Graph Search

Endpoint

GET /api/v1/knowledge/graph/search?q={term}&limit={n}

Usage

Search uses Neo4j fulltext indexing. Handles multi-word queries, partial matches, and fuzzy matching.

# Single term
curl -H "X-API-Key: $KEY" \
  "https://api.marc27.com/api/v1/knowledge/graph/search?q=titanium&limit=5"

# Multi-word (finds entities matching ANY word)
curl -H "X-API-Key: $KEY" \
  "https://api.marc27.com/api/v1/knowledge/graph/search?q=rocket+nozzle+erosion&limit=5"

Response

[
  {
    "name": "Nozzle",
    "entity_type": "APL",
    "label": "Application",
    "tenant": "public"
  },
  {
    "name": "erosion",
    "entity_type": "PRO",
    "label": "Property",
    "tenant": "public"
  }
]

Tips

• Use short, specific terms — “Inconel”, “fatigue”, “nozzle”
• For natural language questions, use Semantic Search instead
• Use entity_type filter in the query to narrow results

Entity Neighbors

Get connected entities (1-hop traversal):

GET /api/v1/knowledge/graph/entity/{name}

curl -H "X-API-Key: $KEY" \
  "https://api.marc27.com/api/v1/knowledge/graph/entity/Inconel%20718"