Neo4j export

The neo4j command writes the graph into a Neo4j database so you can explore it with Cypher and the Neo4j Browser. It uses plain UNWIND … MERGE Cypher and does not require APOC.

Install and run

pip install "graphlens-cli[neo4j]"

graphlens neo4j ./my-project \
  --uri bolt://localhost:7687 \
  --user neo4j \
  --password secret

A quick local Neo4j with Docker:

docker run --rm -p 7474:7474 -p 7687:7687 \
  -e NEO4J_AUTH=neo4j/secret neo4j:5

Options

Flag	Default	Description
--lang	auto	Adapter(s) to use
--uri	bolt://localhost:7687	Neo4j Bolt URI
--user	neo4j	Username
--password	password	Password
--wipe / --no-wipe	--no-wipe	Delete existing :Code nodes before import
--batch-size	500	Items per Cypher batch

Use --wipe for a clean reload; omit it to merge into whatever is already there (nodes are MERGEd by id, so re-imports are idempotent).

Schema

• Every node gets a generic :Code label plus a kind-specific label: :Function, :Method, :Class, :Module, :ExternalSymbol, :Boundary, and so on.
• A uniqueness constraint is created on Code.id.
• Relations are created grouped by kind, so a graphlens CALLS relation becomes a [:CALLS] edge, REFERENCES becomes [:REFERENCES], etc.
• Scalar node metadata (including span fields) is stored as properties prefixed with meta_.

Example queries

// The 20 most-called functions
MATCH (f:Function)<-[:CALLS]-(caller)
RETURN f.qualified_name AS fn, count(caller) AS callers
ORDER BY callers DESC
LIMIT 20;

// Everything a class reaches within 2 hops
MATCH path = (c:Class {name: 'OrderService'})-[*1..2]-(n)
RETURN path;

// Cross-language communication (after running graphlens-link)
MATCH (consumer)-[r:COMMUNICATES_WITH]->(provider)
RETURN consumer.qualified_name, provider.qualified_name, r.mechanism;

// Third-party surface area
MATCH (e:ExternalSymbol {meta_origin: 'third_party'})<-[:RESOLVES_TO|CALLS|HAS_TYPE]-(n)
RETURN e.name, count(n) AS uses
ORDER BY uses DESC;

Programmatic export

Prefer to drive it from Python? Analyze, then store the graph through a backend that implements the GraphBackend contract (store / clear). The repository's examples/neo4j_export.py shows a complete export-and-query script you can adapt.