v2.63.0 — Agentic Advisors, Performance Hot-Loops & Build Hardening

v2.63.0 — Agentic Advisors, Performance Hot-Loops & Build Hardening

This release rolls up 79 commits since v2.62.0, dominated by a new wave of
autonomous "agentic" analysis engines, targeted performance fixes in hot paths,
and a healthier build/CI surface (Java 11, repaired javadoc/javac errors,
tighter CodeQL configuration, expanded test coverage).

✨ New agentic advisors / engines (10+)

• GraphCascadingFailureAdvisor — agentic cascading-failure resilience advisor with 240 unit tests
• GraphInfluenceSeedAdvisor — SPREAD/CONTAINMENT seed-selection advisor (238 tests)
• GraphPrivacyExposureAuditor — autonomous re-identification risk auditor
• GraphAdversaryForecaster — adversarial-intent modeler
• GraphSynchronizationEngine — autonomous Kuramoto oscillator simulation
• GraphFairnessAuditEngine — autonomous network equity analyzer
• GraphInformationDiffusionEngine — autonomous cascade simulation
• GraphSpectralClusteringEngine — autonomous spectral partitioning
• GraphNetworkControllabilityEngine — structural controllability analysis
• GraphPercolationEngine — 7-engine percolation analysis with health scoring & HTML dashboard (46 tests)
• GraphInfluenceCampaignPlanner — strategic influence maximization
• GraphAutonomousRepairEngine — structural weakness detection & repair planning
• GraphOpinionDynamicsEngine — autonomous opinion formation simulation
• GraphTemporalDynamicsEngine — autonomous temporal network analysis
• GraphGameTheoryEngine — 6 engines (Shapley, Nash, coalitions, bargaining power, strategic position) + 28 tests
• GraphKnowledgeExtractor — autonomous knowledge extraction & link prediction
• GraphTopologyHypothesisTester — structural classification
• NetworkImmunizationPlanner — autonomous epidemic containment
• GraphSentinel — autonomous structural drift detector
• GraphEvolutionSimulator — 5 growth models + tipping-point detection
• GraphAutoPilot — 6 weakness diagnostics, 5 action types, health scoring, impact simulation (28 tests)
• GraphIntelligenceAdvisor — proactive recommendations via structural fingerprinting

🚀 Performance

• EdgePersistenceAnalyzer — memoize classify() to eliminate O(W·E) re-scans (#168)
• NMI/ARI — O(n) contingency tables; replaced O(k_A·k_B) set-copy intersections
• IndependentSetAnalyzer — eliminate dead loop, drop to O(Σ|MIS|) accumulation, adjacency-backed IS check
• GraphGameTheoryEngine — cache coalition values in greedy merge
• NetworkImmunizationPlanner — O(|infected|) SIR simulation (was O(V) per step); precompute centrality rankings once
• GraphKnowledgeExtractor — cache edgeSet at construction
• InfluenceSpreadSimulator — eliminate per-edge String allocation in Monte Carlo hot loops
• IndependentSetAnalyzer — adjacency-backed IS check
• LT diffusion — incremental active-neighbor tracking
• EdgeBetweennessAnalyzer — O(1) edge lookup via pre-built index
• NodeSimilarityAnalyzer — O(V·Δ²) 2-hop pair enumeration
• GraphColoringAnalyzer — O(1) adjacency and degree lookups
• GraphDrawingQualityAnalyzer — AABB early rejection for edge-crossing detection
• Dijkstra PQ — replaced String[]/parseDouble entries with typed DijkstraEntry

🔒 Security / correctness

• HTML-escape user-controlled vertex names in HTML reports (CWE-79)
• computeNormalizedCut — use intValue() for Integer comparison (#166)
• Replaced broken QR with Colt EigenvalueDecomposition (#169)
• RandomWalkAnalyzer — guard against null getNeighbors (NPE on isolated vertices)
• Intervention planner — partner selection + coupling sweep reproducibility
• GraphComplementAnalyzer — fix wrong Edge constructor argument order

🧰 Refactor

• IndexedGraph — extract subset-aware constructor; dedupe in WienerIndexCalculator
• GraphSentinel — hoist degree/hub computation out of sub-methods
• GraphProductCalculator.strongProduct — rewrite via edge iteration + tests
• GraphSparsificationAnalyzer — extract newEmptyShell + addEdgeCopy
• Shared GraphUtils — deduplicate buildAdjacency + BFS

🧹 Cleanup

• Remove 83 lines of dead/unreachable private methods + unused imports
• Remove 15 unused imports and 4 redundant escapeHtml wrappers across 15 files
• Remove 9 stale duplicate test files under src/test/gvisual (–2,345 LoC)
• Replace remaining zalenix string refs with sauravbhattacharya001
• @author Javadoc and vendor metadata normalized

🛠 Build / CI

• Repair main javac and pom test exclude; bump CI to Java 11
• Repair two illegal Javadoc escapes blocking mvn compile
• Harden CI build; add QuadTree/EdgeType tests
• Harden CodeQL config; add ThresholdConfig and AnalysisResult tests
• Maven Wrapper added; post-publish verification job in publish.yml
• Cross-platform installer scripts + Scoop manifest

🧪 Tests added / expanded

• GraphCascadingFailureAdvisorTest (240), GraphInfluenceSeedAdvisorTest (238)
• GraphMotifFinderTest (18), GraphDegreeSequenceRandomizerTest (16)
• GraphComplementAnalyzer, DominatingSetAnalyzer, EdgeType, QuadTree,
  NetworkFlowExporter, RandomGraphGenerator, ThresholdConfig,
  AnalysisResult — comprehensive coverage
• coverage.yml + codecov.yml configured
• Auto-discover tests; 3 new test suites

📦 Packaging / release infra

• Issue templates: CI failure + documentation
• Issue auto-triage workflow with content-based labeling
• GHCR / Security Policy / Copilot Ready / Auto Label / Stale Bot badges
• Documentation pages for 19 missing analyzers + Transformation & Special-Classes sections
• Local CI reproduction, benchmarking, and dependency management docs in CONTRIBUTING.md
• Comprehensive CONTRIBUTING.md rewrite with full 147-class module catalog

📈 Stats

• 79 commits since v2.62.0
• 25 feat, 13 perf, 6 fix, 7 refactor, 7 test, 10 docs (plus chore/ci/cleanup/security/deps)

Full diff: v2.62.0...v2.63.0

v2.62.0 — Graph Partitioning Visualizer, GNN Playground & Narrative Generator

What's New

✨ Features

• Interactive Graph Partitioning Visualizer — Kernighan-Lin algorithm with real-time visualization of partition refinement steps
• GNN Playground — interactive Graph Neural Network message-passing visualization for exploring node embeddings and layer propagation
• Graph Narrative Generator — autonomous natural-language report generation from graph structure and metrics

⚡ Performance

• O(1) neighbor index lookups in face-walk dart tracing (planar embedding)
• O(V·Δ²) 4-cycle counting in MotifAnalyzer via 2-path enumeration (previously O(V²·E))

🔧 Refactoring

• Eliminated redundant adjacency map rebuilds in ChordalGraphAnalyzer
• Deduplicated Tarjan bridge-finding into shared \GraphUtils.findBridges\

📚 Documentation

• Comprehensive TESTING.md guide with coverage gap inventory of 40 untested classes
• Updated README testing section

Full Changelog: v2.61.0...v2.62.0

v2.61.0 — Security Hardening, Resilience BFS & Planar Refactor

What's Changed

Security

• HTML-escape vertex names in Swing panel controllers — prevents CWE-79 XSS via malicious vertex names rendered in JLabel HTML (3 controllers patched)
• Pre-read file size check in GraphAnnotationManager — prevents CWE-400 OOM denial-of-service when loading oversized annotation files

Performance

• Single-pass BFS in GraphResilienceAnalyzer — computes component count and largest component size in one traversal instead of separate passes

Refactoring

• Deduplicate PlanarGraphAnalyzer.countComponents — delegates to existing GraphUtils implementation, removing ~30 lines of duplicated BFS code

Full Changelog: v2.60.0...v2.61.0

v2.60.0 — Link Prediction, Chromatic Polynomial Tests & Steiner Tree Performance

What's Changed

Performance

• SteinerTreeAnalyzer: O(1) edge lookup via HashSet + linear-time tree pruning replacing O(V²·E) degree recount loops
• LinkPredictionAnalyzer: 2-hop pair enumeration in predict() and predictEnsemble() — dramatically reduces candidate generation for sparse graphs

Refactoring

• TopologicalSortAnalyzer: Eliminated redundant Kahn's algorithm pass — cycle detection now reuses the primary topological sort traversal instead of running a separate BFS

Testing

• ChromaticPolynomialCalculator: Comprehensive test suite covering deletion-contraction, complete graphs, bipartite graphs, trees, and edge cases

Full Changelog: v2.59.0...v2.60.0

v2.59.0 — Interactive Motif Census with Significance Testing

What's Changed

New Feature

• Interactive Motif Census — Full subgraph pattern analysis tool with statistical significance testing
  • Enumerate 6 motif types: triangle, wedge, 3-star, square, diamond, K4
  • 8 graph presets: Karate Club, social network, star, ring, random, complete, bipartite, scale-free
  • Custom edge list input for user-defined graphs
  • Canvas force-directed visualization with motif instance highlighting
  • Census bar chart for pattern frequency comparison
  • Erdős–Rényi random graph ensemble comparison with z-score significance profiling
  • Autonomous proactive insights engine: clustering coefficient analysis, hub detection, network type classification
  • Interactive features: drag nodes, pan/zoom, click motif chips to highlight specific instances

Full Changelog: v2.58.0...v2.59.0

v2.58.0 — Graph Partitioner KL Refinement Performance

What's Changed

Performance

• GraphPartitioner KL refinement optimization — Two significant improvements to
  efineKLPair:
  1. Pre-built neighbor HashSets: Neighbor sets are now constructed once outside the pass loop instead of being rebuilt per vertex per pass, eliminating O(passes × |A| × avg_degree) redundant allocations and copies
  2. Removed redundant list mutations: Eliminated O(|A|) + O(|B|) linear-scan \ArrayList.remove(Object)\ calls after each swap — the locked set already prevents swapped vertices from being revisited, making the removals unnecessary. For partitions of size P, this saves O(P²) total work across all swap passes

Full Changelog: v2.57.0...v2.58.0

v2.57.0 — What-If Analyzer, Influence Maximizer, Anomaly Detection Performance

What's New

✨ Graph What-If Analyzer

Interactive hypothetical modification tool — simulate edge/vertex additions and removals to see how graph metrics (centrality, connectivity, community structure) change before committing changes.

✨ Influence Maximizer

Autonomous seed selection for maximum information spread using greedy hill-climbing with marginal gain optimization. Supports multiple diffusion models (Independent Cascade, Linear Threshold).

⚡ Performance: GraphAnomalyDetector adjacency pre-computation

Pre-builds adjacency sets once during \computeRawMetrics()\ and reuses across clustering coefficient, neighbor deviation, and degree computations. Eliminates O(V) redundant JUNG \getNeighbors()\ collection allocations per vertex in the triangle-counting inner loop and removes per-vertex ArrayList + HashSet construction overhead.

Full Changelog: v2.56.0...v2.57.0

v2.56.0 — Graph Symmetry Explorer, Community Evolution & Performance Caching

What's New

✨ Features

• Graph Symmetry Explorer — automorphism detection, orbit computation & symmetry analysis for understanding structural symmetries in graphs
• Community Evolution Tracker — interactive community lifecycle visualization showing how communities form, merge, split, and dissolve over time

⚡ Performance

• GraphSparsificationAnalyzer: cached edge betweenness and bridge computation — eliminates redundant O(V·E) recalculations across repeated sparsification queries

🔧 Refactoring

• IndependentSetAnalyzer: cached adjacency map — eliminated 7+ redundant O(V+E) adjacency map rebuilds across greedy, backtracking, and analysis methods

Full Changelog: v2.55.0...v2.56.0

v2.55.0 — Interactive Visualizers, Perf Overhaul & Refactoring

Highlights

🎮 New Interactive Visualizers

• SCC Finder — Tarjan's algorithm with step-by-step animation, stack visualization, and 6 preset directed graphs
• Crossing Number Game — drag nodes to minimize edge crossings with 8 presets (Petersen, K5, K3,3, Dodecahedron, etc.), auto-solve, timer & rating
• Maximum Clique Finder — Bron-Kerbosch visualizer with R/P/X set coloring, 7 presets, and recursion logging
• Hamiltonian Path & Circuit Finder — backtracking with Warnsdorff heuristic, 8 presets, real-time animation
• Graph Storyteller — autonomous narrative generator analyzing structure and producing multi-section reports (communities, vulnerabilities, forecasts)

⚡ Performance Improvements

• Chordal analyzer: O(V+E) bucket-queue MCS replacing O(V²) linear scan; shared adjacency map eliminates up to 9 redundant rebuilds per analysis
• Chromatic polynomial: pendant vertex stripping + tree detection shortcut; batch pendant stripping + exponentiation by squaring for polynomial powers
• Dominating set: bucket-based priority for O(Δ²) amortized greedy instead of O(V²) per-round scan; incremental k-domination scoring
• Motif detection: shared pre-computed neighbor sets across all detectors; O(1) HashSet lookups replace O(degree) isNeighbor calls
• Link prediction: streaming top-K ensemble (O(K) memory vs O(V²)); allocation-free Jaccard via arithmetic union
• Resilience: bucket-queue degree attack replaces O(V²) linear scan
• Distance report: consolidate 8+ O(V²) passes into single pass
• Anomaly: O(1) node lookup via HashMap index; allocation-free counting
• Network profiler: pre-built neighbor sets for O(k²) triangle checks; reuse maxDegree
• Graph query engine: replace nested stream filters with explicit labeled loops (~2-3x faster)
• Vertex cover: eliminate O(E) string allocations in approx algorithm

🔧 Refactoring

• Slater ranking: replace O(n!) permutation enumeration with branch-and-bound search
• LocationResolver: batch SQL updates with transaction control
• Chordal public API: deduplicate adjacency map construction
• DominatingSetAnalyzer: extract helpers, optimize independence check
• NetworkFlowAnalyzer: remove dead code, cache flow decomposition
• RandomGraphGenerator: fix broken Edge constructor
• GraphSampler: extract 4 shared helpers, -76 lines
• Motif detector: share neighbor sets, deduplicate local clustering
• ShortestPathFinder: deduplicate DijkstraEntry with GraphUtils

🐛 Bug Fixes

• Fix broken \degCV\ field name → \degreeCV\ in network classify()
• Fix broken Edge constructor in RandomGraphGenerator
• Fix path traversal, resource leak, and encoding in resilience CSV export
• Fix stray closing brace in computeFillIn (compilation error)

🧪 Testing

• EdgeBetweennessAnalyzer: 17 comprehensive test cases
• DimacsExporter: 11 test cases covering edge cases and null handling
• GraphPowerCalculator: 20 tests (power computation, caching, error handling)
• PerfectGraphAnalyzer: 15 tests (perfect/imperfect classification, odd holes)

🔄 CI

• Bump actions/checkout v4→v6, stale v9→v10, deploy-pages v4→v5, upload-pages-artifact v4→v5, github-script v7→v9, setup-qemu-action v3→v4
• Updated copilot-instructions.md with accurate project state (272 Java files, 105 test suites)

v2.54.0 — Migration Snapshot Performance

What's Changed

Performance

• Reuse previous snapshot map in getMigrationCounts — avoids rebuilding the community-to-node lookup for the prior snapshot on every call, reducing O(V) redundant work per migration count query
• Single-pass getSummary — combined separate iteration passes into one traversal for computing community summary statistics

Full Changelog: v2.53.0...v2.54.0