The Knowledge Graph

┌─────────────────────────────────────────────────────────┐
│                   KNOWLEDGE GRAPH                       │
├─────────────────────────────────────────────────────────┤
│  30,000+ Concepts                                       │
│  80,000+ Relations                                      │
│  10 Canonical Relation Types                            │
│  384-dimensional Embeddings                             │
└─────────────────────────────────────────────────────────┘

┌──────────────────────────────────────┐
│  Concept: "chair"                    │
├──────────────────────────────────────┤
│  ID: G-00142                         │
│  Type: PhysicalObject                │
│  Definition: A piece of furniture    │
│              designed for sitting    │
│  Embedding: [0.12, -0.34, ...]       │
└──────────────────────────────────────┘

      [chair]
         │
    ┌────┴────┬─────────┬──────────┐
    │ is_a    │ has_part│ made_of  │ affords_action
    ▼         ▼         ▼          ▼
[furniture] [legs]    [wood]     [sit]

{
  "source": "chair",
  "target": "furniture",
  "relation": "is_a",
  "strength": "primary",
  "confidence": 0.98
}

Claim: "Dogs are mammals"

Traversal:
  dog ──is_a──> mammal ✓

Result: Direct match, high confidence

Claim: "Poodles are animals"

Traversal:
  poodle ──is_a──> dog ──is_a──> mammal ──is_a──> animal ✓

Result: 3-hop path, good confidence (decreases with hops)

Claim: "Water is essential for life"

Path 1: water ──requires──> life
Path 2: life ──requires──> water
Path 3: water ──produces──> hydration ──requires──> life

Result: Multiple supporting paths, very high confidence

# Conceptual representation
embeddings = {
    "dog": [0.12, -0.34, 0.56, ..., 0.23],    # 384 dims
    "cat": [0.11, -0.31, 0.52, ..., 0.21],    # Similar to dog
    "car": [-0.45, 0.67, -0.12, ..., 0.89],   # Very different
}

def cosine_similarity(a, b):
    return np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b))
 
similarity("dog", "cat")    # 0.85 - related animals
similarity("dog", "puppy")  # 0.92 - very similar
similarity("dog", "car")    # 0.12 - unrelated

validation_rules = {
    "id_format": r"G-\d{5}",           # ID must be G-XXXXX
    "label_format": r"^[a-z_]+$",       # Lowercase with underscores
    "definition_length": (20, 500),     # 20-500 characters
    "min_relations": 1,                 # At least one relation
    "max_parts": 25,                    # Maximum 25 has_part relations
}

response = requests.get(
    "https://api.glyphnet.io/v1/concepts/dog",
    headers={"X-API-Key": API_KEY}
)
 
concept = response.json()
# {
#   "id": "G-00089",
#   "label": "dog",
#   "type": "PhysicalObject",
#   "definition": "A domesticated carnivorous mammal...",
#   "relations": [
#     {"type": "is_a", "target": "mammal", "confidence": 0.98},
#     {"type": "has_part", "target": "tail", "confidence": 0.95},
#     ...
#   ]
# }

response = requests.get(
    "https://api.glyphnet.io/v1/concepts/dog/related",
    headers={"X-API-Key": API_KEY},
    params={"relation": "is_a", "depth": 2}
)
 
related = response.json()
# ["mammal", "animal", "canine", "pet", ...]