]

Testing Notes

• Test on multiple browsers (Chrome, Firefox, Safari)
• Clear localStorage before testing backward compatibility
• Use browser DevTools for performance profiling
• Take screenshots of visual bugs
• Record console logs for errors
• Note any unexpected behavior even if “working”

Last Updated: 2025-11-16 Status: Ready for comprehensive testing