-
Notifications
You must be signed in to change notification settings - Fork 43
Migration guide v6.3.0
Starting from this release, Reporter
and Report
interfaces have been merged into a ReportNode
interface, which contains the same features. The com.powsybl.commons.reporter
package has been renamed into com.powsybl.commons.report
, hence the corresponding imports are impacted.
- The root
ReportNode
can be build thanks to the builder provided byReportNode.newRootReportNode()
, - A
ReportNode
can be added within anotherReportNode
thanks to the adder provided byreportNode.newReportNode()
.
The mentioned adder and builder share the same methods, with the following correspondence to former ReportBuilder
methods:
-
withMessageTemplate(String messageKey, String messageTemplate)
corresponds toreportBuilder.withKey(messageKey).withDefaultMessage(messageTemplate)
- the
withTypedValue(key, value, type)
methods keep the same signature as in the formerReportBuilder
- the
withUntypedValue(key, value, type)
methods corresponds toreportBuilder.withValue(key, value, type)
-
withSeverity(TypedValue severity)
, which corresponds toreportBuilder.withSeverity(TypedValue severity)
-
withSeverity(String severity)
, which corresponds toreportBuilder.withSeverity(new TypedValue(severity, TypedValue.SEVERITY)
Here are some replacement examples to help you migrate:
reporter.report(Report.builder()
.withKey("keyId")
.withDefaultMessage("New report with typed value ${typedValue}, untyped value ${untypedValue} and severity ${reportSeverity}")
.withTypedValue("typedValue", 20, "type")
.withValue("untypedValue", 3.)
.withSeverity(TypedValue.INFO_SEVERITY)
.build());
with
ReportNode childReportNode = reportNode.newReportNode()
.withMessageTemplate("keyId", "New report with typed value ${typedValue}, untyped value ${untypedValue} and severity ${reportSeverity}")
.withTypedValue("typedValue", 20, "type")
.withUntypedValue("untypedValue", 3.)
.withSeverity(TypedValue.INFO_SEVERITY)
.add();
reporter.report(reportNode);
with
reporter.include(reportNode);
Note that the included ReportNode
needs to be a root.
Previously adding a report / sub-reporter with createSubReporter
/ report
, now adding a child ReportNode with an adder
Replace
Reporter subReporter = reporter.createSubReporter(taskKey, defaultName);
or
reporter.report(reportKey, defaultMessage);
with
ReportNode childReportNode = reportNode.newReportNode()
.withMessageTemplate(reportKey, defaultMessage)
.add();
The replacements are similar for the other Reporter::createSubReporter
and Reporter::report
methods.
Replace the root creation
Reporter reporter = new ReporterModel(taskKey, defaultName, Maps.of(valueKey, new TypedValue(value, type)));
with
ReportNode rootReportNode = ReportNode.newRootReportNode()
.withMessageTemplate(taskKey, defaultName)
.withTypedValue(valueKey, value, type)
.build();
As a side-effect, the ReporterContext
interface has been renamed into ReportNodeContext
, and all the methods have been renamed. You need to replace the call to
-
ReporterContext::peekReporter
byReportNodeContext::peekReportNode
-
ReporterContext::getReporter
byReportNodeContext::getReportNode
-
ReporterContext::pushReporter
byReportNodeContext::pushReportNode
-
ReporterContext::popReportNode
byReportNodeContext::popReportNode
The implementations have also been renamed:
-
SimpleReporterContext
intoSimpleReportNodeContext
-
MultiThreadReporterContext
intoMultiThreadReportNodeContext
Lastly, Networks::allowReporterContextMultiThreadAccess
has therefore been renamed into Networks::allowReportNodeContextMultiThreadAccess
.
In the LoadFlowParameters
, the default value of writeSlackBus
parameter is now true
. This is a functional change, that could lead to differences during unit and integration tests.
Some criteria used for contingency definition were moved in a dedicated new module: powsybl-iidm-criteria
.
If you used the following classes, you should add the new module in your project dependencies and update your import statements:
- previously in the
com.powsybl.contingency.contingency.list.criterion
package, now incom.powsybl.iidm.criteria
:Criterion
PropertyCriterion
RegexCriterion
SingleCountryCriterion
SingleNominalVoltageCriterion
ThreeNominalVoltageCriterion
TwoCountriesCriterion
TwoNominalVoltageCriterion
- previously in the
com.powsybl.contingency.json
package, now incom.powsybl.iidm.criteria.json
:CriterionDeserializer
CriterionSerializer
SingleNominalVoltageCriterion.VoltageInterval
was renommed as VoltageInterval
.
The following methods of VoltageInterval
were renamed to match the bean methods syntax:
-
getLowClosed()
=>isLowClosed()
-
getHighClosed()
=>isHighClosed()
Intervals with only one bound are now considered as valid. This have the following impacts:
VoltageInterval.getNominalVoltageLowBound()
and VoltageInterval.getNominalVoltageHighBound()
now return an Optional<Double>
instead of a Double
to indicated if a bound was defined or not.
The concept of "null interval" was removed, as for the VoltageInterval.isNull()
method. If you need to indicate that an interval was not set, use a "null" value instead.
As a consequence, getVoltageInterval...()
methods of TwoNominalVoltageCriterion
and ThreeNominalVoltageCriterion
now return an Optional<VoltageInterval>
instead of a VoltageInterval
.
The historical constructor was removed. Instead, you can use either the build()
static method which returns a builder, or one of the greaterThan
, lowerThan
and between
static methods which return an interval with one or both bounds.
For instance, you can replace:
new VoltageInterval(200.0, 230.0, true, true)
by either
VoltageInterval.builder()
.setLowBound(200.0, true)
.setHighBound(230.0, true)
.build()
or
VoltageInterval.between(200.0, 230.0, true, true)
The following utilitary methods:
-
ValidationUtil.checkLoadingLimits(...)
; -
ValidationUtil.checkPermanentLimit(...)
;
now take two additional parameters:
-
boolean throwException
: should the method throw an exception when a missing permanent limit is detected? -
ReportNode reportNode
: the reportNode to use to log potential errors. You can useReportNode.NO_OP
if you are not interested by this log.
SecurityAnalysisProvider.run(…)
now takes an additional parameter: List<LimitReduction> limitReductions
. You need to add this parameter to your own SecurityAnalysisProvider implementations.
When you used the constructor with all the parameters to create a "LimitReduction", you should use the builder. Default values for the characteristics are:
-
monitoringOnly
: false; -
contingencyContext
:ContingencyContext.all()
; -
networkElementCriteria
:Collections.emptyList()
; -
limitDurationCriteria
:Collections.emptyList()
.
Thus, you can omit setting one of the characteristics if you used its default value.
For instance:
LimitReduction limitReduction = new LimitReduction(LimitType.APPARENT_POWER, 0.5, false,
ContingencyContext.all(),
Collections.emptyList(),
List.of(durationCriterion1, durationCriterion2));
could be rewritten in
LimitReduction limitReduction = LimitReduction.builder(LimitType.APPARENT_POWER, 0.5)
.withLimitDurationCriteria(durationCriterion1, durationCriterion2)
.build();
Note:
-
monitoringOnly
,networkElementCriteria
andcontingencyContext
are not specified as their default values are used; - the
with...Criteria(...)
methods take either aList
or varArgs. You could thus omit the encapsulatingList.of()
for ease of use. But in either case, the previously definied criteria are replaced by the given ones.
-
LimitViolation.getLimitReduction
now returns adouble
instead of afloat
; - If you have defined your own IIDM implementation, you should change the type of the
limitReductionValue
parameter fromfloat
todouble
in thecheckPermanentLimit…
andcheckTemporaryLimit…
methods of yourBranch
's andThreeWindingTransformer
's implementations; - If you have a custom
LimitViolationDetector
's implementation, you should change the type of thelimitReductionValue
parameter fromfloat
todouble
in thecheckPermanentLimit
andcheckTemporaryLimit
methods; -
AbstractBranchActionExpressionNode.getLimitReduction()
now returns adouble
instead of afloat
.
The actions previously spread out in powsybl-core have been regrouped into the module powsybl-action
. If you were using either one of them, you will have to change your imports and dependencies accordingly:
- Package
com.powsybl.security.action
moved tocom.powsybl.action
in the new modulepowsybl-action-api
; - Package
com.powsybl.security.json.action
moved tocom.powsybl.action.json
; - The action part of
SecurityAnalysisJsonModule
has been relocated to the classActionJsonModule
and moved fromcom.powsybl.security.json
tocom.powsybl.action.json
. The rest ofSecurityAnalysisJsonModule
stays the same; - Module with artifactId
powsybl-action
renamedpowsybl-action-ial
; - Module with artifactId
powsybl-action-dsl
renamedpowsybl-action-ial-dsl
and packages moved accordingly; - Module with artifactId
powsybl-action-dsl-spi
renamedpowsybl-action-ial-dsl-spi
and packages moved accordingly; - Module with artifactId
powsybl-action-simulator
renamedpowsybl-action-ial-simulator
and packages moved accordingly; - Module with artifactId
powsybl-action-util
renamedpowsybl-action-ial-util
and packages moved accordingly.
Compatibility is supported. OperatorStrategyResult
object contains now a list of ConditionalActionsResult
objects instead of only the results on final state of the simulation. Each ConditionActionsResult
groups the result after applying the list of actions described in a ConditionalActions
if the condition is satisfied. For an operator strategy with a single ConditionalActions
object ConditionalActionsResult
will have a single object containing the same results as before this feature. For operator strategy with multiple ConditionalActions
objects, ConditionalActionsResult
will have one result per ConditionalActions
object and the last one in the list represents the last stage.
New BUS_REACTIVE_POWER
sensitivity function type has been added in sensitivity factor. You need to add a default behavior in switch cases or add a behavior to the new enum value.
A new CgmesMetadataModels
extension has been added, whose purpose is to store metadata information present in the Model
node of the instance files. Please refer to IEC 61970-552 for more details about these metadata.
The CgmesMetadataModels
may contain several instances of CgmesMetadataModel
, one per CgmesSubset
(EQ, TP, SV, SSH...). The CgmesMetadataModel
class now replaces the former classes CgmesSshMetadata
and CgmesSvMetadata
which have been removed in this version.
To retrieve the Ssh and Sv metadata associated to the network (after reading a CGMES case for instance), this must now be done as follow:
CgmesMetadataModels modelsExtension = network.getExtension(CgmesMetadataModels.class);
CgmesMetadataModel sshModel = modelsExtension.getModelForSubset(CgmesSubset.STEADY_STATE_HYPOTHESIS).orElseThrow();
CgmesMetadataModel svModel = modelsExtension.getModelForSubset(CgmesSubset.STATE_VARIABLES).orElseThrow();
String sshDescription = sshModel.getDescription();
int svVersion = svModel.getVersion();
Please note:
network.getExtension(CgmesMetadataModels.class)
can return a null value.
getModelForSubset
returns an optional.
To create a new CgmesMetadataModels
extension, to set metadata information for a CGMES export, this must be done as follow:
network.newExtension(CgmesMetadataModelsAdder.class)
.newModel()
.setSubset(CgmesSubset.STEADY_STATE_HYPOTHESIS)
.setId("sshId")
.setDescription("SSH description")
.setModelingAuthoritySet("http://powsybl.org")
.setVersion(1)
.addProfile("http://steady-state-hypothesis")
.add()
.add();
If you used the default TimeSeriesCsvConfig
by calling new TimeSeriesCsvConfig()
and imported some versioned TimeSeries with a version number at 0 (old default value for unversioned TimeSeries), it will now consider them as versioned TimeSeries since the new default value has been set to -1.
If you still want to import TimeSeries with a version number and have them considered as unversioned TimeSeries, you have to use the version number -1. However, it will now throw an exception. You need to replace your call from new TimeSeriesCsvConfig()
to new TimeSeriesCsvConfig(ZoneId.systemDefault(), TimeSeriesConstants.DEFAULT_SEPARATOR, true, TimeFormat.DATE_TIME, DEFAULT_MAX_COLUMNS, false)
in order to keep it working (you will still get warnings) or change your TimeSeries version number to another value if it's not meant to be an unversioned one.
Also, a third parameter was added to the static method parseCsv(BufferedReader reader, TimeSeriesCsvConfig timeSeriesCsvConfig)
: you have to use parseCsv(BufferedReader reader, TimeSeriesCsvConfig timeSeriesCsvConfig, ReportNode reportNode)
now.
If you have defined your own IIDM implementation, you should implement the following methods in your DanglingLine
implementations: DanglingLine setPairingKey(String pairingKey)
. Note that in the in-memory implementation, setting the paring key is allowed only for unpaired dangling lines. An exception is thrown otherwise.