// Copyright 2022 Google LLC // // Licensed under the Apache License, Version 2.0 (the "License"); // you may not use this file except in compliance with the License. // You may obtain a copy of the License at // // http://www.apache.org/licenses/LICENSE-2.0 // // Unless required by applicable law or agreed to in writing, software // distributed under the License is distributed on an "AS IS" BASIS, // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. // See the License for the specific language governing permissions and // limitations under the License. syntax = "proto3"; package google.cloud.documentai.v1beta3; import "google/api/resource.proto"; import "google/protobuf/timestamp.proto"; option csharp_namespace = "Google.Cloud.DocumentAI.V1Beta3"; option go_package = "cloud.google.com/go/documentai/apiv1beta3/documentaipb;documentaipb"; option java_multiple_files = true; option java_outer_classname = "DocumentAiEvaluation"; option java_package = "com.google.cloud.documentai.v1beta3"; option php_namespace = "Google\\Cloud\\DocumentAI\\V1beta3"; option ruby_package = "Google::Cloud::DocumentAI::V1beta3"; // Gives a short summary of an evaluation, and links to the evaluation itself. message EvaluationReference { // The resource name of the Long Running Operation for the evaluation. string operation = 1; // The resource name of the evaluation. string evaluation = 2 [(google.api.resource_reference) = { type: "documentai.googleapis.com/Evaluation" }]; // An aggregate of the statistics for the evaluation with fuzzy matching on. Evaluation.Metrics aggregate_metrics = 4; // An aggregate of the statistics for the evaluation with fuzzy matching off. Evaluation.Metrics aggregate_metrics_exact = 5; } // An evaluation of a ProcessorVersion's performance. message Evaluation { option (google.api.resource) = { type: "documentai.googleapis.com/Evaluation" pattern: "projects/{project}/locations/{location}/processors/{processor}/processorVersions/{processor_version}/evaluations/{evaluation}" }; // Evaluation counters for the documents that were used. message Counters { // How many documents were sent for evaluation. int32 input_documents_count = 1; // How many documents were not included in the evaluation as they didn't // pass validation. int32 invalid_documents_count = 2; // How many documents were not included in the evaluation as Document AI // failed to process them. int32 failed_documents_count = 3; // How many documents were used in the evaluation. int32 evaluated_documents_count = 4; } // Evaluation metrics, either in aggregate or about a specific entity. message Metrics { // The calculated precision. float precision = 1; // The calculated recall. float recall = 2; // The calculated f1 score. float f1_score = 3; // The amount of occurrences in predicted documents. int32 predicted_occurrences_count = 4; // The amount of occurrences in ground truth documents. int32 ground_truth_occurrences_count = 5; // The amount of documents with a predicted occurrence. int32 predicted_document_count = 10; // The amount of documents with a ground truth occurrence. int32 ground_truth_document_count = 11; // The amount of true positives. int32 true_positives_count = 6; // The amount of false positives. int32 false_positives_count = 7; // The amount of false negatives. int32 false_negatives_count = 8; // The amount of documents that had an occurrence of this label. int32 total_documents_count = 9; } // Evaluations metrics, at a specific confidence level. message ConfidenceLevelMetrics { // The confidence level. float confidence_level = 1; // The metrics at the specific confidence level. Metrics metrics = 2; } // Metrics across multiple confidence levels. message MultiConfidenceMetrics { // A type that determines how metrics should be interpreted. enum MetricsType { // The metrics type is unspecified. By default, metrics without a // particular specification are for leaf entity types (i.e., top-level // entity types without child types, or child types which are not // parent types themselves). METRICS_TYPE_UNSPECIFIED = 0; // Indicates whether metrics for this particular label type represent an // aggregate of metrics for other types instead of being based on actual // TP/FP/FN values for the label type. Metrics for parent (i.e., non-leaf) // entity types are an aggregate of metrics for their children. AGGREGATE = 1; } // Metrics across confidence levels with fuzzy matching enabled. repeated ConfidenceLevelMetrics confidence_level_metrics = 1; // Metrics across confidence levels with only exact matching. repeated ConfidenceLevelMetrics confidence_level_metrics_exact = 4; // The calculated area under the precision recall curve (AUPRC), computed by // integrating over all confidence thresholds. float auprc = 2; // The Estimated Calibration Error (ECE) of the confidence of the predicted // entities. float estimated_calibration_error = 3; // The AUPRC for metrics with fuzzy matching disabled, i.e., exact matching // only. float auprc_exact = 5; // The ECE for the predicted entities with fuzzy matching disabled, i.e., // exact matching only. float estimated_calibration_error_exact = 6; // The metrics type for the label. MetricsType metrics_type = 7; } // The resource name of the evaluation. // Format: // `projects/{project}/locations/{location}/processors/{processor}/processorVersions/{processor_version}/evaluations/{evaluation}` string name = 1; // The time that the evaluation was created. google.protobuf.Timestamp create_time = 2; // Counters for the documents used in the evaluation. Counters document_counters = 5; // Metrics for all the entities in aggregate. MultiConfidenceMetrics all_entities_metrics = 3; // Metrics across confidence levels, for different entities. map entity_metrics = 4; // The KMS key name used for encryption. string kms_key_name = 6; // The KMS key version with which data is encrypted. string kms_key_version_name = 7; }