Merge pull request #155 from linkml/dmbl_importer

add stub of a dbml importer

docs/packages/importers.rst

@@ -77,6 +77,13 @@ NCI implements a JSON serialization of ISO-11197. You can import this JSON and c
     schemauto import-cadsr "cdes/*.json"
 
 
+Importing from DBML
+--------------------
+
+DBML is a simple DSL for defining database schemas. It is a subset of SQL DDL.
+
+
+
 Packages for importing
 ----------------------  
 

pyproject.toml

@@ -53,6 +53,7 @@ click-default-group = "^1.2.4"
 linkml-runtime = "^1.7.2"
 duckdb = "^0.10.1"
 numpy = "<2.0"
+pydbml = "^1.1.2"
 
 [tool.poetry.dev-dependencies]
 pytest = ">=7.1.1"

schema_automator/importers/dbml_import_engine.py

@@ -0,0 +1,95 @@
+from schema_automator.importers.import_engine import ImportEngine
+from pydbml import PyDBML
+from linkml_runtime.linkml_model import SchemaDefinition, ClassDefinition, SlotDefinition
+from dataclasses import dataclass
+
+
+def _map_dbml_type_to_linkml(dbml_type: str) -> str:
+    """
+    Maps DBML data types to LinkML types.
+
+    :param dbml_type: The DBML column type.
+    :return: Corresponding LinkML type.
+    """
+    type_mapping = {
+        "int": "integer",
+        "varchar": "string",
+        "text": "string",
+        "float": "float",
+        "boolean": "boolean",
+        "date": "date",
+        "datetime": "datetime",
+    }
+    return type_mapping.get(dbml_type.lower(), "string")
+
+
+@dataclass
+class DbmlImportEngine(ImportEngine):
+    """
+    An ImportEngine that introspects a DBML schema to determine a corresponding LinkML schema.
+    """
+
+    def convert(
+            self,
+            file: str,
+            name: str = None,
+            model_uri: str = None,
+            identifier: str = None,
+            **kwargs
+    ) -> SchemaDefinition:
+        """
+        Converts a DBML schema file into a LinkML SchemaDefinition.
+
+        :param file: Path to the DBML schema file.
+        :param name: Optional name for the generated LinkML schema.
+        :param model_uri: Optional URI for the schema.
+        :param identifier: Identifier field for the schema.
+        :return: SchemaDefinition object representing the DBML schema.
+        """
+        # Initialize the schema definition
+        schema_name = name or "GeneratedSchema"
+        schema = SchemaDefinition(name=schema_name, id=model_uri or f"https://example.org/{schema_name}")
+
+        # Parse the DBML file
+        with open(file, 'r', encoding='utf-8') as f:
+            dbml_content = f.read()
+        parsed_dbml = PyDBML(dbml_content)
+
+        # Process tables
+        for table in parsed_dbml.tables:
+            class_def = ClassDefinition(
+                name=table.name,
+                description=table.note or f"Auto-generated class for table '{table.name}'",
+                slots=[],
+                unique_keys=[],  # Initialize unique keys property
+            )
+            processed_slots = set()  # Track processed slot names to avoid duplicates
+
+            # Handle primary key and unique constraints
+            primary_key_columns = [col for col in table.columns if col.pk]
+            unique_columns = [col for col in table.columns if col.unique and not col.pk]
+
+            # Process columns
+            for column in table.columns:
+
+                slot_name = column.name
+                slot_def = SlotDefinition(
+                    name=slot_name,
+                    range=_map_dbml_type_to_linkml(column.type),
+                    description=column.note or f"Column '{slot_name}'",
+                    required=column in primary_key_columns or column.unique,
+                    identifier=column in primary_key_columns,  # Mark primary key columns as identifiers
+                )
+                schema.slots[slot_name] = slot_def
+                class_def.slots.append(slot_name)
+                processed_slots.add(slot_name)
+
+            # Handle single unique column as primary key if no explicit primary key exists
+            if not primary_key_columns and len(unique_columns) == 1:
+                unique_column = unique_columns[0]
+                schema.slots[unique_column.name].identifier = True
+                schema.slots[unique_column.name].required = True
+
+            schema.classes[table.name] = class_def
+
+        return schema

tests/test_importers/test_dbml_importer.py

@@ -0,0 +1,72 @@
+import pytest
+from linkml_runtime.linkml_model import SchemaDefinition
+from schema_automator.importers.dbml_import_engine import DbmlImportEngine
+
+# Sample DBML content for testing
+DBML_SAMPLE = """
+Table Users {
+  id int [primary key, not null]
+  email varchar [unique, not null]
+  username varchar
+}
+
+Table Orders {
+  order_id int [not null]
+  user_id int [not null]
+  product_id int [not null]
+  quantity int
+}
+
+Table Countries {
+  code varchar [primary key, not null]
+  name varchar [not null]
+}
+"""
+
+@pytest.fixture
+def dbml_file(tmp_path):
+    """
+    Fixture to create a temporary DBML file.
+    """
+    dbml_path = tmp_path / "test.dbml"
+    dbml_path.write_text(DBML_SAMPLE)
+    print(dbml_path)
+    return dbml_path
+
+@pytest.fixture
+def importer():
+    """
+    Fixture to initialize the DbmlImportEngine.
+    """
+    return DbmlImportEngine()
+
+def test_dbml_to_linkml_conversion(dbml_file, importer):
+    """
+    Test the basic conversion of DBML to a LinkML schema.
+    """
+    schema = importer.convert(file=str(dbml_file), name="TestSchema")
+
+    # Assert the schema object is created
+    assert isinstance(schema, SchemaDefinition)
+
+    # Check that expected classes are present
+    assert "Users" in schema.classes
+    assert "Orders" in schema.classes
+
+    # Check that expected slots are present
+    assert "id" in schema.slots
+    assert schema.slots["id"].identifier
+    assert schema.slots["id"].required
+
+
+def test_primary_key_handling(dbml_file, importer):
+    """
+    Test correct handling of primary keys and required attributes.
+    """
+    schema = importer.convert(file=str(dbml_file), name="TestSchema")
+
+    # Check that primary keys are marked as required and identifiers
+    users_class = schema.classes["Users"]
+    assert "id" in users_class.slots
+    assert schema.slots["id"].identifier
+    assert schema.slots["id"].required