Fix coclustering and isprint instabilities #124
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
marcboulle
force-pushed
the
122-fix-datagrid-instability
branch
8 times, most recently
from
6029f4f to
90925b8
Compare
marcboulle
force-pushed
the
122-fix-datagrid-instability
branch
from
f99bf16 to
0cc0940
Compare
Choix en dur du repertoire de lancement pour lancer un session de debug depuis Visual C++ 2022 - impact dans MODL.cpp et MODL_Coclustering.cpp
Implementation windows - nettoyage uniquement si, pointeur hFinf non NUL - correction a reporter en V11
Portability.h
- p_isprint: mise en placxe d'une implementation portable
- impacts sur tous les isprint existant:
- KWCLex.lex
- KWDatabaseFormatDetector::ComputeSeparatorPriority
- KWDREncrypt::InitWorkingArrays
- KWTest
Test sur la stabilisation effectives des resultats sur Windows, Linux, Mac
- TestKhiops/Rules/EncryptRules
- TestKhiops/Bugs/DicoSpecialChars
Correction a reporter en V11
marcboulle
force-pushed
the
122-fix-datagrid-instability
branch
from
0cc0940 to
dfd3c81
Compare
marcboulle
changed the title
folmos-at-orange
requested changes
Jan 17, 2024
test/LearningTest/TestCoclustering/Standard-unstable/Adult2varsTiny/results.ref/time.log
Outdated
Show resolved
Hide resolved
test/LearningTest/TestKhiops/Standard-unstable/IrisU/results.ref/time.log
Outdated
Show resolved
Hide resolved
marcboulle
force-pushed
the
122-fix-datagrid-instability
branch
from
dfd3c81 to
48c251d
Compare
Correction dans les algorithmes d'optimisation: - partout, tester l'amelioration du cout en remplacant ~if (dCost < dBestCost)~ par ~if (dCost < dBestCost - dEpsilon)~ - correction principale dans KWDataGridOptimizer::OptimizeDataGrid - impacts pour propager la correction partout - DTDiscretizerMODL::DiscretizeNEW - DTDiscretizerMODL::DiscretizeOLD - DTDiscretizerMODL::DiscretizeGranularizedFrequencyTableNEW (pour le nul cost: < dBestCost + dEpsilon) - DTGrouperMODL::GroupPreprocessedTable - DTGrouperMODL::SmallSourceNumberGroup - KWDiscretizerMODL::Discretize - KWGrouperMODL::GroupPreprocessedTable - KWGrouperMODL::SmallSourceNumberGroup - KWDensityEstimationTest::SearchBestInstanceGridSize - MHDiscretizerHistogramMODL::GranularizedDiscretizeValues - MHDiscretizerHistogramMODL_fp::OptimizeGranularity - MHFloatingPointFrequencyTableBuilder::InitializeDomainBounds Stabilisation lie au probleme de choix d'une partition aleatoire, base sur un Shuffle, puis un Sort - le sort est instable entre Windows et Linux en cas d'egalite du critere - correction en memoirsant un index aleatoire suite au Shuffle, puis en utilisant cet index en critere de tri secondaire - KWDataGridManager::SortAttributeParts: reimplementation du tri avec random index en cas d'egalite - KWSortableSymbolCompareValue: utilisation de l'index comme critere de tri secondaire Teste sur les 12 jeux donnees instables entre Windows, Linux, Mac - quelques jeux de tests de reference ont change - les resultats sont maintenant identique sur les trois OS Teste sur LearningTest entier Memorisation des jeux de test pour les test de non regression sur git - test\LearningTest\TestKhiops\Standard\IrisU - deplace depuis Standard-unstable, desormais supprime - test\LearningTest\TestCoclustering\Standard\Adult2varsTiny - deplace depuis Standard-unstable, desormais supprime Correction a reporter en V11
marcboulle
force-pushed
the
122-fix-datagrid-instability
branch
from
48c251d to
133ec96
Compare
folmos-at-orange
approved these changes
Jan 17, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Quatre commits, avec corrections principales: