What you built
Victory lap. You built, from empty files: a property-graph record store, index-free adjacency (the O(1) hop), label and property indexes, four traversal algorithms (BFS, DFS, shortest path, variable-length with relationship isomorphism), and a complete tokenize → parse → plan → execute Cypher engine that answers the Kevin-Bacon co-star question correctly on the real graph.
Honest scope — what production graph databases add on top: durable on-disk record stores (the 15-byte/34-byte fixed records and doubly-linked relationship chains from Chapter 1's deep-dive), transactions / MVCC, a cost-based planner with real statistics, aggregation (count, collect), WITH pipelining, write clauses (CREATE / MERGE / SET), and clustering. Every one of those is an extension of a structure you now understand from the inside.
- openCypher — spec, grammar & TCK
- Cypher Manual — patterns reference
- Cypher Manual — pattern uniqueness / relationship isomorphism
- Neo4j — 15 Rules of a Native Graph Database (index-free adjacency)
- Neo4j Movies dataset (the real graph)
- "Cypher: An Evolving Query Language for Property Graphs" (SIGMOD 2018)
- "Graph Databases" (Robinson, Webber, Eifrem)