CSV Tree Exporter

CSV exporter plugin for vcti-tree-exporter: writes a vcti tree as a directory tree of CSV tables.

CsvExporter serializes any vcti-tree tree whose node payloads are vcti-datanode DataNodes as a directory of CSV tables mirroring the hierarchy, and ships a get_exporter_descriptor() factory so a consumer registers it in a catalog and resolves it by id ("csv") or by attribute lookup.

CSV suits light 2-D tables of settings and metadata that load straight into pandas — not heavy result arrays. Use HDF5 or NPZ for complete numeric data.

Tree → CSV mapping

The output path is a directory:

out/
  _attributes.json          # the root node's attributes
  info.attributes.json      # a metadata-only leaf -> sidecar only
  materials/                # a group node -> sub-directory
    props.csv               # a structured-array leaf -> table with a header row
    props.attributes.json   # its attributes
  vec.csv                   # a plain 1-D array -> one value per row
  grid.csv                  # a plain 2-D array -> rows

Names are sanitized to filesystem-safe segments and de-collided case-insensitively (Mat, mat → Mat, mat~1) so nothing is silently overwritten. An array with more than two dimensions raises UnrepresentableNodeError. CSV does not preserve dtype and the export is one-way.

Usage

from pathlib import Path
from vcti.tree.exporter.core import build_registry, get_exporter
from vcti.tree.exporter.csv import get_exporter_descriptor

registry = build_registry([get_exporter_descriptor()])   # plus any other format plugins
get_exporter(registry, "csv").export(tree, Path("out_dir"), overwrite=True)

overwrite=True replaces the existing output directory — point it at a dedicated path.

Dependencies

• vcti-tree-exporter — the Exporter protocol and the ExporterDescriptor catalog.
• vcti-tree / vcti-datanode — the tree and payload types.
• numpy. (CSV writing itself uses only the standard library.)