Investigation: Verify the value sets against the ORE code

The curve leaves BusinessDayConvention and DayCounter are declared as catalogue sets over ores_refdata_business_day_convention_catalogue_v and ores_refdata_day_counter_catalogue_v. Those views expose the registries' code, and the registries hold words: Following, 1/1. An ORE document states its own spelling, F and ACT/365, so the guard refuses a document that ORE accepts. The two sets are text until the registries carry the spelling the document writes; the finding is the reason, and it is not one of the thirty-one below, which compare sets with the code that reads a parameter and not with the rows a catalogue holds.

• Set code: calibrationstrategy
• Parameter: ModelParameters/CalibrationStrategy
• Set as it stands: enumerated, CoterminalDealStrike, CoterminalATM, DeltaGammaAdjusted, None.
• What ORE accepts: six values. The parser accepts CoterminalATM, CoterminalDealStrike, DeltaGammaAdjusted, UnderlyingATM, UnderlyingDealStrike and None. The enum holds the same six values.

• Evidence:

  // OREData/ored/model/irmodeldata.cpp:76
  CalibrationStrategy parseCalibrationStrategy(const string& s) {
      if (boost::algorithm::to_upper_copy(s) == "COTERMINALATM")
          return CalibrationStrategy::CoterminalATM;
      ...
      else if (boost::algorithm::to_upper_copy(s) == "UNDERLYINGATM")
          return CalibrationStrategy::UnderlyingATM;
      else if (boost::algorithm::to_upper_copy(s) == "UNDERLYINGDEALSTRIKE")
          return CalibrationStrategy::UnderlyingDealStrike;
  

  // OREData/ored/model/irmodeldata.hpp:79
  enum class CalibrationStrategy { CoterminalATM, CoterminalDealStrike,
      UnderlyingATM, UnderlyingDealStrike, DeltaGammaAdjusted, None };
  

• Case handling: the parser uses boost::algorithm::to_upper_copy. The exact spelling of a set value does not matter to ORE.
• Direction: too narrow. The screen refuses UnderlyingATM and UnderlyingDealStrike, and ORE accepts both.
• Correction: in projects/ores.seeder/datasets/ore_pricing_engines/curation/value-sets-guide-enums.json, under sets, entry calibrationstrategy, add UnderlyingATM and UnderlyingDealStrike to values, with order 4 and order 5.

• Set code: scheme
• Parameter: EngineParameters/Scheme
• Set as it stands: enumerated, the nine FDM schemes plus Segment: CrankNicolson, Hundsdorfer, Douglas, CraigSneyd, ModifiedCraigSneyd, ImplicitEuler, ExplicitEuler, MethodOfLines, TrBDF2, Segment.

• What ORE accepts: ORE reads Scheme in two engine families, and the accepted lists differ.

  The finite difference engines parse nine schemes:

  // OREData/ored/utilities/parsers.cpp:754
  FdmSchemeDesc parseFdmSchemeDesc(const std::string& s) {
      static std::map<std::string, FdmSchemeDesc> m = {
          {"CrankNicolson", ...}, {"Hundsdorfer", ...}, {"Douglas", ...},
          {"CraigSneyd", ...}, {"ModifiedCraigSneyd", ...},
          {"ImplicitEuler", ...}, {"ExplicitEuler", ...},
          {"MethodOfLines", ...}, {"TrBDF2", ...}};
  

  The variance swap engine accepts GaussLobatto and Segment only:

  // OREData/ored/portfolio/builders/varianceswap.hpp:107
  std::string schemeStr = engineParameter("Scheme", {}, false, "GaussLobatto");
  if (schemeStr == "GaussLobatto")
      settings.scheme = ...::Scheme::GaussLobatto;
  else if (schemeStr == "Segment")
      settings.scheme = ...::Scheme::Segment;
  else
      QL_FAIL("invalid var swap pricing engine parameter Scheme (...)");
  

• Case handling: both comparison functions are case-sensitive.
• Direction: too narrow and too wide. The set refuses GaussLobatto, which the variance swap engine accepts and uses as its default. The set accepts Segment, which every finite difference engine refuses.
• Correction: in projects/ores.seeder/datasets/ore_pricing_engines/curation/value-sets-code.json, under sets, entry scheme, add GaussLobatto to values. State in the description that GaussLobatto and Segment apply to the variance swap engine, and that Segment does not apply to a finite difference engine. A single set cannot be exact for a name that two engine families read. The union of the two lists is the closest exact set that the model can hold.

• Set code: calibration
• Parameter: ModelParameters/Calibration
• Set as it stands: text, no values.

• What ORE accepts: five values, compared after upper casing.

  // OREData/ored/model/irmodeldata.cpp:47
  CalibrationType parseCalibrationType(const string& s) {
      if (boost::algorithm::to_upper_copy(s) == "BOOTSTRAP")
      else if (boost::algorithm::to_upper_copy(s) == "BESTFIT")
      else if (boost::algorithm::to_upper_copy(s) == "FIRSTBESTFITTHENBOOTSTRAP")
      else if (boost::algorithm::to_upper_copy(s) == "STATISTICALWITHRISKNEUTRALVOLATILITY")
      else if (boost::algorithm::to_upper_copy(s) == "NONE")
      else QL_FAIL("Calibration type " << s << " not recognized");
  

• Case handling: case-insensitive.
• Direction: wrong kind. The screen shows a text field, so a person can save Deal or any other word. ORE refuses that word.
• Correction: in projects/ores.seeder/datasets/ore_pricing_engines/curation/value-sets-guide-enums.json, under sets, entry calibration, set source_kind to enumerated and add the five values Bootstrap, BestFit, FirstBestFitThenBootstrap, StatisticalWithRiskNeutralVolatility and None.

• Set code: model
• Parameter: ModelParameters/Model
• Set as it stands: text, no values.

• What ORE accepts: five model names. Any other name reaches the final QL_FAIL.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:276
  if (modelParam_ == "BlackScholes" && engineParam_ == "MC") {
  } else if ((modelParam_ == "LocalVolDupire" || modelParam_ == "LocalVolAndreasenHuge") && ...) {
  } else if (modelParam_ == "Heston" && ...) {
  } else if (modelParam_ == "GaussianCam" && ...) {
  } else {
      QL_FAIL("model '" << modelParam_ << "' / engine '" << engineParam_
                        << "' not recognised, ...");
  }
  

• Case handling: case-sensitive.
• Direction: wrong kind.
• Correction: in projects/ores.seeder/datasets/ore_pricing_engines/curation/value-sets-code.json, under sets, entry model, set source_kind to enumerated and add BlackScholes, LocalVolDupire, LocalVolAndreasenHuge, Heston and GaussianCam.

• Set code: engine
• Parameter: EngineParameters/Engine
• Set as it stands: text, no values.

• What ORE accepts: the values MC, FD, AMC and AMCCG. The builder reads the parameter at scriptedtrade.cpp:126 and fails on any other pairing.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:305
  } else {
      QL_FAIL("model '" << modelParam_ << "' / engine '" << engineParam_
                        << "' not recognised, expected BlackScholes/[MC|FD], ...");
  }
  

  AMC and AMCCG are the builder names that set buildingAmc_ and buildingAmcCg_ (same file, lines 129-133).

• Case handling: case-sensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry engine, set source_kind to enumerated and add MC, FD, AMC and AMCCG.

• Set code: bounds
• Parameter: EngineParameters/Bounds
• Set as it stands: text, no values.

• What ORE accepts: Fixed and PriceThreshold.

  // OREData/ored/portfolio/builders/varianceswap.hpp:108
  std::string boundsStr = engineParameter("Bounds", {}, false, "PriceThreshold");
  if (boundsStr == "Fixed") ...
  else if (boundsStr == "PriceThreshold") ...
  else QL_FAIL("invalid var swap pricing engine parameter Bounds (...)");
  

• Case handling: case-sensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry bounds, set source_kind to enumerated and add Fixed and PriceThreshold.

• Set code: timingadjustment
• Parameter: EngineParameters/TimingAdjustment
• Set as it stands: text, no values.

• What ORE accepts: Black76 and BivariateLognormal.

  // OREData/ored/portfolio/builders/capfloorediborleg.cpp:47
  typename T::TimingAdjustment timingAdjustment = T::Black76;
  if (timingAdjustmentStr == "Black76")
      timingAdjustment = T::Black76;
  else if (timingAdjustmentStr == "BivariateLognormal")
      timingAdjustment = T::BivariateLognormal;
  else QL_FAIL("timing adjustment parameter (...) not recognised.");
  

• Case handling: case-sensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry timingadjustment, set source_kind to enumerated and add Black76 and BivariateLognormal.

• Set code: heston_processdiscretization
• Parameter: EngineParameters/Heston.ProcessDiscretization
• Set as it stands: text, no values.

• What ORE accepts: the nine values that parseHestonProcessDiscretization lists.

  // OREData/ored/utilities/parsers.cpp:1911
  {"PartialTruncation", ...}, {"FullTruncation", ...}, {"Reflection", ...},
  {"NonCentralChiSquareVariance", ...}, {"QuadraticExponential", ...},
  {"QuadraticExponentialMartingale", ...},
  {"BroadieKayaExactSchemeLobatto", ...},
  {"BroadieKayaExactSchemeLaguerre", ...},
  {"BroadieKayaExactSchemeTrapezoidal", ...}
  

  The call site is scriptedtrade.cpp:612.

• Case handling: case-sensitive.
• Direction: wrong kind. The set heston_quantoprocessdiscretization already holds these nine values as an enumerated set.
• Correction: in value-sets-code.json, under sets, entry heston_processdiscretization, set source_kind to enumerated and add the nine values that heston_quantoprocessdiscretization holds.

• Set code: heston_calibrationmethod
• Parameter: EngineParameters/Heston.CalibrationMethod
• Set as it stands: text, no values.

• What ORE accepts: ConstantBestFit, None, PiecewiseBootstrap and PiecewiseBestFit.

  // OREData/ored/model/hestonmodelcalibration.cpp:588
  if (calibrationMethod_ == "PiecewiseBootstrap") {
  ...
  // OREData/ored/model/hestonmodelcalibration.cpp:622
  } else if (calibrationMethod_ == "PiecewiseBestFit") {
  ...
  // OREData/ored/model/hestonmodelcalibration.cpp:643
  } else { QL_FAIL("unknown calibration method \"" << calibrationMethod_ << "\""); }
  

  ConstantBestFit and None are the two further branches at hestonmodelbuilder.cpp:74 and hestonmodelcalibration.cpp:255.

• Case handling: case-sensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry heston_calibrationmethod, set source_kind to enumerated and add the four values.

• Set code: bootstrap_mode
• Parameter: EngineParameters/Bootstrap.Mode
• Set as it stands: text, no values.

• What ORE accepts: Alternating and Simultaneously.

  // OREData/ored/portfolio/builders/convertiblebond.cpp:75
  auto bootstrapModeStr = engineParameter("Bootstrap.Mode", {}, true);
  if (bootstrapModeStr == "Alternating") ...
  else if (bootstrapModeStr == "Simultaneously") ...
  else QL_FAIL("invalid Bootstrap.Mode, expected Alternating or Simultaneously");
  

• Case handling: case-sensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry bootstrap_mode, set source_kind to enumerated and add Alternating and Simultaneously.

• Set code: yieldcurvemodel
• Parameter: EngineParameters/YieldCurveModel
• Set as it stands: text, no values.

• What ORE accepts: Standard, ExactYield, ParallelShifts and NonParallelShifts.

  // OREData/ored/portfolio/builders/cms.cpp:30
  GFunctionFactory::YieldCurveModel ycmFromString(const string& s) {
      if (s == "Standard") ... else if (s == "ExactYield") ...
      else if (s == "ParallelShifts") ...
      else if (s == "NonParallelShifts") ...
      else QL_FAIL("unknown string for YieldCurveModel");
  }
  

• Case handling: case-sensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry yieldcurvemodel, set source_kind to enumerated and add the four values.

• Set code: infmodeltype
• Parameter: ModelParameters/InfModelType
• Set as it stands: text, no values.

• What ORE accepts: DK and JY.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:580
  infModelType_ = modelParameter("InfModelType", getModelEngineQualifiers(), false, "DK");
  // OREData/ored/portfolio/builders/scriptedtrade.cpp:873
  if (infModelType_ == "DK") { ... }
  // OREData/ored/model/camcorrelationsfromindexcorrelations.cpp:69
  if (infModelType == "JY") { ... }
  

• Case handling: case-sensitive.
• Direction: wrong kind. A third word passes the reader and then silently takes the JY branch or builds no inflation model.
• Correction: in value-sets-code.json, under sets, entry infmodeltype, set source_kind to enumerated and add DK and JY.

• Set code: floatspreadmapping
• Parameters: ModelParameters/FloatSpreadMapping and EngineParameters/FloatSpreadMapping
• Set as it stands: text, no values.

• What ORE accepts: NextCoupon, proRata and Simple.

  // OREData/ored/model/lgmdata.cpp:100
  QuantExt::AnalyticLgmSwaptionEngine::FloatSpreadMapping parseFloatSpreadMapping(const string& s) {
      if (boost::algorithm::to_upper_copy(s) == "NEXTCOUPON")
      else if (boost::algorithm::to_upper_copy(s) == "PRORATA")
      else if (boost::algorithm::to_upper_copy(s) == "SIMPLE")
      else QL_FAIL("FloatSpreadMapping '" << s << "' not recognized");
  }
  

• Case handling: case-insensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry floatspreadmapping, set source_kind to enumerated and add NextCoupon, proRata and Simple.

• Set code: irvolatilitytype
• Parameter: ModelParameters/IrVolatilityType
• Set as it stands: text, no values.

• What ORE accepts: Hagan and HullWhite.

  // OREData/ored/model/lgmdata.cpp:81
  LgmData::VolatilityType parseVolatilityType(const string& s) {
      if (boost::algorithm::to_upper_copy(s) == "HULLWHITE")
      else if (boost::algorithm::to_upper_copy(s) == "HAGAN")
      else QL_FAIL("Volatility type " << s << " not recognized");
  }
  

  The call site is scriptedtrade.cpp:1637. The set volatilitytype already holds these two values.

• Case handling: case-insensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry irvolatilitytype, set source_kind to enumerated and add Hagan and HullWhite.

• Set code: regression_vargroupmode
• Parameter: EngineParameters/Regression.VarGroupMode
• Set as it stands: text, no values.

• What ORE accepts: Global and Trivial.

  // OREData/ored/utilities/parsers.cpp:1657
  QuantExt::McRegressionModel::VarGroupMode parseVarGroupMode(const std::string& s) {
      if (s == "Global") ... else if (s == "Trivial") ...
      else QL_FAIL("VarGroupMode '" << s << "' not recognized, expected Global, Trivial");
  }
  

• Case handling: case-sensitive.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry regression_vargroupmode, set source_kind to enumerated and add Global and Trivial.

• Set code: globalmodel
• Parameter: ModelParameters/GlobalModel
• Set as it stands: text, no values.

• What ORE accepts: a bool. parseBool accepts Y, YES, TRUE, True, true, 1, N, NO, FALSE, False, false and 0.

  // OREData/ored/portfolio/flexiswap.cpp:205
  bool useGlobalModel = parseBool(flexiSwapBuilder->modelParameter("GlobalModel", {}, true));
  

  The bool parser is parsers.cpp:153.

• Case handling: the bool parser has a fixed spelling list.
• Direction: wrong kind. The screen shows a text field for a switch.
• Correction: in value-sets-code.json, under sets, entry globalmodel, set source_kind to boolean and remove values.

• Set code: usead
• Parameter: EngineParameters/UseAD
• Set as it stands: text, no values.

• What ORE accepts: a bool.

  // OREData/ored/portfolio/builders/swaption.cpp:125
  bool useCachedSensis = parseBool(builder->engineParameter("UseAD", {}, true)) && ...;
  // OREData/ored/portfolio/builders/scriptedtrade.cpp:538
  useAd_ = parseBool(engineParameter("UseAD", getModelEngineQualifiers(), false, "false"));
  

• Case handling: the bool parser has a fixed spelling list.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry usead, set source_kind to boolean and remove values.

• Set code: useindexcurveforindexcds
• Parameter: EngineParameters/useIndexCurveForIndexCDS
• Set as it stands: text, no values.

• What ORE accepts: a bool.

  // OREData/ored/portfolio/cdo.cpp:597
  bool useIndexCurve =
      parseBool(cdoEngineBuilder->engineParameter("useIndexCurveForIndexCDS", {}, false, "false"));
  

• Case handling: the bool parser has a fixed spelling list.
• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry useindexcurveforindexcds, set source_kind to boolean and remove values.

• Set code: beta
• Parameter: EngineParameters/beta
• Set as it stands: number, no minimum and no maximum.

• What ORE accepts: any value that is greater than or equal to zero.

  // OREData/ored/portfolio/builders/commodityswaption.hpp:60
  Real beta = parseReal(engineParameter("beta"));
  QL_REQUIRE(beta >= 0.0, "CommoditySwaptionAnalyticalEngineBuilder: beta must be non-negative");
  

  The same check is at line 79.

• Direction: wrong bounds. The screen accepts -0.5, and ORE refuses it.
• Correction: in value-sets-code.json, under sets, entry beta, add "minimum": 0.

These sets are integer with no minimum and no maximum. ORE reads each value with parseInteger and stores it in a QuantLib::Size. A Size is unsigned. ORE silently converts a negative value to a very large positive value. The value set must refuse a negative value.

// OREData/ored/utilities/parsers.cpp:145
Integer parseInteger(const string& s) { ... boost::lexical_cast<Integer>(s.c_str()) ... }

// OREData/ored/portfolio/builders/cdo.hpp:281
Size nBuckets = parseInteger(engineParameter("buckets"));
// OREData/ored/portfolio/builders/cdo.hpp:358
Size nSimulations = parseInteger(engineParameter("nSimulations"));
// OREData/ored/portfolio/builders/callablebond.cpp:392
Size stateGridPoints = parseInteger(engineParameter("StateGridPoints"));
// OREData/ored/portfolio/builders/callablebond.cpp:408
Size ny = parseInteger(engineParameter("ny"));
// OREData/ored/portfolio/builders/riskparticipationagreement.cpp:140
Size maxGapDays = parseInteger(engineParameter("MaxGapDays"));

The sets and their parameters:

• Direction: wrong bounds. The screen accepts a negative number, and ORE does not reject it. ORE converts it to a huge Size and the run behaves wrongly.
• Correction: in value-sets-code.json, under sets, add "minimum": 0 to each entry above. Do not add 0 to the seed, pricing_seed, training_seed and trainingseed entries. ORE reads a seed as a signed long (cbo.cpp:45).

• Set code: trainingsamples
• Parameter: EngineParameters/TrainingSamples
• Set as it stands: text, no values.

• What ORE accepts: a whole number.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:634
  if (auto tmp = engineParameter("TrainingSamples", getModelEngineQualifiers(), false, ""); !tmp.empty()) {
      params_.trainingSamples = parseInteger(tmp);
  

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry trainingsamples, set source_kind to integer and add "minimum": 0.

• Set code: regressionvariancecutoff
• Parameter: EngineParameters/RegressionVarianceCutoff
• Set as it stands: text, no values.

• What ORE accepts: a real number or an empty string.

  // OREData/ored/portfolio/builders/callablebond.cpp:477
  parseRealOrNull(engineParameter("RegressionVarianceCutoff", {}, false, std::string())),
  

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry regressionvariancecutoff, set source_kind to number. Add the description note that an empty value is allowed.

• Set code: indicatorsmoothingforvalues
• Parameter: EngineParameters/IndicatorSmoothingForValues
• Set as it stands: text, no values.

• What ORE accepts: a real number.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:552
  parseReal(engineParameter("IndicatorSmoothingForValues", ..., false, "0.0"));
  // OREData/ored/portfolio/builders/swaption.cpp:128
  parseReal(builder->engineParameter("IndicatorSmoothingForValues", {}, false, "0.0")),
  

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry indicatorsmoothingforvalues, set source_kind to number.

• Set code: indicatorsmoothingforderivatives
• Parameter: EngineParameters/IndicatorSmoothingForDerivatives
• Set as it stands: text, no values.

• What ORE accepts: a real number.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:554
  parseReal(engineParameter("IndicatorSmoothingForDerivatives", ..., false, "0.2"));
  // OREData/ored/portfolio/builders/swaption.cpp:129
  parseReal(builder->engineParameter("IndicatorSmoothingForDerivatives", {}, false, "0.0")),
  

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry indicatorsmoothingforderivatives, set source_kind to number.

• Set code: sqrtsmoothingforderivatives
• Parameter: EngineParameters/SqrtSmoothingForDerivatives
• Set as it stands: text, no values.

• What ORE accepts: a real number.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:556
  parseReal(engineParameter("SqrtSmoothingForDerivatives", ..., false, "1E-8"));
  // OREData/ored/portfolio/builders/swaption.cpp:130
  parseReal(builder->engineParameter("SqrtSmoothingForDerivatives", {}, false, "0.0")),
  

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry sqrtsmoothingforderivatives, set source_kind to number.

• Set code: displacement
• Parameter: ModelParameters/Displacement
• Set as it stands: text, no values.

• What ORE accepts: a real number.

  // OREData/ored/portfolio/builders/vanillaoption.hpp:164
  auto displacementStr = this->modelParameter("Displacement", {assetName}, false, "0.0");
  

  The value is then parsed as a real in the same function.

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry displacement, set source_kind to number.

• Set code: irreversion
• Parameters: ModelParameters/IrReversion and ModelParameters/IrReversion_
• Set as it stands: text, no values.

• What ORE accepts: a real number.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:1155
  string revStr = modelParameter("IrReversion_" + ccy, ..., false, "");
  if (revStr.empty())
      revStr = modelParameter("IrReversion", ..., false, "");
  ...
  irReversions_[ccy] = parseReal(revStr);
  

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry irreversion, set source_kind to number.

• Set code: infjyrealratereversion
• Parameter: ModelParameters/InfJyRealRateReversion
• Set as it stands: text, no values.

• What ORE accepts: a real number. The sibling times and volumes are comma lists, but the reversion is one value.

  // OREData/ored/portfolio/builders/scriptedtrade.cpp:1752
  auto tmp4 = modelParameter("InfJyRealRateReversion", ..., true);
  ...
  ReversionParameter realRateRev(..., {parseReal(tmp4)});
  

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry infjyrealratereversion, set source_kind to number.

• Set code: bootstrap_mesherconcentration
• Parameter: EngineParameters/Bootstrap.MesherConcentration
• Set as it stands: text, no values.

• What ORE accepts: a real number or an empty string.

  // OREData/ored/portfolio/builders/convertiblebond.cpp:71
  auto modelMesherConcentrationStr = engineParameter("Bootstrap.MesherConcentration", {}, false, "");
  if (!modelMesherConcentrationStr.empty())
      modelMesherConcentration = parseReal(modelMesherConcentrationStr);
  

• Direction: wrong kind.
• Correction: in value-sets-code.json, under sets, entry bootstrap_mesherconcentration, set source_kind to number.

• Set codes: convention.period and curve.period
• Parameters: convention.period serves 17 terms, for example InflationSwap|ObservationLag, FXOption|SwitchTenor and CrossCurrencyBasis|SpreadTenor. curve.period serves InflationCapFloorVolatilities|ObservationLag.
• Set as it stands: regex, pattern ^([0-9]+[DdWwMmYy])+$.

• What ORE accepts: the QuantLib::PeriodParser grammar. The grammar takes one or more number-unit pairs and it allows a leading sign.

  // QuantLib/ql/utilities/dataparsers.cpp:40
  Period PeriodParser::parse(const std::string& str) {
      QL_REQUIRE(str.length()>1, ...);
      ... find_first_of("DdWwMmYy") ...
  // QuantLib/ql/utilities/dataparsers.cpp:77
  Size nPos = str.find_first_of("-+0123456789");
  QL_REQUIRE(nPos<iPos, "no numbers of " << units << " provided");
  nInteger n = std::stoi(str.substr(nPos,iPos));
  

• Case handling: the parser accepts an upper case unit and a lower case unit.
• Direction: too narrow. ORE parses -3M and +1Y. The pattern refuses both. A negative observation lag is meaningful for an inflation index.
• Correction: this pattern is this repository's own choice. ORE enforces a parser, not a pattern. In projects/ores.sql/utility/convention_value_sets.py, OWN_SETS, entry convention.period, and in projects/ores.sql/utility/curve_value_sets.py, OWN_SETS, entry curve.period, change pattern to ^[+-]?([0-9]+[DdWwMmYy])+$. A simpler option is to state text for these two sets, because ORE holds no pattern.

• Set code: curve.yield_curve_interpolation_method
• Parameter: YieldCurves|InterpolationMethod
• Set as it stands: enumerated, 20 values. It holds KrugerLogMixedLinearCubic and LogMixedLinearCubicNaturalSpline.

• What ORE accepts: 21 values. The parser also accepts DefaultLogMixedLinearCubic as a deprecated alias of KrugerLogMixedLinearCubic.

  // OREData/ored/marketdata/yieldcurve.cpp:336
  else if (s == "DefaultLogMixedLinearCubic") // deprecated,for backwards compatibility
      return YieldCurve::InterpolationMethod::KrugerLogMixedLinearCubic;
  

• Case handling: case-sensitive.
• Direction: too narrow. The alias is deprecated, so a configuration should not use it. ORE still accepts it.
• Correction: the value list is generated from projects/ores.sql/utility/curve_terms.json, key parsers, entry parseYieldCurveInterpolationMethod, list values. Add {"means": "KrugerLogMixedLinearCubic", "value": "DefaultLogMixedLinearCubic"} and regenerate the terms file.

The evidence is one builder's comparison for a scripted trade. The same element names a builder from ORE's registry for every other product type, and the registry holds far more names than the five a scripted trade accepts. The seeded population states LGM, BlackScholesMerton, DiscountedCashflows, CrossAssetModel, GarmanKohlhagen and eleven more, and the valid name depends on the product type. A set of five would refuse every valid interest rate configuration, so both sets stay text and their descriptions say what they name.

The IR and FX builders read it through parseCalibrationType, which accepts five values. The scripted trade builder stores the same parameter as a plain string with a default of Deal and never compares it, so no list bounds it there. The union of a five-value list and an unvalidated string is text, and the set says so.

The audit did not reach the catalogue contents. ORE's parser maps 26 spellings onto seven conventions, and the registry holds the seven; the catalogue view states the crosswalk, so a document that says F and one that says Following are the same rule. The day counter registry already held ORE's spellings.

The table lists every set that this audit verified, with one verdict for each set. The seven catalogue sets are not in the table. They hold a relation name and no values, so this audit does not compare them.

Verdict counts for the 217 verified sets:

ORE never reads this value from a configuration and never validates it. ORE writes the value itself, for example at analytic.cpp:243 (NPV), xvaanalytic.cpp:573 (Exposure), stresstest.cpp:149 (Stress) and sensitivityanalysis.cpp:213 (SensitivityDelta or SensitivityDeltaGamma). The builders compare the value with a few words, for example cdo.hpp:89 and swaption.cpp:125. A user can write any word in the XML and ORE does not refuse it. The parameter is therefore neither a fixed list nor free text. Decide whether the setting belongs in the model at all before a value set is authored.

ORE sets this value and never parses it. xvaanalytic.cpp:574 sets Classic, xvaanalytic.cpp:986 sets American, and capfloor.cpp:141 compares it with American. The audit found no reader that accepts or refuses a user value. Decide whether the setting belongs in the model before a value set is authored.

The audit found no reader of this name in the ORE source. The link may come from the ORE guide only. A search of OREData, OREAnalytics and QuantExt found no string literal calibrationIndexTerms.

EngineParameters/ConversionRatioDiscretizationGrid

The audit found no reader of this spelling. ORE reads Pricing.ConversionRatioDiscretisationGrid with an s at convertiblebond.cpp:89, and that name has its own set. The two names look like a duplicate. Confirm which name ORE uses before a value set is authored.