Metadata Validation

CSVW-EO provides two validation layers:

1. Internal schema validation
2. SHACL validation

---

Internal Validation

The validate_metadata.py utility validates metadata against the internal Pydantic schema.

It checks:

• Required fields
• Datatypes
• Structural consistency
• Metadata object validity

Usage

python validate_metadata.py metadata.json

SHACL Validation

The validate_metadata_shacl.py utility validates metadata against RDF SHACL constraints.

This is the strictest validation layer.

It checks:

• RDF consistency
• SHACL constraints
• Cross-field consistency
• Structural restrictions

TODO: As the dataset is not published yet, it is not functional for now (or put the .ttl within the library..).

Usage

python validate_metadata_shacl.py \
  metadata.json \
  csvw-eo-constraints.ttl

Validation Recommendations

Validator	Purpose
validate_metadata.py	Fast schema validation
validate_metadata_shacl.py	Formal RDF validation

Both validators should be used before publishing metadata.

Structure Validation

The assert_same_structure.py utility validates that a generated dummy dataset matches the structure of the original dataset.

It checks:

• Column names
• Column ordering
• Datatypes
• Nullable constraints
• Optional categorical domains

Usage

python assert_same_structure.py \
  original.csv \
  dummy.csv

Important Notes

Validation ensures structural correctness only.

Validation does not guarantee:

• Metadata safety
• Privacy preservation
• Statistical validity