ruvector/crates/ruvector-postgres/tests/routing_tests.rs
rUv 073ce73612
feat(postgres): Add 53 SQL function definitions for all advanced modules (#46)
* feat(postgres): Add 7 advanced AI modules to ruvector-postgres

Comprehensive implementation of advanced AI capabilities:

## New Modules (23,541 lines of code)

### 1. Self-Learning / ReasoningBank (`src/learning/`)
- Trajectory tracking for query optimization
- Pattern extraction using K-means clustering
- ReasoningBank for pattern storage and matching
- Adaptive search parameter optimization

### 2. Attention Mechanisms (`src/attention/`)
- Scaled dot-product attention (core)
- Multi-head attention with parallel heads
- Flash Attention v2 (memory-efficient)
- 10 attention types with PostgresEnum support

### 3. GNN Layers (`src/gnn/`)
- Message passing framework
- GCN (Graph Convolutional Network)
- GraphSAGE with mean/max aggregation
- Configurable aggregation methods

### 4. Hyperbolic Embeddings (`src/hyperbolic/`)
- Poincaré ball model
- Lorentz hyperboloid model
- Hyperbolic distance metrics
- Möbius operations

### 5. Sparse Vectors (`src/sparse/`)
- COO format sparse vector type
- Efficient sparse-sparse distance functions
- BM25/SPLADE compatible
- Top-k pruning operations

### 6. Graph Operations & Cypher (`src/graph/`)
- Property graph storage (nodes/edges)
- BFS, DFS, Dijkstra traversal
- Cypher query parser (AST-based)
- Query executor with pattern matching

### 7. Tiny Dancer Routing (`src/routing/`)
- FastGRNN neural network
- Agent registry with capabilities
- Multi-objective routing optimization
- Cost/latency/quality balancing

## Docker Infrastructure
- Dockerfile with pgrx 0.12.6 and PostgreSQL 16
- docker-compose.yml with test runner
- Initialization SQL with test tables
- Shell scripts for dev/test/benchmark

## Feature Flags
- `learning`, `attention`, `gnn`, `hyperbolic`
- `sparse`, `graph`, `routing`
- `ai-complete` and `graph-complete` bundles

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* fix(docker): Copy entire workspace for pgrx build

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* fix(docker): Build standalone crate without workspace

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* docs: Update README to enhance clarity and structure

* fix(postgres): Resolve compilation errors and Docker build issues

- Fix simsimd Option/Result type mismatch in scaled_dot.rs
- Fix f32/f64 type conversions in poincare.rs and lorentz.rs
- Fix AVX512 missing wrapper functions by using AVX2 fallback
- Fix Vec<Vec<f32>> to JsonB for pgrx pg_extern compatibility
- Fix DashMap get() to get_mut() for mutable access
- Fix router.rs dereference for best_score comparison
- Update Dockerfile to copy pre-written SQL file for pgrx
- Simplify init.sql to use correct function names
- Add postgres-cli npm package for CLI tooling

All changes tested successfully in Docker with:
- Extension loads with AVX2 SIMD support (8 floats/op)
- Distance functions verified working
- PostgreSQL 16 container runs successfully

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* feat: Add ruvLLM examples and enhanced postgres-cli

Added from claude/ruvector-lfm2-llm-01YS5Tc7i64PyYCLecT9L1dN branch:
- examples/ruvLLM: Complete LLM inference system with SIMD optimization
  - Pretraining, benchmarking, and optimization system
  - Real SIMD-optimized CPU inference engine
  - Comprehensive SOTA benchmark suite
  - Attention mechanisms, memory management, router

Enhanced postgres-cli with full ruvector-postgres integration:
- Sparse vector operations (BM25, top-k, prune, conversions)
- Hyperbolic geometry (Poincare, Lorentz, Mobius operations)
- Agent routing (Tiny Dancer system)
- Vector quantization (binary, scalar, product)
- Enhanced graph and learning commands

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* fix(postgres-cli): Use native ruvector type instead of pgvector

- Change createVectorTable to use ruvector type (native RuVector extension)
- Add dimensions column for metadata since ruvector is variable-length
- Update index creation to use simple btree (HNSW/IVFFlat TBD)
- Tested against Docker container with ruvector extension

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* feat(postgres): Add 53 SQL function definitions for all advanced modules

Enable all advanced PostgreSQL extension functions by adding their SQL
definitions to the extension file. This exposes all Rust #[pg_extern]
functions to PostgreSQL.

## New SQL Functions (53 total)

### Hyperbolic Geometry (8 functions)
- ruvector_poincare_distance, ruvector_lorentz_distance
- ruvector_mobius_add, ruvector_exp_map, ruvector_log_map
- ruvector_poincare_to_lorentz, ruvector_lorentz_to_poincare
- ruvector_minkowski_dot

### Sparse Vectors (14 functions)
- ruvector_sparse_create, ruvector_sparse_from_dense
- ruvector_sparse_dot, ruvector_sparse_cosine, ruvector_sparse_l2_distance
- ruvector_sparse_add, ruvector_sparse_scale, ruvector_sparse_to_dense
- ruvector_sparse_nnz, ruvector_sparse_dim
- ruvector_bm25_score, ruvector_tf_idf, ruvector_sparse_normalize
- ruvector_sparse_topk

### GNN - Graph Neural Networks (5 functions)
- ruvector_gnn_gcn_layer, ruvector_gnn_graphsage_layer
- ruvector_gnn_gat_layer, ruvector_gnn_message_pass
- ruvector_gnn_aggregate

### Routing/Agents - "Tiny Dancer" (11 functions)
- ruvector_route_query, ruvector_route_with_context
- ruvector_calculate_agent_affinity, ruvector_select_best_agent
- ruvector_multi_agent_route, ruvector_create_agent_embedding
- ruvector_get_routing_stats, ruvector_register_agent
- ruvector_update_agent_performance, ruvector_adaptive_route
- ruvector_fastgrnn_forward

### Learning/ReasoningBank (7 functions)
- ruvector_record_trajectory, ruvector_get_verdict
- ruvector_distill_memory, ruvector_adaptive_search
- ruvector_learning_feedback, ruvector_get_learning_patterns
- ruvector_optimize_search_params

### Graph/Cypher (8 functions)
- ruvector_graph_create_node, ruvector_graph_create_edge
- ruvector_graph_get_neighbors, ruvector_graph_shortest_path
- ruvector_graph_pagerank, ruvector_cypher_query
- ruvector_graph_traverse, ruvector_graph_similarity_search

## CLI Updates
- Enabled hyperbolic geometry commands in postgres-cli
- Added vector distance and normalize commands
- Enhanced client with connection pooling and retry logic

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
2025-12-02 22:49:29 -05:00

269 lines
8.9 KiB
Rust

// Integration tests for Tiny Dancer Routing module
//
// These tests validate the complete routing functionality including
// agent registration, FastGRNN neural network, and routing decisions.
#[cfg(test)]
mod routing_tests {
use ruvector_postgres::routing::{
agents::{Agent, AgentRegistry, AgentType},
fastgrnn::FastGRNN,
router::{OptimizationTarget, Router, RoutingConstraints},
};
#[test]
fn test_complete_routing_workflow() {
// Create registry and router
let registry = AgentRegistry::new();
let router = Router::with_registry(std::sync::Arc::new(registry));
// Register diverse agents
let agents = vec![
create_agent("gpt-4", 0.03, 500.0, 0.95, vec!["coding", "reasoning"]),
create_agent("claude-3", 0.025, 400.0, 0.93, vec!["coding", "writing"]),
create_agent("gpt-3.5", 0.002, 200.0, 0.75, vec!["general", "fast"]),
create_agent("llama-2", 0.0, 800.0, 0.70, vec!["local", "private"]),
];
for agent in agents {
router.registry().register(agent).unwrap();
}
// Test cost-optimized routing
let request_emb = vec![0.1; 384];
let decision = router
.route(&request_emb, &RoutingConstraints::new(), OptimizationTarget::Cost)
.unwrap();
assert_eq!(decision.agent_name, "llama-2"); // Free option
assert!(decision.confidence > 0.0);
// Test quality-optimized routing
let decision = router
.route(&request_emb, &RoutingConstraints::new(), OptimizationTarget::Quality)
.unwrap();
assert_eq!(decision.agent_name, "gpt-4"); // Highest quality
// Test latency-optimized routing
let decision = router
.route(&request_emb, &RoutingConstraints::new(), OptimizationTarget::Latency)
.unwrap();
assert_eq!(decision.agent_name, "gpt-3.5"); // Fastest
}
#[test]
fn test_routing_with_constraints() {
let registry = AgentRegistry::new();
let router = Router::with_registry(std::sync::Arc::new(registry));
router.registry().register(
create_agent("expensive-high-quality", 1.0, 200.0, 0.99, vec!["coding"])
).unwrap();
router.registry().register(
create_agent("cheap-medium-quality", 0.01, 200.0, 0.75, vec!["coding"])
).unwrap();
let request_emb = vec![0.1; 384];
// Constrain by max cost
let constraints = RoutingConstraints::new()
.with_max_cost(0.5)
.with_min_quality(0.7);
let decision = router
.route(&request_emb, &constraints, OptimizationTarget::Quality)
.unwrap();
// Should pick cheap option due to cost constraint
assert_eq!(decision.agent_name, "cheap-medium-quality");
}
#[test]
fn test_fastgrnn_routing() {
let mut router = Router::new();
router.init_grnn(64);
router.registry().register(
create_agent("agent1", 0.05, 200.0, 0.85, vec!["coding"])
).unwrap();
let request_emb = vec![0.1; 384];
let decision = router
.route(&request_emb, &RoutingConstraints::new(), OptimizationTarget::Balanced)
.unwrap();
// Verify neural network enhanced confidence
assert!(decision.confidence > 0.0);
assert!(decision.confidence <= 1.0);
}
#[test]
fn test_capability_based_routing() {
let registry = AgentRegistry::new();
let router = Router::with_registry(std::sync::Arc::new(registry));
router.registry().register(
create_agent("coder", 0.05, 200.0, 0.90, vec!["coding", "debugging"])
).unwrap();
router.registry().register(
create_agent("writer", 0.03, 150.0, 0.85, vec!["writing", "translation"])
).unwrap();
router.registry().register(
create_agent("generalist", 0.02, 300.0, 0.70, vec!["coding", "writing", "general"])
).unwrap();
let request_emb = vec![0.1; 384];
// Require coding capability
let constraints = RoutingConstraints::new()
.with_capability("coding".to_string());
let decision = router
.route(&request_emb, &constraints, OptimizationTarget::Quality)
.unwrap();
// Should pick specialized coder (highest quality with coding)
assert!(decision.agent_name == "coder" || decision.agent_name == "generalist");
// Verify writer was not selected
assert_ne!(decision.agent_name, "writer");
}
#[test]
fn test_agent_metrics_update() {
let registry = AgentRegistry::new();
let mut agent = create_agent("test-agent", 0.05, 200.0, 0.80, vec!["test"]);
// Initial state
assert_eq!(agent.performance.total_requests, 0);
assert_eq!(agent.performance.avg_latency_ms, 200.0);
// Update with better latency
agent.update_metrics(150.0, true, Some(0.85));
assert_eq!(agent.performance.total_requests, 1);
assert_eq!(agent.performance.avg_latency_ms, 150.0);
assert_eq!(agent.performance.success_rate, 1.0);
// Update with worse latency
agent.update_metrics(250.0, true, Some(0.75));
assert_eq!(agent.performance.total_requests, 2);
assert_eq!(agent.performance.avg_latency_ms, 200.0); // Average of 150 and 250
assert_eq!(agent.performance.success_rate, 1.0);
// Failed request
agent.update_metrics(300.0, false, None);
assert_eq!(agent.performance.total_requests, 3);
assert!(agent.performance.success_rate < 1.0);
}
#[test]
fn test_fastgrnn_sequence_processing() {
let grnn = FastGRNN::new(10, 5);
let sequence = vec![
vec![1.0, 0.0, 0.0, 0.5, -0.5, 0.2, -0.2, 0.8, -0.8, 0.0],
vec![0.0, 1.0, 0.0, -0.5, 0.5, -0.2, 0.2, -0.8, 0.8, 0.0],
vec![0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0],
];
let outputs = grnn.forward_sequence(&sequence);
assert_eq!(outputs.len(), 3);
assert_eq!(outputs[0].len(), 5);
// Verify state evolution (later states should be different from first)
let first_state = &outputs[0];
let last_state = &outputs[2];
let diff: f32 = first_state
.iter()
.zip(last_state.iter())
.map(|(a, b)| (a - b).abs())
.sum();
assert!(diff > 0.0, "Hidden state should evolve across sequence");
}
#[test]
fn test_routing_alternatives() {
let registry = AgentRegistry::new();
let router = Router::with_registry(std::sync::Arc::new(registry));
// Register multiple similar agents
for i in 0..5 {
let quality = 0.7 + (i as f32 * 0.05);
let cost = 0.01 + (i as f32 * 0.01);
router.registry().register(
create_agent(&format!("agent-{}", i), cost, 200.0, quality, vec!["test"])
).unwrap();
}
let request_emb = vec![0.1; 384];
let decision = router
.route(&request_emb, &RoutingConstraints::new(), OptimizationTarget::Quality)
.unwrap();
// Should have alternatives listed
assert!(!decision.alternatives.is_empty());
assert!(decision.alternatives.len() <= 3); // Max 3 alternatives
// Alternatives should have lower scores
for alt in &decision.alternatives {
assert!(alt.score < 1.0);
assert!(!alt.reason.is_empty());
}
}
#[test]
fn test_excluded_agents() {
let registry = AgentRegistry::new();
let router = Router::with_registry(std::sync::Arc::new(registry));
router.registry().register(
create_agent("agent-a", 0.05, 200.0, 0.90, vec!["test"])
).unwrap();
router.registry().register(
create_agent("agent-b", 0.05, 200.0, 0.85, vec!["test"])
).unwrap();
let request_emb = vec![0.1; 384];
// Exclude the best agent
let constraints = RoutingConstraints::new()
.with_excluded_agent("agent-a".to_string());
let decision = router
.route(&request_emb, &constraints, OptimizationTarget::Quality)
.unwrap();
assert_eq!(decision.agent_name, "agent-b");
}
// Helper function to create test agents
fn create_agent(
name: &str,
cost: f32,
latency: f32,
quality: f32,
capabilities: Vec<&str>,
) -> Agent {
let mut agent = Agent::new(
name.to_string(),
AgentType::LLM,
capabilities.iter().map(|s| s.to_string()).collect(),
);
agent.cost_model.per_request = cost;
agent.performance.avg_latency_ms = latency;
agent.performance.quality_score = quality;
agent.embedding = Some(vec![0.1; 384]); // Default embedding
agent
}
}