rust-kgdb 0.6.15 → 0.6.16
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +89 -0
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -12,6 +12,95 @@ Native Rust RDF/SPARQL engine with graph analytics, embeddings, and rule-based r
|
|
|
12
12
|
|
|
13
13
|
---
|
|
14
14
|
|
|
15
|
+
## The Power of Abstraction: Making LLMs Deterministic
|
|
16
|
+
|
|
17
|
+
**The Problem**: Large Language Models are fundamentally non-deterministic. Same question, different answers. No way to verify correctness. No audit trail. No reproducibility.
|
|
18
|
+
|
|
19
|
+
**The Solution**: Mathematical abstraction layers that transform probabilistic LLM outputs into deterministic, verifiable reasoning chains.
|
|
20
|
+
|
|
21
|
+
```
|
|
22
|
+
┌─────────────────────────────────────────────────────────────────────────────┐
|
|
23
|
+
│ FROM PROBABILISTIC LLM TO DETERMINISTIC REASONING │
|
|
24
|
+
│ │
|
|
25
|
+
│ USER QUERY ──────────────────────────────────────────────────────────────▶│
|
|
26
|
+
│ "Find suspicious providers" │
|
|
27
|
+
│ │
|
|
28
|
+
│ ┌─────────────────────────────────────────────────────────────────────┐ │
|
|
29
|
+
│ │ INTELLIGENCE CONTROL PLANE │ │
|
|
30
|
+
│ │ │ │
|
|
31
|
+
│ │ ┌────────────────┐ ┌────────────────┐ ┌────────────────┐ │ │
|
|
32
|
+
│ │ │ CONTEXT THEORY │ │ TYPE THEORY │ │ PROOF THEORY │ │ │
|
|
33
|
+
│ │ │ │ │ │ │ │ │ │
|
|
34
|
+
│ │ │ Spivak's Ologs │ │ Hindley-Milner │ │ Curry-Howard │ │ │
|
|
35
|
+
│ │ │ │ │ │ │ │ │ │
|
|
36
|
+
│ │ │ • Schema as │ │ • Typed tool │ │ • Proofs are │ │ │
|
|
37
|
+
│ │ │ Category │ │ signatures │ │ programs │ │ │
|
|
38
|
+
│ │ │ • Morphisms = │ │ • Composition │ │ • Types are │ │ │
|
|
39
|
+
│ │ │ Properties │ │ validation │ │ propositions │ │ │
|
|
40
|
+
│ │ │ • Functors = │ │ • Compile-time │ │ • Derivation │ │ │
|
|
41
|
+
│ │ │ Transforms │ │ rejection │ │ chains │ │ │
|
|
42
|
+
│ │ └───────┬────────┘ └───────┬────────┘ └───────┬────────┘ │ │
|
|
43
|
+
│ │ │ │ │ │ │
|
|
44
|
+
│ │ └────────────────────┼────────────────────┘ │ │
|
|
45
|
+
│ │ │ │ │
|
|
46
|
+
│ │ ▼ │ │
|
|
47
|
+
│ │ ┌────────────────────────────────────────────────────────────┐ │ │
|
|
48
|
+
│ │ │ HYPERMIND AGENT FRAMEWORK │ │ │
|
|
49
|
+
│ │ │ │ │ │
|
|
50
|
+
│ │ │ User Query → LLM Planner → Typed Execution Plan → Tools │ │ │
|
|
51
|
+
│ │ │ │ │ │
|
|
52
|
+
│ │ │ "Find suspicious" → kg.sparql.query → kg.datalog.apply │ │ │
|
|
53
|
+
│ │ │ → kg.embeddings.search → COMBINE │ │ │
|
|
54
|
+
│ │ └────────────────────────────────────────────────────────────┘ │ │
|
|
55
|
+
│ │ │ │ │
|
|
56
|
+
│ │ ▼ │ │
|
|
57
|
+
│ │ ┌────────────────────────────────────────────────────────────┐ │ │
|
|
58
|
+
│ │ │ rust-kgdb ENGINE │ │ │
|
|
59
|
+
│ │ │ │ │ │
|
|
60
|
+
│ │ │ • GraphDB: SPARQL 1.1 + RDF 1.2 + Hypergraph │ │ │
|
|
61
|
+
│ │ │ • GraphFrames: Distributed analytics (no Spark needed) │ │ │
|
|
62
|
+
│ │ │ • Datalog: Semi-naive evaluation + stratified negation │ │ │
|
|
63
|
+
│ │ │ • Embeddings: HNSW + ARCADE 1-hop cache │ │ │
|
|
64
|
+
│ │ └────────────────────────────────────────────────────────────┘ │ │
|
|
65
|
+
│ │ │ │ │
|
|
66
|
+
│ └────────────────────────────────┼────────────────────────────────────┘ │
|
|
67
|
+
│ │ │
|
|
68
|
+
│ ▼ │
|
|
69
|
+
│ ◀──────────────────────────────────────────────────────────────── OUTPUT │
|
|
70
|
+
│ ProofDAG: Cryptographically-signed derivation chain │
|
|
71
|
+
│ Hash: sha256:8f3a2b1c... (Reproducible, Auditable, Verifiable) │
|
|
72
|
+
└─────────────────────────────────────────────────────────────────────────────┘
|
|
73
|
+
```
|
|
74
|
+
|
|
75
|
+
**How Mathematical Foundations Make This Possible**:
|
|
76
|
+
|
|
77
|
+
| Foundation | Role | What It Provides |
|
|
78
|
+
|------------|------|-----------------|
|
|
79
|
+
| **Context Theory** (Spivak's Ologs) | Schema as Category | Automatic schema detection, semantic validation, consistent interpretation |
|
|
80
|
+
| **Type Theory** (Hindley-Milner) | Typed Tool Signatures | Compile-time validation, prevents invalid tool compositions |
|
|
81
|
+
| **Proof Theory** (Curry-Howard) | Proofs = Programs | Every conclusion has a derivation chain, reproducible reasoning |
|
|
82
|
+
| **Category Theory** | Morphism Composition | Tools as morphisms, validated composition, guaranteed well-formedness |
|
|
83
|
+
|
|
84
|
+
**The Three-Layer Stack**:
|
|
85
|
+
|
|
86
|
+
1. **rust-kgdb** (Foundation) - High-performance knowledge graph database
|
|
87
|
+
- 2.78µs lookup speed (35x faster than RDFox)
|
|
88
|
+
- Native Rust, zero-copy semantics, 24 bytes/triple
|
|
89
|
+
|
|
90
|
+
2. **HyperMind Agent** (Execution) - Schema-aware agent framework
|
|
91
|
+
- LLM Planner with schema injection
|
|
92
|
+
- Typed tool composition (kg.sparql.query, kg.datalog.apply, etc.)
|
|
93
|
+
- Memory management (working, episodic, long-term)
|
|
94
|
+
|
|
95
|
+
3. **Intelligence Control Plane** (Orchestration) - Neuro-symbolic integration
|
|
96
|
+
- Mathematical foundations (Context + Type + Proof Theory)
|
|
97
|
+
- ProofDAG generation for auditability
|
|
98
|
+
- Deterministic LLM outputs through symbolic grounding
|
|
99
|
+
|
|
100
|
+
**Result**: Transform any LLM from a "black box" into a **verifiable reasoning system** where every answer comes with mathematical proof of correctness.
|
|
101
|
+
|
|
102
|
+
---
|
|
103
|
+
|
|
15
104
|
## The ProofDAG: Verifiable AI Reasoning
|
|
16
105
|
|
|
17
106
|
Every HyperMind answer comes with a **ProofDAG** - a cryptographically-signed derivation graph that makes LLM outputs auditable and reproducible.
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "rust-kgdb",
|
|
3
|
-
"version": "0.6.
|
|
3
|
+
"version": "0.6.16",
|
|
4
4
|
"description": "Production-grade Neuro-Symbolic AI Framework with Schema-Aware GraphDB, Context Theory, and Memory Hypergraph: +86.4% accuracy over vanilla LLMs. Features Schema-Aware GraphDB (auto schema extraction), BYOO (Bring Your Own Ontology) for enterprise, cross-agent schema caching, LLM Planner for natural language to typed SPARQL, ProofDAG with Curry-Howard witnesses. High-performance (2.78µs lookups, 35x faster than RDFox). W3C SPARQL 1.1 compliant.",
|
|
5
5
|
"main": "index.js",
|
|
6
6
|
"types": "index.d.ts",
|