v0.9.4

What's Changed

• Fix hardcoded dataloader detection in preview endpoints by @Irozuku in #601
• Feat/button consistency by @Creylay in #606
• Add ARFF dataset loader support by @Irozuku in #608
• Refactor converter row count metadata and add frontend warnings for row-changing converters by @Irozuku in #609
• Fix WebSocket URL Resolution for Live Metrics and Hardware Monitor by @Irozuku in #611
• Fix prediction preview serialization and manual input row initialization bugs by @Irozuku in #614
• fix: replace Paper with Accordion for model comparison panel by @Creylay in #610
• Add Portuguese language support for various components and tours by @Felipedino in #615
• fix: clear notebook state and update SaveDatasetModal style on dataset save by @Creylay in #616
• fix: implement search threshold for displaying search bars in components by @Creylay in #617
• Update benchmark documentation by @Irozuku in #618
• Reduce redundant dataset fetches and standardize notebook API endpoints by @Irozuku in #620
• Feat/fix and redesign onboarding by @Creylay in #621
• Reduce redundant HTTP requests in the notebooks/datasets UI by @Irozuku in #622
• Fix/generative session modal and ordering by @Creylay in #624
• Fix stale dataset cache and converter updates for Arrow-backed datasets by @Irozuku in #625
• Fix/light mode white colors and confirmation modal styles by @Creylay in #626
• fix: remove union_type from schema_field definitions by @Creylay in #627
• Fix/right panel padding and chip overflow by @Creylay in #628
• fix: Add step 0.1 to 0.1 in splits by @Felipedino in #636
• Migrate frontend theme to DashAI blue branding by @Irozuku in #639
• refactor: "Run all" doesnt expand show results in card models by @Felipedino in #630
• demo:refactor: round correlation matrix results and heatmap color scheme by @Felipedino in #640
• demo:feat: enhance explorers to support numeric categorical types by @Felipedino in #641
• Update DashAI branding with new SVG logo and UI consistency changes by @Irozuku in #642
• demo: adjust input field width for manual prediction by @Felipedino in #644
• Feat/integrate image classification by @Creylay in #612
• fix: stop infinite job queue polling in edge cases by @Felipedino in #629
• Fix optimizer parameter switching and Hyperopt trial parameter casting by @Irozuku in #631
• Feat/new model highlight by @Creylay in #632
• fix: adjust layout and spacing for model comparison header buttons by @Creylay in #634
• Fix/reduce requests by @Felipedino in #637
• refactor/Remove deprecated by @Felipedino in #638
• fix: enhance SaveDatasetModal with delete confirmation and scrollable… by @Creylay in #645
• Persist dataset row/column counts and reduce frontend dataset requests by @Irozuku in #619
• Fix: ItemsToDeleteList unreadable in light mode by @Creylay in #635
• Add delete confirmation modal to ModelComparisonTable and RunCard com… by @Creylay in #633
• Feat/explainer prediction redesign by @Creylay in #643
• Fix dataset name overridden by auto-generated value after file upload by @Creylay in #646
• Normalize and centralize spacing of components by @Felipedino in #613
• feat: Implement drag-and-drop functionality for models, converters and explorers by @Felipedino in #647
• Fix pagination reset on column selection by @Creylay in #648
• Fix correct image preview filename for Skewed Chi2 Sampler by @Creylay in #649
• Fix dataset export filename by @Creylay in #650
• feat: add German (Deutsch) language support by @nicolasnast-png in #651
• Add Hub for browsing and importing datasets from external sources by @Irozuku in #652
• Clear stale localStorage state when deleting or recreating a run by @Creylay in #654
• Add native type extraction support to BaseDataLoader and ARFF by @Irozuku in #655
• Add image classification models by @Creylay in #653
• Fix/hpo with one-hot encoder and score system by @Felipedino in #656
• Fix type change validation for hub import datasets without local files by @Irozuku in #657
• Improve performance and usability of analysis tabs and insights by @Irozuku in #658
• Include translation function in model fetching dependencies by @Creylay in #659

New Contributors

• @nicolasnast-png made their first contribution in #651

Full Changelog: v0.9.3...v0.9.4

v0.9.3

What's Changed

• Feat/add models to all tasks by @Felipedino in #586
• Standardize panel borders across modules using theme divider by @Irozuku in #587
• fix text score profiles by @Felipedino in #590
• Fix/grid layout by @Creylay in #591
• Standardize module workflows with a unified step based UI and reusable component selection by @Irozuku in #592
• refactor: update icon sizes and improve theme toggle accessibility by @Creylay in #593
• style: redesign bar chart hover highlighting in dataset tabs by @Creylay in #594
• Fix broken anchor links by converting docs to absolute paths by @Irozuku in #595
• Add Spanish translations for build and plugin development documentation by @Irozuku in #596
• Fix i18n reactivity in RightBar (converters and explorers not updating on language change) by @Irozuku in #597
• Fix column selector for Explorers and Converters. by @Irozuku in #598
• fix: show step numbers instead of check icons in steppers by @Creylay in #583
• Refactor typography variants across components for consistency by @Felipedino in #599
• Remove duplicate step text in converter/explorer modal by @Felipedino in #600
• Fixed async form loading by @cristian-tamblay in #602
• Remove forced reinstall of llama-cpp-python in publish workflow by @Irozuku in #604
• Remove ARM force reinstall of llama-cpp-python by @Irozuku in #605

Full Changelog: v0.9.2...v0.9.3

v0.9.2

What's Changed

• Refactor app navigation with route based workflows and stepper support by @Irozuku in #582
• fix: improve alert visibility and styling in dataset components by @Creylay in #584
• fix: include pywebview in dependencies and update macOS build configurations by @Creylay in #588

Full Changelog: v0.9.1...v0.9.2

v0.9.1

What's Changed

• Updated pypi version to pass new CI/CD by @cristian-tamblay in #275
• Removed llama-cpp from requirements by @cristian-tamblay in #279
• Notebooks module by @cristian-tamblay in #289
• Minor fix by @cristian-tamblay in #291
• Fix scroll in dataset summary table by @cristian-tamblay in #294
• Fixed converters by @cristian-tamblay in #299
• DashAI 0.3.0 Release by @cristian-tamblay in #382
• Added download for the frontend artifact so the upload works by @cristian-tamblay in #383
• DashAI 0.9.0 by @cristian-tamblay in #579
• Minor fixes for production publish actions by @cristian-tamblay in #580
• Add comprehensive platform benchmark and improve documentation layout by @Irozuku in #581

Full Changelog: v0.9.0...v0.9.1

v0.9.0

What's Changed

• Pre-Release 0.1.5 by @cristian-tamblay in #259
• Fix/torch cuda by @Felipedino in #260
• Revert requirements changes from 6eff065, keeping Distilbert updates by @cristian-tamblay in #262
• Change installation instructions and enchance explorer plots by @Felipedino in #261
• Project/pipelines by @CarlaHayler in #267
• Project/generative plugin by @Irozuku in #255
• Improvement dataloaders and added plugins in the documentation by @Felipedino in #263
• fix: refactor import statements in ConfigureExplorersStep and NodeSid… by @Irozuku in #270
• feat/pipelines by @CarlaHayler in #268
• DashAI 0.2.0 by @cristian-tamblay in #272
• Optimize dataset info retrieval by @Irozuku in #273
• Fix/nodes file by @cristian-tamblay in #276
• Project/generative plugin llm fix by @cristian-tamblay in #277
• Removed llama-cpp from base installation, added pip install dashai[fu… by @cristian-tamblay in https://github.com//pull/278
• Fix bug when changing from string to null in forms by @Irozuku in #281
• fix: update device handling in OpusMtEnESTransformer by @Irozuku in #284
• fix: add Hugging Face API key to StableDiffusionV3 parameters and remove deprecated model option by @Irozuku in #283
• Replace explorations with notebook-centric API and redesign dataset module by @Irozuku in #271
• Remove target column selector for unsupervised converters by @Irozuku in #285
• fix: refactor get_dataset_file to obtain total rows from metadata by @Irozuku in #287
• feat: delete converters and explorers by @Irozuku in #286
• fix: refactor DashAIDataset creation for imbalanced converter by @Irozuku in #290
• Feat/ux dataset by @Irozuku in #288
• fix: adjust height and overflow properties in DatasetPreviewNotebook by @Irozuku in #293
• Reduce memory usage and load time by skipping costly fingerprint generation by @Irozuku in #296
• Add types to dataset table by @Felipedino in #292
• fix: increase height of dataset preview box for better visibility by @Irozuku in #298
• Fix label encoder and add default value to simpleimputer by @Felipedino in #297
• Refactor dataset creation flow by @Creylay in #295
• Remove unnecessary steps in explorer/converter modals by @Irozuku in #300
• Default names in forms by @Creylay in #306
• Fix dataset name duplication crash and improve validation error handling by @Creylay in #307
• Add nan remover converter by @Irozuku in #304
• Feat/dataset experiment chaining by @Irozuku in #301
• Unify notebook dataset creation under DatasetJob workflow by @Irozuku in #305
• Display dataset name instead of ID in experiments table by @Creylay in #308
• Add action icons for Predictions and Explainability in Experiments module by @Irozuku in #309
• Fix LLVMPY issue by @Riul1999 in #312
• Fix trained model view when using it for predictions by @Irozuku in #319
• Refactor NewSessionButton into generic NewItemButton and standardize button styling across dataset visualization by @Irozuku in #316
• Add dataset split and percentage selection for local explanations, update KernelShap sampling to use fractions by @Irozuku in #313
• Add NaN detection and user warnings for experiment creation by @Irozuku in #310
• Improve UI with Standardized Explorer/Converter Display Names by @Irozuku in #315
• Refactor prediction dataset filtering to allow datasets without target column by @Irozuku in #320
• Feat/update mui resize by @Creylay in #317
• Feat/huey job queue by @cristian-tamblay in #269
• Implement polling for message status updates in Generative module by @Irozuku in #318
• fix: remove default column selection in ColumnSelector by @Creylay in #323
• Fix prediction target encoded by @Irozuku in #322
• Refactor ColumnSelector to improve UI feedback for Explorers by @Irozuku in #325
• Add cache-control headers to front_api by @cristian-tamblay in #324
• Enhance explorer and converter filtering to include short description by @Irozuku in #326
• Update db migration configuration by @Creylay in #303
• Revert "Update db migration configuration" by @cristian-tamblay in #329
• Enhanced CI/CD on PR by @cristian-tamblay in #327
• Feat: prevent modal close by @Creylay in #330
• Refactor: replace valueFormatter with valueGetter for date formatting by @Creylay in #331
• Feat: differentiate notebook view by @Creylay in #332
• Replace dynamic imports with component registry lookup in converter job class by @Irozuku in #333
• Ensure proper initialization of superclass in TruncatedSVD by @Irozuku in #334
• Fix default values in RBFSamplerSchema by @Irozuku in #336
• Interactive Dataloader (back) by @Riul1999 in #321
• Metadata explorer converter by @Irozuku in #335
• Enhance experiment configuration UI with improved metric selection and clearer optimizer handling by @Irozuku in #337
• Set default placeholder for max_features in DecisionTreeClassifier schema by @Irozuku in #338
• Conditionally render delete button in ConverterBox and ExplorerBox component by @Irozuku in #339
• Update PermutationFeatureImportance to select samples using fraction by @Irozuku in #341
• Fix excel dataloader by @Irozuku in #343
• Ensure hover effect for NewItemButton background color by @Irozuku in #344
• Fix column replacement logic in converter job by @Irozuku in #345
• Refactor model job by @Irozuku in #347
• Add resizable functionality to ConfigureToolModal by @Irozuku in #340
• Fix converters by @Irozuku in #346
• Fix regression models by @Irozuku in #348
• Add max_features field with support for percentage and None type for DecisionTreeClassifier by @Irozuku in #349
• Update dependencies in RightBar to include explorersAndConverters by @Irozuku in #351
• Move run finished status update to the end of the job by @Irozuku in #352
• Enhance parameter type handling in OptunaOptimizer by @Irozuku in #354
• Add display name and color attributes to components by @Irozuku in #356
• Add Bag-of-Words converter by @Irozuku in #358
• Fix split experiment by @Irozuku in #357
• Add Tokenizer converter by @Irozuku in #359
• Add TF-IDF converter by @Irozuku in #360
• Fix bow model by @Irozuku in #363
• Fix/dataset table refresh by @Creylay in #361
• Improve dataset upload error by @Irozuku in #37...

DashAI 0.3.0

What's Changed

• Pre-Release 0.1.5 by @cristian-tamblay in #259
• Fix/torch cuda by @Felipedino in #260
• Revert requirements changes from 6eff065, keeping Distilbert updates by @cristian-tamblay in #262
• Change installation instructions and enchance explorer plots by @Felipedino in #261
• Project/pipelines by @CarlaHayler in #267
• Project/generative plugin by @Irozuku in #255
• Improvement dataloaders and added plugins in the documentation by @Felipedino in #263
• fix: refactor import statements in ConfigureExplorersStep and NodeSid… by @Irozuku in #270
• feat/pipelines by @CarlaHayler in #268
• DashAI 0.2.0 by @cristian-tamblay in #272
• Updated pypi version to pass new CI/CD by @cristian-tamblay in #275
• Optimize dataset info retrieval by @Irozuku in #273
• Fix/nodes file by @cristian-tamblay in #276
• Project/generative plugin llm fix by @cristian-tamblay in #277
• Removed llama-cpp from base installation, added pip install dashai[fu… by @cristian-tamblay in https://github.com//pull/278
• Removed llama-cpp from requirements by @cristian-tamblay in #279
• Fix bug when changing from string to null in forms by @Irozuku in #281
• fix: update device handling in OpusMtEnESTransformer by @Irozuku in #284
• fix: add Hugging Face API key to StableDiffusionV3 parameters and remove deprecated model option by @Irozuku in #283
• Replace explorations with notebook-centric API and redesign dataset module by @Irozuku in #271
• Remove target column selector for unsupervised converters by @Irozuku in #285
• fix: refactor get_dataset_file to obtain total rows from metadata by @Irozuku in #287
• feat: delete converters and explorers by @Irozuku in #286
• Notebooks module by @cristian-tamblay in #289
• fix: refactor DashAIDataset creation for imbalanced converter by @Irozuku in #290
• Feat/ux dataset by @Irozuku in #288
• Minor fix by @cristian-tamblay in #291
• fix: adjust height and overflow properties in DatasetPreviewNotebook by @Irozuku in #293
• Fix scroll in dataset summary table by @cristian-tamblay in #294
• Reduce memory usage and load time by skipping costly fingerprint generation by @Irozuku in #296
• Add types to dataset table by @Felipedino in #292
• fix: increase height of dataset preview box for better visibility by @Irozuku in #298
• Fix label encoder and add default value to simpleimputer by @Felipedino in #297
• Fixed converters by @cristian-tamblay in #299
• Refactor dataset creation flow by @Creylay in #295
• Remove unnecessary steps in explorer/converter modals by @Irozuku in #300
• Default names in forms by @Creylay in #306
• Fix dataset name duplication crash and improve validation error handling by @Creylay in #307
• Add nan remover converter by @Irozuku in #304
• Feat/dataset experiment chaining by @Irozuku in #301
• Unify notebook dataset creation under DatasetJob workflow by @Irozuku in #305
• Display dataset name instead of ID in experiments table by @Creylay in #308
• Add action icons for Predictions and Explainability in Experiments module by @Irozuku in #309
• Fix LLVMPY issue by @Riul1999 in #312
• Fix trained model view when using it for predictions by @Irozuku in #319
• Refactor NewSessionButton into generic NewItemButton and standardize button styling across dataset visualization by @Irozuku in #316
• Add dataset split and percentage selection for local explanations, update KernelShap sampling to use fractions by @Irozuku in #313
• Add NaN detection and user warnings for experiment creation by @Irozuku in #310
• Improve UI with Standardized Explorer/Converter Display Names by @Irozuku in #315
• Refactor prediction dataset filtering to allow datasets without target column by @Irozuku in #320
• Feat/update mui resize by @Creylay in #317
• Feat/huey job queue by @cristian-tamblay in #269
• Implement polling for message status updates in Generative module by @Irozuku in #318
• fix: remove default column selection in ColumnSelector by @Creylay in #323
• Fix prediction target encoded by @Irozuku in #322
• Refactor ColumnSelector to improve UI feedback for Explorers by @Irozuku in #325
• Add cache-control headers to front_api by @cristian-tamblay in #324
• Enhance explorer and converter filtering to include short description by @Irozuku in #326
• Update db migration configuration by @Creylay in #303
• Revert "Update db migration configuration" by @cristian-tamblay in #329
• Enhanced CI/CD on PR by @cristian-tamblay in #327
• Feat: prevent modal close by @Creylay in #330
• Refactor: replace valueFormatter with valueGetter for date formatting by @Creylay in #331
• Feat: differentiate notebook view by @Creylay in #332
• Replace dynamic imports with component registry lookup in converter job class by @Irozuku in #333
• Ensure proper initialization of superclass in TruncatedSVD by @Irozuku in #334
• Fix default values in RBFSamplerSchema by @Irozuku in #336
• Interactive Dataloader (back) by @Riul1999 in #321
• Metadata explorer converter by @Irozuku in #335
• Enhance experiment configuration UI with improved metric selection and clearer optimizer handling by @Irozuku in #337
• Set default placeholder for max_features in DecisionTreeClassifier schema by @Irozuku in #338
• Conditionally render delete button in ConverterBox and ExplorerBox component by @Irozuku in #339
• Update PermutationFeatureImportance to select samples using fraction by @Irozuku in #341
• Fix excel dataloader by @Irozuku in #343
• Ensure hover effect for NewItemButton background color by @Irozuku in #344
• Fix column replacement logic in converter job by @Irozuku in #345
• Refactor model job by @Irozuku in #347
• Add resizable functionality to ConfigureToolModal by @Irozuku in #340
• Fix converters by @Irozuku in #346
• Fix regression models by @Irozuku in #348
• Add max_features field with support for percentage and None type for DecisionTreeClassifier by @Irozuku in #349
• Update dependencies in RightBar to include explorersAndConverters by @Irozuku in #351
• Move run finished status update to the end of the job by @Irozuku in #352
• Enhance parameter type handling in OptunaOptimizer by @Irozuku in #354
• Add display name and color attributes to component...

v0.1.5

0.1.5

On this release, we aim at distribution our software with more implemented capabilities, Hyper Parameter Optimization, Explainability, Converters and Explorers.

Tabular classification task and Regression Tasks are fully functional.
Translation and Text classification are unstable.
Image classification is implemented as a plugin but untested.

To test the software, we highly recommend using our tested example datasets attached on this pre-release.

DashAI Pre-release

0.0.14

On this release, we aim at distribution our software with the most basic capabilities.

• Tabular classification task is fully functional.
• Translation and Text classification are unstable.
• Image classification is not yet implemented.

To test the software, we highly recommend using our tested example datasets attached on this pre-release.