Export Module Specification

Required: json , xml.etree.ElementTree , pickle , pathlib
Optional: networkx (for GraphML/GEXF graph exports)

Overview

Multi-format export generation from parsed GNN models. Pipeline Step 7.

The export module uses a layered design:

processor.py — Orchestrates multi-file and multi-format export workflows (generate_exports, export_single_gnn_file, export_model, export_gnn_model). Includes a built-in GNN content parser (parse_gnn_content).
formatters.py — Format-specific serializers that write parsed GNN data to disk. Each function takes a data dict and an output Path, returning bool for success.
format_exporters.py — Advanced GNN-aware exporters with full section parsing (_gnn_model_to_dict), NetworkX graph construction, and logged output. Functions return Tuple[bool, str].
core.py — Pipeline integration adapter invoked by 7_export.py.
utils.py — Module introspection (get_module_info, get_supported_formats).
mcp.py — Model Context Protocol tool registrations.

Format	Extension	Module	Notes
JSON	`.json`	formatters / format_exporters	Human-readable, portable
XML	`.xml`	formatters / format_exporters	Hierarchical with minidom pretty-print
GraphML	`.graphml`	formatters / format_exporters	Requires NetworkX (optional)
GEXF	`.gexf`	formatters / format_exporters	Gephi-compatible, requires NetworkX
Pickle	`.pkl`	formatters / format_exporters	Python binary serialization
Plaintext Summary	`.txt`	formatters / format_exporters	Human-readable model summary
Plaintext DSL	`.dsl`	formatters / format_exporters	Round-trip GNN-like text

from export import generate_exports, export_single_gnn_file, export_model
from export import get_supported_formats, get_module_info

uv run uv run --extra dev python -m pytest src/tests/export/test_export_overall.py src/tests/export/test_export_roundtrip.py -v