How graphgeeks-lab/graphfaker works
GraphFaker generates synthetic graph datasets from reusable schemas or built-in domains, and can also turn selected real-world sources into graphs. The main workflow is to choose a domain or schema, generate entities, relationships and events, retain planted-pattern truth and run metadata, then export tables or load them into a graph store. Fraud, coordination and supply-chain domains add labelled patterns and hardness evaluation; social generation and schema drafting are additional paths. The graph follows documented behavior and indexed references; unsampled implementations are represented only at subsystem level where the tree supports them.
Main entry points: cli.py CLI · tables.py Graph Tables · run.py Schema Engine · graph.py Graph Schema
Components 27
User Entry
- CLI
graphfaker/cli.py - Python API
graphfaker/core.py
Schemas And Domains
- Graph Schema
graphfaker/schema/graph.py - Attribute Samplers
graphfaker/schema/samplers.py - Pattern Schema
graphfaker/schema/patterns.py - Social Domain
graphfaker/domains/social.py - Fraud Domain
graphfaker/domains/fraud - Coordination Domain
graphfaker/domains/coordination - Supply Chain Domain
graphfaker/domains/supply_chain - Schema Drafter
graphfaker/llm/draft.py
Generation Runtime
- Schema Engine
graphfaker/engine/run.py - Attribute Sampling
graphfaker/engine/sampling.py - Topology Builder
graphfaker/engine/topology.py - Process Generator
graphfaker/engine/process.py - Pattern Injection
graphfaker/engine/injection.py - Fraud Process
graphfaker/domains/fraud/process.py
Data And Integrations
- Graph Tables
graphfaker/backends/tables.py - File Export
graphfaker/export.py - Neo4j Sink
graphfaker/sinks/neo4j.py - Live Neo4j Sink
graphfaker/sinks/neo4j_live.py - Other Sinks
graphfaker/sinks - Real-World Fetchers
graphfaker/fetchers
Truth And Analysis
- Ground Truth
graphfaker/sinks/truth.py - Load Verification
graphfaker/sinks/verify.py - Entity Resolution
graphfaker/resolve.py
Other components
- User
- External Data
Connections 30
- User → CLI: runs commands
README.md - User → Python API: calls API
README.md - CLI → Fraud Domain: selects domain
graphfaker/cli.py - CLI → Coordination Domain: selects domain
- CLI → Supply Chain Domain: selects domain
- CLI → Social Domain: selects domain
- CLI → Schema Engine: runs schema
graphfaker/cli.py - Python API → Schema Engine: generates graph
- Social Domain → Graph Schema: declares graph
graphfaker/domains/social.py - Graph Schema → Attribute Samplers: defines attributes
graphfaker/schema/graph.py - Graph Schema → Pattern Schema: declares patterns
- Schema Drafter → Graph Schema: creates schema
graphfaker/llm/draft.py - Schema Engine → Attribute Sampling: samples attributes
graphfaker/engine/run.py - Schema Engine → Topology Builder: builds edges
graphfaker/engine/run.py - Schema Engine → Graph Tables: produces tables
graphfaker/engine/run.py - Process Generator → Pattern Schema: uses declarations
graphfaker/engine/process.py - Fraud Domain → Fraud Process: generates activity
graphfaker/domains/fraud/process.py - Fraud Domain → Pattern Injection: plants typologies
graphfaker/domains/fraud/typologies.py - Coordination Domain → Pattern Injection: plants playbooks
graphfaker/domains/coordination/playbooks.py - Supply Chain Domain → Pattern Injection: plants patterns
graphfaker/domains/supply_chain/patterns.py - CLI → File Export: writes dataset
graphfaker/cli.py - Graph Tables → Neo4j Sink: loads tables
graphfaker/sinks/neo4j.py - Graph Tables → Live Neo4j Sink: streams batches
graphfaker/sinks/neo4j_live.py - Graph Tables → Other Sinks: exports tables
graphfaker/sinks/duckdb.py - External Data → Real-World Fetchers: provides records
README.md - Real-World Fetchers → CLI: feeds graph
graphfaker/cli.py - Schema Engine → Ground Truth: records planted truth
- Graph Tables → Load Verification: checks loaded graph
graphfaker/sinks/verify.py - Ground Truth → Load Verification: checks ground truth
graphfaker/sinks/verify.py - Graph Tables → Entity Resolution: resolves entities