Crate datasynth_eval 

Synthetic Data Evaluation Framework

This crate provides comprehensive evaluation capabilities for validating the quality and correctness of generated synthetic financial data.

Features

• Statistical Quality: Benford’s Law, amount distributions, line item patterns
• Semantic Coherence: Balance sheet validation, subledger reconciliation
• Data Quality: Uniqueness, completeness, format consistency
• ML-Readiness: Feature distributions, label quality, graph structure
• Reporting: HTML and JSON reports with pass/fail thresholds

Example

use datasynth_eval::{Evaluator, EvaluationConfig};

let config = EvaluationConfig::default();
let evaluator = Evaluator::new(config);

// Evaluate generated data
let result = evaluator.evaluate(&generation_result)?;

// Generate report
result.generate_html_report("evaluation_report.html")?;

Re-exports

pub use config::EvaluationConfig;

pub use config::EvaluationThresholds;

pub use config::PrivacyEvaluationConfig;

pub use error::EvalError;

pub use error::EvalResult;

pub use statistical::AmountDistributionAnalysis;

pub use statistical::AmountDistributionAnalyzer;

pub use statistical::AnomalyRealismEvaluation;

pub use statistical::AnomalyRealismEvaluator;

pub use statistical::BenfordAnalysis;

pub use statistical::BenfordAnalyzer;

pub use statistical::BenfordConformity;

pub use statistical::DetectionDifficulty;

pub use statistical::DriftDetectionAnalysis;

pub use statistical::DriftDetectionAnalyzer;

pub use statistical::DriftDetectionEntry;

pub use statistical::DriftDetectionMetrics;

pub use statistical::DriftEventCategory;

pub use statistical::LabeledDriftEvent;

pub use statistical::LabeledEventAnalysis;

pub use statistical::LineItemAnalysis;

pub use statistical::LineItemAnalyzer;

pub use statistical::LineItemEntry;

pub use statistical::StatisticalEvaluation;

pub use statistical::TemporalAnalysis;

pub use statistical::TemporalAnalyzer;

pub use statistical::TemporalEntry;

pub use coherence::AccountType;

pub use coherence::ApprovalLevelData;

pub use coherence::AuditEvaluation;

pub use coherence::AuditEvaluator;

pub use coherence::AuditFindingData;

pub use coherence::AuditRiskData;

pub use coherence::AuditTrailEvaluation;

pub use coherence::AuditTrailGap;

pub use coherence::BalanceSheetEvaluation;

pub use coherence::BalanceSheetEvaluator;

pub use coherence::BalanceSnapshot;

pub use coherence::BankReconciliationEvaluation;

pub use coherence::BankReconciliationEvaluator;

pub use coherence::BidEvaluationData;

pub use coherence::BudgetVarianceData;

pub use coherence::CashPositionData;

pub use coherence::CoherenceEvaluation;

pub use coherence::ConcentrationMetrics;

pub use coherence::CountryPackData;

pub use coherence::CountryPackEvaluation;

pub use coherence::CountryPackEvaluator;

pub use coherence::CountryPackThresholds;

pub use coherence::CovenantData;

pub use coherence::CrossProcessEvaluation;

pub use coherence::CrossProcessEvaluator;

pub use coherence::CycleCountData;

pub use coherence::DocumentChainEvaluation;

pub use coherence::DocumentChainEvaluator;

pub use coherence::DocumentReferenceData;

pub use coherence::EarnedValueData;

pub use coherence::EntityReferenceData;

pub use coherence::EsgEvaluation;

pub use coherence::EsgEvaluator;

pub use coherence::EsgThresholds;

pub use coherence::ExpenseReportData;

pub use coherence::FairValueEvaluation;

pub use coherence::FinancialReportingEvaluation;

pub use coherence::FinancialReportingEvaluator;

pub use coherence::FinancialStatementData;

pub use coherence::FrameworkViolation;

pub use coherence::GovernanceData;

pub use coherence::HedgeEffectivenessData;

pub use coherence::HolidayData;

pub use coherence::HrPayrollEvaluation;

pub use coherence::HrPayrollEvaluator;

pub use coherence::ICMatchingData;

pub use coherence::ICMatchingEvaluation;

pub use coherence::ICMatchingEvaluator;

pub use coherence::ImpairmentEvaluation;

pub use coherence::IsaComplianceEvaluation;

pub use coherence::KpiData;

pub use coherence::LeaseAccountingEvaluation;

pub use coherence::LeaseAccountingEvaluator;

pub use coherence::LeaseEvaluation;

pub use coherence::ManufacturingEvaluation;

pub use coherence::ManufacturingEvaluator;

pub use coherence::MaterialityData;

pub use coherence::NettingData;

pub use coherence::NetworkEdge;

pub use coherence::NetworkEvaluation;

pub use coherence::NetworkEvaluator;

pub use coherence::NetworkNode;

pub use coherence::NetworkThresholds;

pub use coherence::O2CChainData;

pub use coherence::P2PChainData;

pub use coherence::PayrollHoursData;

pub use coherence::PayrollLineItemData;

pub use coherence::PayrollRunData;

pub use coherence::PcaobComplianceEvaluation;

pub use coherence::PerformanceObligation;

pub use coherence::ProductionOrderData;

pub use coherence::ProjectAccountingEvaluation;

pub use coherence::ProjectAccountingEvaluator;

pub use coherence::ProjectAccountingThresholds;

pub use coherence::ProjectRevenueData;

pub use coherence::QualityInspectionData;

pub use coherence::QuoteLineData;

pub use coherence::ReconciliationData;

pub use coherence::ReferentialData;

pub use coherence::ReferentialIntegrityEvaluation;

pub use coherence::ReferentialIntegrityEvaluator;

pub use coherence::RetainageData;

pub use coherence::RevenueContract;

pub use coherence::RevenueRecognitionEvaluation;

pub use coherence::RevenueRecognitionEvaluator;

pub use coherence::RoutingOperationData;

pub use coherence::SafetyMetricData;

pub use coherence::SalesQuoteData;

pub use coherence::SalesQuoteEvaluation;

pub use coherence::SalesQuoteEvaluator;

pub use coherence::SalesQuoteThresholds;

pub use coherence::ScorecardCoverageData;

pub use coherence::SourcingEvaluation;

pub use coherence::SourcingEvaluator;

pub use coherence::SourcingProjectData;

pub use coherence::SoxComplianceEvaluation;

pub use coherence::SpendAnalysisData;

pub use coherence::StandardsComplianceEvaluation;

pub use coherence::StandardsThresholds;

pub use coherence::StrengthStats;

pub use coherence::SubledgerEvaluator;

pub use coherence::SubledgerReconciliationEvaluation;

pub use coherence::SupplierEsgData;

pub use coherence::TaxEvaluation;

pub use coherence::TaxEvaluator;

pub use coherence::TaxLineData;

pub use coherence::TaxRateData;

pub use coherence::TaxReturnData;

pub use coherence::TaxThresholds;

pub use coherence::TimeEntryData;

pub use coherence::TreasuryEvaluation;

pub use coherence::TreasuryEvaluator;

pub use coherence::TreasuryThresholds;

pub use coherence::VariableConsideration;

pub use coherence::ViolationSeverity;

pub use coherence::WaterUsageData;

pub use coherence::WithholdingData;

pub use coherence::WorkpaperData;

pub use quality::CompletenessAnalysis;

pub use quality::CompletenessAnalyzer;

pub use quality::ConsistencyAnalysis;

pub use quality::ConsistencyAnalyzer;

pub use quality::ConsistencyRule;

pub use quality::DuplicateInfo;

pub use quality::FieldCompleteness;

pub use quality::FieldDefinition;

pub use quality::FieldValue;

pub use quality::FormatAnalysis;

pub use quality::FormatAnalyzer;

pub use quality::FormatVariation;

pub use quality::QualityEvaluation;

pub use quality::UniqueRecord;

pub use quality::UniquenessAnalysis;

pub use quality::UniquenessAnalyzer;

pub use ml::AnomalyScoringAnalysis;

pub use ml::AnomalyScoringAnalyzer;

pub use ml::CrossModalAnalysis;

pub use ml::CrossModalAnalyzer;

pub use ml::DomainGapAnalysis;

pub use ml::DomainGapAnalyzer;

pub use ml::EmbeddingReadinessAnalysis;

pub use ml::EmbeddingReadinessAnalyzer;

pub use ml::FeatureAnalysis;

pub use ml::FeatureAnalyzer;

pub use ml::FeatureQualityAnalysis;

pub use ml::FeatureQualityAnalyzer;

pub use ml::FeatureStats;

pub use ml::GnnReadinessAnalysis;

pub use ml::GnnReadinessAnalyzer;

pub use ml::GraphAnalysis;

pub use ml::GraphAnalyzer;

pub use ml::GraphMetrics;

pub use ml::LabelAnalysis;

pub use ml::LabelAnalyzer;

pub use ml::LabelDistribution;

pub use ml::MLReadinessEvaluation;

pub use ml::SchemeDetectabilityAnalysis;

pub use ml::SchemeDetectabilityAnalyzer;

pub use ml::SplitAnalysis;

pub use ml::SplitAnalyzer;

pub use ml::SplitMetrics;

pub use ml::TemporalFidelityAnalysis;

pub use ml::TemporalFidelityAnalyzer;

pub use report::BaselineComparison;

pub use report::ComparisonResult;

pub use report::EvaluationReport;

pub use report::HtmlReportGenerator;

pub use report::JsonReportGenerator;

pub use report::MetricChange;

pub use report::ReportMetadata;

pub use report::ThresholdChecker;

pub use report::ThresholdResult;

pub use tuning::ConfigSuggestion;

pub use tuning::ConfigSuggestionGenerator;

pub use tuning::TuningAnalyzer;

pub use tuning::TuningCategory;

pub use tuning::TuningOpportunity;

pub use enhancement::AutoTuneResult;

pub use enhancement::AutoTuner;

pub use enhancement::ConfigPatch;

pub use enhancement::EnhancementReport;

pub use enhancement::Recommendation;

pub use enhancement::RecommendationCategory;

pub use enhancement::RecommendationEngine;

pub use enhancement::RecommendationPriority;

pub use enhancement::RootCause;

pub use enhancement::SuggestedAction;

pub use privacy::LinkageAttack;

pub use privacy::LinkageConfig;

pub use privacy::LinkageResults;

pub use privacy::MembershipInferenceAttack;

pub use privacy::MiaConfig;

pub use privacy::MiaResults;

pub use privacy::NistAlignmentReport;

pub use privacy::NistCriterion;

pub use privacy::PrivacyEvaluation;

pub use privacy::SynQPMatrix;

pub use privacy::SynQPQuadrant;

pub use benchmarks::acfe_calibrated_1k;

pub use benchmarks::acfe_collusion_5k;

pub use benchmarks::acfe_management_override_2k;

pub use benchmarks::all_acfe_benchmarks;

pub use benchmarks::all_benchmarks;

pub use benchmarks::all_industry_benchmarks;

pub use benchmarks::anomaly_bench_1k;

pub use benchmarks::data_quality_100k;

pub use benchmarks::entity_match_5k;

pub use benchmarks::financial_services_fraud_5k;

pub use benchmarks::fraud_detect_10k;

pub use benchmarks::get_benchmark;

pub use benchmarks::get_industry_benchmark;

pub use benchmarks::graph_fraud_10k;

pub use benchmarks::healthcare_fraud_5k;

pub use benchmarks::manufacturing_fraud_5k;

pub use benchmarks::retail_fraud_10k;

pub use benchmarks::technology_fraud_3k;

pub use benchmarks::AcfeAlignment;

pub use benchmarks::AcfeCalibration;

pub use benchmarks::AcfeCategoryDistribution;

pub use benchmarks::BaselineModelType;

pub use benchmarks::BaselineResult;

pub use benchmarks::BenchmarkBuilder;

pub use benchmarks::BenchmarkSuite;

pub use benchmarks::BenchmarkTaskType;

pub use benchmarks::CostMatrix;

pub use benchmarks::DatasetSpec;

pub use benchmarks::EvaluationSpec;

pub use benchmarks::FeatureSet;

pub use benchmarks::IndustryBenchmarkAnalysis;

pub use benchmarks::LeaderboardEntry;

pub use benchmarks::MetricType;

pub use benchmarks::SplitRatios;

pub use banking::AmlDetectabilityAnalysis;

pub use banking::AmlDetectabilityAnalyzer;

pub use banking::AmlTransactionData;

pub use banking::BankingEvaluation;

pub use banking::KycCompletenessAnalysis;

pub use banking::KycCompletenessAnalyzer;

pub use banking::KycProfileData;

pub use banking::TypologyData;

pub use process_mining::EventSequenceAnalysis;

pub use process_mining::EventSequenceAnalyzer;

pub use process_mining::ProcessEventData;

pub use process_mining::ProcessMiningEvaluation;

pub use process_mining::VariantAnalysis;

pub use process_mining::VariantAnalyzer;

pub use process_mining::VariantData;

pub use causal::CausalModelEvaluation;

pub use causal::CausalModelEvaluator;

pub use enrichment::EnrichmentQualityEvaluation;

pub use enrichment::EnrichmentQualityEvaluator;

Modules

banking

Banking/KYC/AML evaluation module.

benchmarks

Benchmark suite definitions for ML evaluation.

causal

Causal model evaluator.

coherence

Semantic coherence evaluation module.

config

Configuration for the evaluation framework.

diff_engine

Diff engine for comparing baseline vs counterfactual output directories.

enhancement

Enhancement derivation module for automatic configuration optimization.

enrichment

LLM enrichment quality evaluator.

error

Error types for the evaluation framework.

gates

Quality gate engine for pass/fail criteria on generation runs.

ml

ML-readiness evaluation module.

privacy

Privacy evaluation module.

process_mining

OCEL 2.0 process mining evaluation module.

quality

Data quality evaluation module.

report

Report generation module.

scenario_diff

Scenario diff types for baseline vs counterfactual comparison.

statistical

Statistical quality evaluation module.

tuning

Configuration tuning and optimization suggestions.

Structs

ComprehensiveEvaluation

Comprehensive evaluation result combining all evaluation modules.

Evaluator

Main evaluator that coordinates all evaluation modules.