From 4489cd5f184bd7ddae2898c01436cd664601fba2 Mon Sep 17 00:00:00 2001
From: GitHub Action Website Snapshot <>
Date: Wed, 30 Oct 2024 09:20:36 +0000
Subject: [PATCH] Refreshing website content from main repo.

  Source commit:
  https://github.com/OpenLineage/OpenLineage/commit/50ba05b3466ffa082e50ea3ff645e154e489195e
---
 .../spark/spark_column_lineage.md             | 43 -------------------
 1 file changed, 43 deletions(-)

diff --git a/docs/integrations/spark/spark_column_lineage.md b/docs/integrations/spark/spark_column_lineage.md
index 05b811f..5e53a76 100644
--- a/docs/integrations/spark/spark_column_lineage.md
+++ b/docs/integrations/spark/spark_column_lineage.md
@@ -92,46 +92,3 @@ To unravel two dependencies implement following logic:
 
 The inputs are also mapped for all dataset dependencies. The result is added to each output. 
 Finally, the list of outputs with all their inputs is mapped to `ColumnLineageDatasetFacetFields` object.
-
-## Writing custom extensions
-
-Spark framework is known for its great ability to be extended by custom libraries capable of reading or writing to anything. In case of having a custom implementation, we prepared an ability to extend column-level lineage implementation to be able to retrieve information from other input or output LogicalPlan nodes. 
-
-Creating such an extension requires implementing a following interface: 
-
-```
-/** Interface for implementing custom collectors of column-level lineage. */
-interface CustomColumnLineageVisitor {
-
-  /**
-   * Collect inputs for a given {@link LogicalPlan}. Column-level lineage mechanism traverses
-   * LogicalPlan on its node. This method will be called for each traversed node. Input information
-   * should be put into builder.
-   *
-   * @param node
-   * @param builder
-   */
-  void collectInputs(LogicalPlan node, ColumnLevelLineageBuilder builder);
-
-  /**
-   * Collect outputs for a given {@link LogicalPlan}. Column-level lineage mechanism traverses
-   * LogicalPlan on its node. This method will be called for each traversed node. Output information
-   * should be put into builder.
-   *
-   * @param node
-   * @param builder
-   */
-  void collectOutputs(LogicalPlan node, ColumnLevelLineageBuilder builder);
-
-  /**
-   * Collect expressions for a given {@link LogicalPlan}. Column-level lineage mechanism traverses
-   * LogicalPlan on its node. This method will be called for each traversed node. Expression
-   * dependency information should be put into builder.
-   *
-   * @param node
-   * @param builder
-   */
-  void collectExpressionDependencies(LogicalPlan node, ColumnLevelLineageBuilder builder);
-}
-```
-and making it available for Service Loader (implementation class name has to be put in a resource file `META-INF/services/io.openlineage.spark.agent.lifecycle.plan.column.CustomColumnLineageVisitor`).