Add m_includeZeroValues option to RimEnsembleStatistics (default on, preserving existing behavior). When disabled, exact zero values (e.g. from shut-in wells/groups) are excluded from the mean, P10/P50/P90 and custom percentile calculations in RimEnsembleStatisticsCase::calculate(). This is useful for derived vectors such as WCT, GOR and PI, where zeros caused by a shut well/group should not skew the statistical curves.
Exact equality (not an epsilon) is used, since shut-in values are reported as exact 0.0 by the simulator and non-accumulated vectors use STEP_RIGHT interpolation, which does not introduce floating-point noise.
Switch the message parameter from const QString& to std::string_view so
callers with std::string in hand no longer need a QString::fromStdString
wrapper. The internal duration formatting still uses QString::arg, fed
from a single QString::fromUtf8 conversion.
Callers passing QString locals are updated with .toStdString() at the
call site.
If percentiles could not be calculated (e.g. not enough data points or infinity in any of the input values), fill with infinity to indicate missing data.
The previous solution caused no statistical curves to be displayed due to wrong number of elements in the percentile vector.
Seen in regression test "ProjectFiles2024Tests\TestCase_Ensemble"
Introduced custom percentile handling for ensemble statistics:
- Added "CUSTOM" statistics type and percentile field to RifEclipseSummaryAddress
- Refactored RimEnsembleStatisticsCase to store and compute arbitrary percentiles
- RimEnsembleStatistics now parses custom percentiles from UI input
- RimEnsembleCurveSet and related interfaces updated to handle custom percentile curves and labels
- Hashing and statistics calculations now include custom percentiles for cache validity
Enhances the robustness of ensemble data processing by:
- Displaying a tooltip when data is missing for a realization.
- Checks if an enhanced summary file exists before attempting to open it.
- Refines the creation of the realization filter to accurately represent the range of available data.
* Moves multi reader initialization
* Log error message if not able to open file
* Make keywordCount pure virtual
The base class `RifSummaryReaderInterface` is changed to be abstract and the `keywordCount` is declared as a pure virtual method, therefore the derived classes must implement the method.
This will ensure that all readers are responsible for reporting a keyword count.
* Ensures addresses are created if required
Calls `createAddressesIfRequired` on the summary reader to ensure addresses are created before accessing keyword count.
This prevents potential issues when the addresses haven't been initialized yet.
Changes in the summary system has introduced a bug causing two identical results to be displayed, the correct statistical vector and the error. If error is asked for, return no values.
* Do not create ensemble in initAfterRead
* Check if filter is active before update
* Do not recompute statistics on every loadData
* Removes unnecessary resampling
* MSVC: Add support for dynamic deoptimalization
* Make sure RFT plots are initialized based on curves in project file
* Make sure the progress dialog is always on top
* Add more progress info when loading project file
* Make sure names in RFT plots are updated
Removes the dedicated `SUMMARY_ENSEMBLE_STATISTICS` category and instead uses a `statisticsType` field within `RifEclipseSummaryAddress` to represent ensemble statistics.
This change simplifies the code and avoids potential confusion by representing statistics as attributes of existing summary types, rather than a separate category. It also improves code robustness.
Make sure statistics curves are recomputed on flag change
When toggling "Discard Missing or Incomplete Realizations", make sure the statistics is recomputed and the ensemble plots updated.
Add hash functions
Add serial number to SummaryReaderInterface
Add hash for RimSummaryCase
Use hash based on input parameters to control caching of data in RimEnsembleStatisticsCase::calculate and RimEnsembleCurveSet::appendOptionItemsForSummaryAddresses
There can be incomplete realizations in an ensemble. Make sure that the calculation is robust for all variants of incomplete curves.
Use RiaTimeHistoryCurveMerger to ensure one definition of time steps and extraction of curve values to be used to compute ensemble statistics.
Remove unused function.
* Add statistics enum
* Support ensemble statistics curve as data source for regression curves
* Allow creation of regression curves from statistics curves
* Make sure regression curves are updated after source curves
* Add state to use full or user defined range for regression source data
* Add isRegressionCurve
* Make sure source stepping works when regression curves are present
* Support cross plot in plot templates
* Fix source stepping for cross plots
* Show cross plot data in "Show Plot Data"
* Use bin size and realization count when computing cross plot statistics
* Remove use of RifEnsembleStatisticsReader
* Add RimSummaryAddressSelector and RimEnsembleCrossPlotStatisticsCase
* Remove RifEnsembleStatisticsReader
* Add support for cross plot x-axis selection
* Support source stepping on ensemble cross plot curves
* Add isXAxisSummaryVector
* #9773 Refactor ensemble import dialog.
* #9773 Add support for ensemble import of StimPlan/Reveal summary data.
* #9773 Add method for finding min and max time steps.
* #9773 Add option for resampling to hour intervals.
* #9773 Add option for resampling to minute intervals.
* #9773 Ensemble statistics: determine sub-sampling period dynamically.
* Use RiaWeightedMeanCalculator for curve resampling.
* Add HOUR and MINUTE to date time app enum.
* #9773 Generate better ensembles for StimPlan summaries.
* #9773 Fix parameters.txt lookup for StimPlan summaries
* Refactor: Improve interface of RicImportSummaryCasesFreature::createSummaryCasesFromFiles
Co-authored-by: Magne Sjaastad <magne.sjaastad@ceetronsolutions.com>
* Update titles after dropping vectors into plot
* Do not replace _Name variables
* Make sure signals are wired when opening a project file
* Add auto name field and set default on
* Remove includes from header files
When computing statistics curves for an ensemble, we computed these values based on curves having identical same timestamp as the first summary case in the ensemble. If the simulator had numerical issues, this first time step can be slightly different from case to case.
Remove this test on first time step, but keep the optional test on last time step.
* Moved UnitSystem from RiaEclipseUnitTools to RiaDefines.
- Renamed UnitSystem to EclipseUnitSystem
- Replaced header includes and removed obsolete includes of RiaEclipseUnitTools.h
* Moved result name functions into separate file.
* Minor cleanup
Co-authored-by: rubenthoms <rubenthoms@users.noreply.github.com>
Co-authored-by: Magne Sjaastad <magne.sjaastad@ceetronsolutions.com>
Co-authored-by: magnesj <magnesj@users.noreply.github.com>
Magne Sjaastad
Kristian Bendiksen
magnesj
Gaute Lindkvist