datasynth-graph 0.2.3

Graph/network export for synthetic accounting data - supports PyTorch Geometric, Neo4j, and DGL formats
Documentation

datasynth-graph

Graph/network export for synthetic accounting data with ML-ready formats.

Overview

datasynth-graph provides graph construction and export capabilities:

  • Graph Builders: Transaction, approval, and entity relationship graphs
  • ML Export: PyTorch Geometric, Neo4j, and DGL formats
  • Feature Engineering: Temporal, amount, structural, and categorical features
  • Data Splits: Train/validation/test split generation

Graph Types

Graph Nodes Edges Use Case
Transaction Network Accounts, Entities Transactions Anomaly detection
Approval Network Users Approvals SoD analysis
Entity Relationship Legal Entities Ownership Consolidation analysis

Export Formats

PyTorch Geometric

graphs/transaction_network/pytorch_geometric/
├── node_features.pt    # [num_nodes, num_features]
├── edge_index.pt       # [2, num_edges]
├── edge_attr.pt        # [num_edges, num_edge_features]
├── labels.pt           # [num_nodes] or [num_edges]
├── train_mask.pt       # Boolean mask
├── val_mask.pt
└── test_mask.pt

Neo4j

graphs/entity_relationship/neo4j/
├── nodes_account.csv
├── nodes_entity.csv
├── edges_transaction.csv
└── import.cypher

Features

Category Features
Temporal weekday, period, is_month_end, is_quarter_end, is_year_end
Amount log(amount), benford_probability, is_round_number
Structural line_count, unique_accounts, has_intercompany
Categorical business_process (one-hot), source_type (one-hot)

Usage

use datasynth_graph::{TransactionGraphBuilder, PyTorchGeometricExporter};

let builder = TransactionGraphBuilder::new();
let graph = builder.build(&entries)?;

let exporter = PyTorchGeometricExporter::new("output/graphs");
exporter.export(&graph, split_config)?;

License

Apache-2.0 - See LICENSE for details.