fluentd-openlineage-parser 0.1.0

data/events/event_full.json

@@ -0,0 +1,206 @@
+{
+  "eventType": "COMPLETE",
+  "eventTime": "2020-12-28T19:51:01.641Z",
+  "run": {
+    "runId": "ea041791-68bc-4ae1-bd89-4c8106a157e4",
+    "facets": {
+      "nominalTime": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "nominalStartTime": "2020-12-17T03:00:00.001Z",
+        "nominalEndTime": "2020-12-17T04:00:00.001Z"
+      },
+      "parent": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "run": {
+          "runId": "3f5e83fa-3480-44ff-99c5-ff943904e5e8"
+        },
+        "job": {
+          "namespace": "my-scheduler-namespace",
+          "name": "myjob.mytask"
+        }
+      },
+      "additionalProp1": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "additionalProp1": {}
+      },
+      "additionalProp2": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "additionalProp1": {}
+      },
+      "additionalProp3": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "additionalProp1": {}
+      }
+    }
+  },
+  "job": {
+    "namespace": "my-scheduler-namespace",
+    "name": "myjob.mytask",
+    "facets": {
+      "documentation": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "description": "string"
+      },
+      "sourceCodeLocation": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "type": "git",
+        "url": "http://example.com"
+      },
+      "sql": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "additionalPropExample": {
+          "example": true
+        },
+        "query": "SELECT * FROM foo"
+      },
+      "additionalProp1": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "additionalProp1": {}
+      },
+      "additionalProp2": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "additionalProp1": {}
+      },
+      "additionalProp3": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+        "additionalProp1": {}
+      }
+    }
+  },
+  "inputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.table",
+      "inputFacets": {
+        "dataQualityMetrics": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/DataQualityMetricsInputDatasetFacet",
+          "rowCount": 1000,
+          "bytes": 1048576,
+          "fileCount": 5
+        },
+        "dataQualityAssertions": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/DataQualityAssertionsDatasetFacet",
+          "assertions": [
+            {
+              "assertion": "row_count_equal_to",
+              "success": true
+            },
+            {
+              "assertion": "no_null_values",
+              "success": true,
+              "column": "id"
+            }
+          ]
+        }
+      },
+      "facets": {
+        "documentation": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "description": "canonical representation of entity Foo"
+        },
+        "schema": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "fields": [
+            {
+              "name": "column1",
+              "type": "VARCHAR",
+              "description": "string"
+            }
+          ]
+        },
+        "dataSource": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "name": "string",
+          "uri": "string"
+        },
+        "additionalProp1": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "additionalProp1": {}
+        },
+        "additionalProp2": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "additionalProp1": {}
+        },
+        "additionalProp3": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "additionalProp1": {}
+        }
+      }
+    }
+  ],
+  "outputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.table",
+      "outputFacets": {
+        "outputStatistics": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/OutputStatisticsOutputDatasetFacet",
+          "rowCount": 2000,
+          "size": 2097152,
+          "fileCount": 5
+        }
+      },
+      "facets": {
+        "documentation": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "description": "canonical representation of entity Foo"
+        },
+        "schema": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "fields": [
+            {
+              "name": "column1",
+              "type": "VARCHAR",
+              "description": "string"
+            }
+          ]
+        },
+        "dataSource": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "name": "string",
+          "uri": "string"
+        },
+        "additionalProp1": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "additionalProp1": {}
+        },
+        "additionalProp2": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "additionalProp1": {}
+        },
+        "additionalProp3": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#MyCustomJobFacet",
+          "additionalProp1": {}
+        }
+      }
+    }
+  ],
+  "producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+  "schemaURL": "https://openlineage.io/spec/1-0-5/OpenLineage.json#/definitions/RunEvent"
+}

data/events/event_invalid_dataset_facet.json

@@ -0,0 +1,31 @@
+{
+  "eventType": "COMPLETE",
+  "eventTime": "2020-12-28T19:52:00.001+10:00",
+  "run": {
+    "runId": "41fb5137-f0fd-4ee5-ba5c-56f8571d1bd7"
+  },
+  "job": {
+    "namespace": "my-scheduler-namespace",
+    "name": "myjob"
+  },
+  "inputs": [  ],
+  "outputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.output-1",
+      "facets": {
+        "ownership": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#whatever",
+          "owners": [
+            {
+              "no-name": "no-name-owner"
+            }
+          ]
+        }
+      }
+    }
+  ],
+  "producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+  "schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/RunEvent"
+}

data/events/event_invalid_input_dataset_facet.json

@@ -0,0 +1,29 @@
+{
+  "eventType": "COMPLETE",
+  "eventTime": "2020-12-28T19:52:00.001+10:00",
+  "run": {
+    "runId": "41fb5137-f0fd-4ee5-ba5c-56f8571d1bd7"
+  },
+  "job": {
+    "namespace": "my-scheduler-namespace",
+    "name": "myjob"
+  },
+  "inputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.input-1",
+      "inputFacets": {
+        "dataQualityMetrics": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/DataQualityMetricsInputDatasetFacet",
+          "noRowCount": 1000,
+          "bytes": 1048576,
+          "fileCount": 5
+        }
+      }
+    }
+  ],
+  "outputs": [],
+  "producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+  "schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/RunEvent"
+}

data/events/event_invalid_job_facet.json

@@ -0,0 +1,26 @@
+{
+  "eventType": "COMPLETE",
+  "eventTime": "2020-12-28T19:52:00.001+10:00",
+  "run": {
+    "runId": "41fb5137-f0fd-4ee5-ba5c-56f8571d1bd7"
+  },
+  "job": {
+    "namespace": "my-scheduler-namespace",
+    "name": "myjob",
+    "facets": {
+      "ownership": {
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/spec/OpenLineage.yml#whatever",
+        "owners": [
+          {
+            "no-name": "no-name-owner"
+          }
+        ]
+      }
+    }
+  },
+  "inputs": [ ],
+  "outputs": [ ],
+  "producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+  "schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/RunEvent"
+}

data/events/event_invalid_output_dataset_facet.json

@@ -0,0 +1,29 @@
+{
+  "eventType": "COMPLETE",
+  "eventTime": "2020-12-28T19:52:00.001+10:00",
+  "run": {
+    "runId": "41fb5137-f0fd-4ee5-ba5c-56f8571d1bd7"
+  },
+  "job": {
+    "namespace": "my-scheduler-namespace",
+    "name": "myjob"
+  },
+  "inputs": [],
+  "outputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.output-1",
+      "outputFacets": {
+        "outputStatistics": {
+          "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+          "_schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/OutputStatisticsOutputDatasetFacet",
+          "rowCount": "wrong",
+          "size": 2097152,
+          "fileCount": 5
+        }
+      }
+    }
+  ],
+  "producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+  "schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/RunEvent"
+}

data/events/event_invalid_run_facet.json

@@ -0,0 +1,28 @@
+{
+  "eventType": "COMPLETE",
+  "eventTime": "2020-12-28T19:52:00.001+10:00",
+  "run": {
+    "runId": "41fb5137-f0fd-4ee5-ba5c-56f8571d1bd7",
+    "facets": {
+      "parent": {
+        "run": {
+          "noRunId": "invalid run id"
+        },
+        "job": {
+          "namespace": "parent_namespace",
+          "name": "parent_name"
+        },
+        "_producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+        "_schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/RunFacet"
+      }
+    }
+  },
+  "job": {
+    "namespace": "my-scheduler-namespace",
+    "name": "myjob"
+  },
+  "inputs": [ ],
+  "outputs": [ ],
+  "producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+  "schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/RunEvent"
+}

data/events/event_no_run_id.json

@@ -0,0 +1,28 @@
+{
+  "eventType": "COMPLETE",
+  "eventTime": "2020-12-28T19:52:00.001+10:00",
+  "run": {
+  },
+  "job": {
+    "namespace": "my-scheduler-namespace",
+    "name": "myjob"
+  },
+  "inputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.input-1"
+    },
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.input-2"
+    }
+  ],
+  "outputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.output-1"
+    }
+  ],
+  "producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+  "schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/RunEvent"
+}

data/events/event_simple.json

@@ -0,0 +1,29 @@
+{
+  "eventType": "COMPLETE",
+  "eventTime": "2020-12-28T19:52:00.001+10:00",
+  "run": {
+    "runId": "41fb5137-f0fd-4ee5-ba5c-56f8571d1bd7"
+  },
+  "job": {
+    "namespace": "my-scheduler-namespace",
+    "name": "myjob"
+  },
+  "inputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.input-1"
+    },
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.input-2"
+    }
+  ],
+  "outputs": [
+    {
+      "namespace": "my-datasource-namespace",
+      "name": "instance.schema.output-1"
+    }
+  ],
+  "producer": "https://github.com/OpenLineage/OpenLineage/blob/v1-0-0/client",
+  "schemaURL": "https://openlineage.io/spec/1-0-1/OpenLineage.json#/definitions/RunEvent"
+}

data/fluentd-openlineage-parser.gemspec

@@ -0,0 +1,28 @@
+lib = File.expand_path("../lib", __FILE__)
+$LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
+
+Gem::Specification.new do |spec|
+  spec.name    = "fluentd-openlineage-parser"
+  spec.version = "0.1.0"
+  spec.authors = ["Pawel Leszczynski"]
+  spec.email   = ["leszczynski.pawel@gmail.com"]
+
+  spec.summary       = %q{Parser to validate Openlineage events.}
+  spec.description   = %q{Fluentd parser that validates if JSON is a valid Openlineage event.}
+  spec.homepage      = "http://openlineage.io"
+  spec.license       = "Apache-2.0"
+
+  test_files, files  = `git ls-files -z`.split("\x0").partition do |f|
+    f.match(%r{^(test|spec|features)/})
+  end
+  spec.files         = files
+  spec.executables   = files.grep(%r{^bin/}) { |f| File.basename(f) }
+  spec.test_files    = test_files
+  spec.require_paths = ["lib"]
+
+  spec.add_development_dependency "bundler"
+  spec.add_development_dependency "rake"
+  spec.add_development_dependency "test-unit"
+  spec.add_runtime_dependency "fluentd", [">= 0.14.10", "< 2"]
+  spec.add_dependency "rusty_json_schema"
+end

data/lib/fluent/plugin/fluentd-openlineage-parser.rb

@@ -0,0 +1,182 @@
+require "fluent/plugin/parser"
+require 'fluent/plugin/parser_json'
+require 'json'
+require "rusty_json_schema"
+
+module Fluent
+  module Plugin
+    class OpenlineageParser < Fluent::Plugin::JSONParser
+      Fluent::Plugin.register_parser("openlineage", self)
+
+      DEFAULT_SPEC_DIRECTORY="/etc/spec"
+
+      def configure(conf)
+        if conf.has_key?('spec_directory')
+          @spec_directory = conf['spec_directory']
+        else
+          @spec_directory = DEFAULT_SPEC_DIRECTORY
+        end
+        if (not @spec_directory.end_with?("/"))
+          @spec_directory += "/"
+        end
+        @validate_input_dataset_facets = conf.fetch('validate_input_dataset_facets', false)
+        @validate_output_dataset_facets = conf.fetch('validate_output_dataset_facets', false)
+        @validate_dataset_facets = conf.fetch('validate_dataset_facets', false)
+        @validate_run_facets = conf.fetch('validate_run_facets', true)
+        @validate_job_facets = conf.fetch('validate_job_facets', true)
+        @schema = load_schema()
+        @validator = RustyJSONSchema.build(@schema)
+        super
+      end
+
+      # https://docs.fluentd.org/plugin-development/api-plugin-parser
+      def parse(text)
+        # parse JSON with default JSONParser
+        super(text) { | time, json |
+          validate_openlineage(json)
+          yield time, json
+        }
+      end
+
+      private
+
+
+      def validate_openlineage(json)
+        if json == nil
+          raise ParserError, "Openlineage validation failed: invalid json provided"
+        end
+
+        # https://github.com/driv3r/rusty_json_schema
+        # Rust json parser ported to ruby that supports Draft 2020-12
+        errors = @validator.validate(json)
+
+        if errors.join(", ").include? "is not valid under any of the given schemas"
+          errors = enrich_oneOf_errors(json)
+        end
+        if !errors.empty?
+          raise ParserError, "Openlineage validation failed: " + errors.join(", ")
+        end
+      end
+
+      # Validator returns very generic OneOfNotValid error message
+      # We try to find better reason for mismatch with each candidate.
+      def enrich_oneOf_errors(json)
+        errors = []
+        @schema["oneOf"].each { |ref|
+          changed_schema = @schema
+          changed_schema.delete("oneOf")
+          changed_schema["$ref"] = ref["$ref"]
+          validator = RustyJSONSchema.build(changed_schema)
+          error = validator.validate(json)
+          if !error.empty?
+            errors.append("#{ref}: #{error.join(", ")}")
+          end
+        }
+        return errors
+      end
+
+      def load_schema()
+        schemaFile = @spec_directory + "OpenLineage.json"
+
+        if (not File.exist?(schemaFile))
+          raise ParserError, "Couldn't find Openlineage.json file within a defined spec directory: " + schemaFile
+        end
+
+        schema = File.read(schemaFile)
+        schema = rewrite_schema_to_include_facets(schema)
+        return schema
+      end
+
+      # Current Openlineage schema contains references to facets' definitions stored in files
+      # in facets directory which are not valid schemas for json_schema.
+      # In this step we rewrite Openlineage schema to contain facets definitions within it
+      def rewrite_schema_to_include_facets(schema)
+        # replace all the refs in schema to local refs
+        # "facets/ColumnLineageDatasetFacet.json" -> "#/defs/ColumnLineageDatasetFacet"
+        schema = schema.gsub(
+          /"facets\/([a-zA-Z]+)\.json"/,
+          '"#/$defs/\1"'
+        )
+        schema_json = JSON.parse(schema)
+        facets_path = @spec_directory + "facets/"
+
+        
+        # list all the facets
+        Dir.glob("#{facets_path}/*.json").each { |facet_file|
+            facet_schema =  JSON.parse(
+              File.read(facet_file).gsub(
+                /"https:\/\/openlineage\.io\/spec\/\d-\d-\d\/OpenLineage\.json#\/\$defs\/([a-zA-Z]+)"/,
+                '"#/$defs/\1"'
+              )
+            )
+
+            facet_schema["properties"].each { |property, ref|
+              facet_name =  ref["$ref"]&.gsub("#/$defs/", "")
+              parents = []
+              facet_schema["$defs"][facet_name]["allOf"]&.each { |definition|
+                unless definition["$ref"].nil?
+                  parents.append(definition["$ref"].gsub("#/$defs/", ""))
+                end
+              }
+              parents.each {|parent|
+              add_ref_as_parent_property(schema_json, parent, facet_name, property)
+              }
+            }
+            # include facets' definitions within schema
+            schema_json["$defs"] = schema_json["$defs"].merge(facet_schema["$defs"])
+        }
+        return schema_json
+      end
+
+      def add_ref_as_parent_property(schema, parent, facet_name, property)
+        getter = find_parent_object_getter(parent)
+        if getter.nil?
+          return
+        end
+        properties = getter.call(schema)["properties"] || {}
+        properties[property] = {"$ref" => "#/$defs/" + facet_name}
+        getter.call(schema)["properties"] = properties
+      end
+
+      # Based on facet name find path to object facets
+      def find_parent_object_getter(parent)
+        getter = nil
+        case parent
+        when "JobFacet"
+          if @validate_job_facets
+            getter = ->(schema) { schema["$defs"]["Job"]["properties"]["facets"] }
+          end
+        when "RunFacet"
+          if @validate_run_facets
+            getter = ->(schema) { schema["$defs"]["Run"]["properties"]["facets"] }
+          end
+        when "DatasetFacet"
+          if @validate_dataset_facets
+            getter = ->(schema) { schema["$defs"]["Dataset"]["properties"]["facets"] }
+          end
+        when "OutputDatasetFacet"
+          if @validate_output_dataset_facets
+            getter = ->(schema) { schema \
+            ["$defs"] \
+            ["OutputDataset"] \
+            ["allOf"].select {|el| el.key?("type") } \
+            [0] \
+            ["properties"] \
+            ["outputFacets"] }
+          end
+        when "InputDatasetFacet"
+          if @validate_input_dataset_facets
+            getter = ->(schema) { schema \
+            ["$defs"] \
+            ["InputDataset"] \
+            ["allOf"].select {|el| el.key?("type") } \
+            [0] \
+            ["properties"] \
+            ["inputFacets"] }
+          end
+        end
+        return getter
+      end
+    end
+  end
+end