Diffing¶

Compare two model versions to find additions, removals, and field-level changes.

diff_models()¶

diff_models() compares a baseline model against a revised model:

from etcion import diff_models

diff = diff_models(model_v1, model_v2)
print(diff.summary())
# "ModelDiff: 2 added, 1 removed, 3 modified"

By default, concepts are matched by id. Use match_by="type_name" to match by (type, name) tuple instead:

diff = diff_models(model_v1, model_v2, match_by="type_name")

ModelDiff¶

The returned ModelDiff is a frozen dataclass with three fields:

diff.added -- tuple of concepts present only in model_b
diff.removed -- tuple of concepts present only in model_a
diff.modified -- tuple of ConceptChange for concepts with field differences

for concept in diff.added:
    print(f"+ {type(concept).__name__}: {concept.name}")

for concept in diff.removed:
    print(f"- {type(concept).__name__}: {concept.name}")

ModelDiff is falsy when models are identical:

if not diff:
    print("Models are identical")

ConceptChange and FieldChange¶

Each modified concept produces a ConceptChange containing a dict of FieldChange objects:

for change in diff.modified:
    print(f"~ {change.concept_type} ({change.concept_id}):")
    for field_name, fc in change.changes.items():
        print(f"    {field_name}: {fc.old!r} -> {fc.new!r}")

Serializing Diffs¶

Convert a diff to a JSON-compatible dict:

import json
print(json.dumps(diff.to_dict(), indent=2))

Or get a one-line summary:

print(diff.summary())