mirror of
https://github.com/ruvnet/RuVector.git
synced 2026-05-28 18:13:33 +00:00
cfdcb8bb12
AC-3a now publishes full-partition ARI alongside the 2-way
coarsening. Measured on the default N=1024 SBM:
2-way coarsened ARI (inherited, backward-compat):
mincut : -0.001 greedy : 0.174
louvain : 0.000 leiden : 0.089
**Full-partition ARI (new, correct metric):**
greedy full_ari : **0.308** ← surprising
louvain full_ari : 0.000 (collapses)
leiden full_ari : 0.107
cpm@γ=2.25 : **0.425** ← still best
**20th discovery: Leiden's aggregation+refinement actively HURTS
full-partition ARI vs greedy level-1 on this substrate.** Greedy
modularity (one pass of local moves, no aggregation) scores 0.308;
adding the aggregation + Traag refinement steps drops that to
0.107 — a 2.9× regression from the more sophisticated algorithm.
The refinement preserves well-connectedness (leiden_refinement.rs
tests still pass) but does so at the cost of merging structurally-
distinct communities from the level-1 output.
This flips the expected order: on hub-heavy SBMs, *more algorithm
is worse* when the objective is modularity and the target is
module recovery. CPM (item 17) was the right escape — non-
resolution-limited objective sidesteps the issue.
Final ranking on default SBM, full-partition ARI:
CPM @ γ=2.25 : 0.425 (non-modularity objective)
greedy L1 : 0.308 (minimal-algorithm modularity)
Leiden : 0.107 (maximal-algorithm modularity)
Louvain : 0.000 (aggregation collapses)
The pattern echoes item 11 (multi-level Louvain collapse on
hub-heavy SBMs) but at a finer granularity: item 11 said
'aggregation breaks', item 20 says 'even Leiden's refinement
can't fully repair it because the underlying modularity
objective has the resolution-limit issue'. The fix (item 17)
was a different objective, not a better algorithm.
Engineering implication: **for AC-3a on this substrate, level-1
greedy modularity is a stronger baseline than multi-level
Leiden.** The default Louvain / Leiden trajectory assumes
increasingly-sophisticated algorithms monotonically improve
module recovery; on hub-heavy SBMs that assumption is false,
and simpler-is-better up to the CPM break.
Files:
- tests/acceptance_partition.rs: full_partition_ari helper,
new eprintln publishing four full-ARI values against ground-
truth module labels. No assertion change (ADR §14 threshold
discipline: coarsening choices are decisions, not knobs).
- docs/adr/ADR-154: §17 item 20 added with the surprising
level-1 vs Leiden inversion and the 'more algorithm is
worse' framing on this substrate.
All 95 prior tests unchanged.
Co-Authored-By: claude-flow <ruv@ruv.net>
EOF
)
|
||
|---|---|---|
| .. | ||
| agentic-jujutsu | ||
| apify | ||
| app-clip | ||
| benchmarks | ||
| boundary-discovery | ||
| brain-boundary-discovery | ||
| climate-consciousness | ||
| cmb-boundary-discovery | ||
| cmb-consciousness | ||
| connectome-fly | ||
| data | ||
| decompiler-dashboard | ||
| delta-behavior | ||
| dna | ||
| docs | ||
| dragnes | ||
| earthquake-boundary-discovery | ||
| ecosystem-consciousness | ||
| edge | ||
| edge-full/pkg | ||
| edge-net | ||
| exo-ai-2025 | ||
| frb-boundary-discovery | ||
| gene-consciousness | ||
| google-cloud | ||
| graph | ||
| gw-consciousness | ||
| health-boundary-discovery | ||
| infrastructure-boundary-discovery | ||
| market-boundary-discovery | ||
| meta-cognition-spiking-neural-network | ||
| mincut | ||
| music-boundary-discovery | ||
| neural-trader | ||
| nodejs | ||
| onnx-embeddings | ||
| onnx-embeddings-wasm | ||
| OSpipe | ||
| pandemic-boundary-discovery | ||
| prime-radiant | ||
| pwa-loader | ||
| quantum-consciousness | ||
| real-eeg-analysis | ||
| real-eeg-multi-seizure | ||
| refrag-pipeline | ||
| robotics | ||
| rust | ||
| ruvLLM | ||
| rvf | ||
| rvf-desktop | ||
| rvf-kernel-optimized | ||
| scipix | ||
| seizure-clinical-report | ||
| seizure-therapeutic-sim | ||
| seti-boundary-discovery | ||
| seti-exotic-signals | ||
| spiking-network | ||
| subpolynomial-time | ||
| temporal-attractor-discovery | ||
| train-discoveries | ||
| ultra-low-latency-sim | ||
| verified-applications | ||
| vibecast-7sense | ||
| void-boundary-discovery | ||
| vwm-viewer | ||
| wasm/ios | ||
| wasm-react | ||
| wasm-vanilla | ||
| weather-boundary-discovery | ||
| bounded_instance_demo.rs | ||
| README.md | ||
| security_hardened.rvf | ||
RuVector Examples
Comprehensive examples demonstrating RuVector's capabilities across multiple platforms and use cases.
Directory Structure
examples/
├── rust/ # Rust SDK examples
├── nodejs/ # Node.js SDK examples
├── graph/ # Graph database features
├── wasm-react/ # React + WebAssembly integration
├── wasm-vanilla/ # Vanilla JS + WebAssembly
├── agentic-jujutsu/ # AI agent version control
├── exo-ai-2025/ # Advanced cognitive substrate
├── refrag-pipeline/ # Document processing pipeline
└── docs/ # Additional documentation
Quick Start by Platform
Rust
cd rust
cargo run --example basic_usage
cargo run --example advanced_features
cargo run --example agenticdb_demo
Node.js
cd nodejs
npm install
node basic_usage.js
node semantic_search.js
WebAssembly (React)
cd wasm-react
npm install
npm run dev
WebAssembly (Vanilla)
cd wasm-vanilla
# Open index.html in browser
Example Categories
| Category | Directory | Description |
|---|---|---|
| Core API | rust/basic_usage.rs |
Vector DB fundamentals |
| Batch Ops | rust/batch_operations.rs |
High-throughput ingestion |
| RAG Pipeline | rust/rag_pipeline.rs |
Retrieval-Augmented Generation |
| Advanced | rust/advanced_features.rs |
Hypergraphs, neural hashing |
| AgenticDB | rust/agenticdb_demo.rs |
AI agent memory system |
| GNN | rust/gnn_example.rs |
Graph Neural Networks |
| Graph | graph/ |
Cypher queries, clustering |
| Node.js | nodejs/ |
JavaScript integration |
| WASM React | wasm-react/ |
Modern React apps |
| WASM Vanilla | wasm-vanilla/ |
Browser without framework |
| Agentic Jujutsu | agentic-jujutsu/ |
Multi-agent version control |
| EXO-AI 2025 | exo-ai-2025/ |
Cognitive substrate research |
| Refrag | refrag-pipeline/ |
Document fragmentation |
Feature Highlights
Vector Database Core
- High-performance similarity search
- Multiple distance metrics (Cosine, Euclidean, Dot Product)
- Metadata filtering
- Batch operations
Advanced Features
- Hypergraph Index: Multi-entity relationships
- Temporal Hypergraph: Time-aware relationships
- Causal Memory: Cause-effect chains
- Learned Index: ML-optimized indexing
- Neural Hash: Locality-sensitive hashing
- Topological Analysis: Persistent homology
AgenticDB
- Reflexion episodes (self-critique)
- Skill library (consolidated patterns)
- Causal memory (hypergraph relationships)
- Learning sessions (RL training data)
- Vector embeddings (core storage)
EXO-AI Cognitive Substrate
- exo-core: IIT consciousness, thermodynamics
- exo-temporal: Causal memory coordination
- exo-hypergraph: Topological structures
- exo-manifold: Continuous deformation
- exo-exotic: 10 cutting-edge experiments
- exo-wasm: Browser deployment
- exo-federation: Distributed consensus
- exo-node: Native bindings
- exo-backend-classical: Classical compute
Running Benchmarks
# Rust benchmarks
cargo bench --example advanced_features
# Refrag pipeline benchmarks
cd refrag-pipeline
cargo bench
# EXO-AI benchmarks
cd exo-ai-2025
cargo bench
Related Documentation
License
MIT OR Apache-2.0