[DOP-22366] Basic column lineage handling in consumer

docs/changelog/next_release/155.feature.rst

@@ -0,0 +1,13 @@
+Implement basic column lineage handler in consumer.
+
+There are 2 DTO classes:
+
+    ``ColumnLineageDTO`` - describes tuple (operation, source_dataset, target_dataset).
+    ``DatasetColumnRelationDTO`` - describes source_column -> target_column (optional) relations within ``ColumnLineageDTO``
+
+Extractor builds list of these ``ColumnLineageDTO`` for each operation, and then perform bulk insert.
+Table rows are immutable, using ``ON CONFLICT DO NOTHING`` with few additional checks to reduce database IO.
+
+Open Lineage integration for Spark before v1.23 (or with columnLineage.datasetLineageEnabled=false, which is still default)
+produced INDIRECT lineage for each combination source_column x target_column,
+which is amlost the cartesian join. It is VERY expensive to handle, so we ignore this cases.