5c97c9b | Dennis Bader | 04 April 2024, 14:09:31 UTC | Refactor/metrics (#2284) | 04 April 2024, 14:09:31 UTC |
91c7087 | Alicja Krzeminska-Sciga | 16 March 2024, 11:05:51 UTC | Add optional inverse transform in historical forecast (#2267) * Add optional inverse transform in historical forecast * Update variables names and docstrings * Move the inverse transform to InvertibleDataTransformer * Fix single element list * Update docstrings * Move the inverse transform of list of lists to inverse_transform method * make invertible transformers act on list of lists of series * add tests * update changelog --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 16 March 2024, 11:05:51 UTC |
d764bc4 | Dennis Bader | 15 March 2024, 10:55:34 UTC | fix type hinting for _with_sanity_checks (#2286) * fix type hinting for _with_sanity_checks * update changelog | 15 March 2024, 10:55:34 UTC |
7986348 | Felix Divo | 12 March 2024, 08:13:00 UTC | Add `ForecastingModel.supports_probabilistic_prediction` (#2259) (#2269) * Remove unnessesary `pass` statements * Rename ForecastingModel_is_probabilistic to supports_probabilistic_prediction, rearrange some documentation * Remove redundant overrides * Reformat * Add CHANGELOG entry --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 12 March 2024, 08:13:00 UTC |
2264cca | Dennis Bader | 11 March 2024, 08:17:52 UTC | Repo/update code owners (#2275) * update code owners * udpated PR template | 11 March 2024, 08:17:52 UTC |
4de0b68 | dennisbader | 05 March 2024, 10:03:57 UTC | Release 0.28.0 | 05 March 2024, 10:03:57 UTC |
c3d79ba | Dennis Bader | 05 March 2024, 09:02:54 UTC | Release 0.28.0 (#2268) * update changelog * bump u8darts 0.27.2 to 0.28.0 * update changelog | 05 March 2024, 09:02:54 UTC |
4744835 | Dennis Bader | 04 March 2024, 18:15:59 UTC | fix torch baseline model import (#2266) | 04 March 2024, 18:15:59 UTC |
3803cbe | Dennis Bader | 04 March 2024, 18:07:44 UTC | Feat/global naive models (#2261) | 04 March 2024, 18:07:44 UTC |
4de71e4 | Dennis Bader | 04 March 2024, 10:37:31 UTC | fix failing dataset unit tests for py38 (#2263) | 04 March 2024, 10:37:31 UTC |
2117bf3 | madtoinou | 02 March 2024, 12:59:51 UTC | Fix: estimator getter and lagged_label_name (#2246) * feat: adding docstring and check to get_multioutput_estimator * fix: added lowbound check * fix: update docstring, indexing account for multi_models param * feat: added corresponding test * feat: added tests for estimator getter * feat: store and expose the lagged label names (for each model estimator) * fix: rephrasing docstring * update changelog * fix: linting * fix: replaced ocl with hrz in naming of the lagged label * fix: update error messages * feat: simplify test, overfit XGB on only one training example * feat: added a method to get estimator for models supporting multi-output natively * feat: added corresponding test * update changelog * fix: linting * Update CHANGELOG.md --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 02 March 2024, 12:59:51 UTC |
62d9200 | madtoinou | 01 March 2024, 16:06:07 UTC | fix: datetime_attribute account for 0 or 1-indexing of the attributes (#2242) * fix: datetime_attribute account for 0 or 1-indexing of the attributes * feat: 1-indexed date attribute are shifted to enforce 0-indexing for all the generated encodings * updated changelog * fix: remove commented lines * fix: typo in comment * make ONE_INDEXED_FREQS a constant * fix: simplified test by using year 2001 * feat: better handling of years with 53 weeks or 366 days * fix: properly take the index length when adding the extra week * fix: simplifying test * fix: update tests to account for the forced 0-indexing of the datetime attributes encoding * fix: passing lmbda parameter as BoxCox doesn't converge when encodings contains a 0 --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 01 March 2024, 16:06:07 UTC |
ccd0d42 | Dennis Bader | 29 February 2024, 14:02:13 UTC | Feat/shifted output (#2176) * add output chunk shift to lightning modeuls * training torch model with shifted output * first shifted output inference works for mixed covariates models * full covariateds support for shifted mixed covariates dataset * add shift support to all torch models * update torch model extreme lags with shift * update torch model encoder settings with shift * update torch model encoder settings with shift * add unit test for shifted torch mmodel with encoders * add unit tests for tft model * add unit tests for all torch models * update output_chunk_shift description * apply suggestions from PR review * add output chunk shift to extreme lags * udpate historical forecasts to work with shifted output * update historical forecasts start description for shifted output * apply suggestions from PR review * prepare regression models for output chunk shift * fix failing unit tests * prepare regression models for output chunk shift part 2 * update hist fc for regression models with output shift * update tabularization * add test for comparing results between output shift and normal multi models * historical forecasts for shifted regression models * update tabularization training tests * update tabulirazion get feature times tests * update tabularization get shared times tests * update tabularization get shared bounds tests * update tabularization get lagged prediction data tests * add tests for tabularization without target lags but only covariate lags * update n_steps_between docs * update changelog * add unit tests for inference datasets * add unit tests for sequential training datasts * update changelog * make ocs property non optional * skip output_chunk_shift checks when loading weights since not relevant for parameter shape * apply suggestions from PR review | 29 February 2024, 14:02:13 UTC |
b9e6d8b | Thomas Kientz | 26 February 2024, 14:21:57 UTC | Change default `gridsearch` kwarg value (#2243) * Change default kwarg * Update CHANGELOG.md --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 26 February 2024, 14:21:57 UTC |
ec53511 | madtoinou | 26 February 2024, 09:02:38 UTC | Fix: Using gridsearch with use_fitted_values=True raises unexpected error (#2222) * fix: arguments must be provided to model cls in order to check presence of the fitted_values attribute * fix: added a check that parameters is indeed a dict * updated changelog * fix: update test to pass the new sanity checks * fix: addressing review comments --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 26 February 2024, 09:02:38 UTC |
2d1919d | Dennis Bader | 24 February 2024, 21:21:12 UTC | Dep/relax ptl (#2251) * remove pytorch lightning upper version cap * fix failing unit test and update changelog | 24 February 2024, 21:21:12 UTC |
2bc6319 | Dennis Bader | 24 February 2024, 16:52:16 UTC | fix ElectricityConsumptionZurich dataset hash (#2250) | 24 February 2024, 16:52:16 UTC |
24ae0e1 | madtoinou | 24 February 2024, 12:37:30 UTC | fix: update hierarchy for single transform window_transform (#2207) * fix: update hierarchy for single transform window_transform * update changelog * update changelog * fix: using set to check overlap * fix: corrected logic to update the hierarchy after window_transform * fix: hierarchy can be conserved when applying non-overlapping transforms * feat: add new argument, improve logic * feat: adding tests * fix: expected argument match docstring in resample() * fix: addressing review comments * fix: linting issue * fix: linting * linting * update changelog and remane keep_old_names to keep_names --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 24 February 2024, 12:37:30 UTC |
bf51476 | madtoinou | 24 February 2024, 12:36:12 UTC | Fix/ts prepend (#2237) * fix: append/prepend correctul retain components names and hierarchy * updated changelog * fix: revert unecessary change * update changelog --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 24 February 2024, 12:36:12 UTC |
6fbb670 | Dennis Bader | 24 February 2024, 12:07:13 UTC | Reformat / lint repository with new dev dependency versions (#2248) * update dev requirements with new pre commit hook lint dependency versions * black reformatting * fix flake8 checks | 24 February 2024, 12:07:13 UTC |
4600453 | Marc Bresson | 16 February 2024, 14:05:03 UTC | Pre commit hooks upgrade (#2228) | 16 February 2024, 14:05:03 UTC |
9e43e3c | Marc Bresson | 16 February 2024, 14:04:01 UTC | Improve description for ARIMA parameters (p, q, seasonal_orders and trend) (#2142) | 16 February 2024, 14:04:01 UTC |
8073de4 | Dennis Bader | 08 February 2024, 22:29:47 UTC | add support for more lr scheduler config parameters to torch models (#2218) * add support for more lr scheduler config parameters to torch models * update changelog | 08 February 2024, 22:29:47 UTC |
5b05d2b | madtoinou | 05 February 2024, 13:26:36 UTC | Fix/RegressionEnsemble with single model regressor and coef access in LinearRegressionModel (#2205) * fix: overwrite the self.model attribute with the model container * fix: prevent creation of RegressionEnsemble with a regression model created with multi_models=False * update changelog * Update darts/models/forecasting/regression_ensemble_model.py Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * rephrasing changelog * fix: enforce multi_models=True when ocl=1 --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 05 February 2024, 13:26:36 UTC |
1d7d854 | madtoinou | 05 February 2024, 12:42:55 UTC | fix: time index intersection for coefficient of variation (#2202) * fix: properly take the intersected time indexes for the coefficient of variation * fix: computing rmse on ndarray directly * fix: forgot sqrt for rmse in coef of variation * fix: update type of return in docstring, taking into consideration the multi_ts and multivariate decorator, which convert arrays into list * update changelog * update changelog --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 05 February 2024, 12:42:55 UTC |
8cb04f6 | Dennis Bader | 02 February 2024, 10:12:29 UTC | Feat/improve timeseries (#2196) * found major peformance boost for time series creation * first boosted time series version * improve slicing with integers * improve slicing with time stamps * improve slicing with time stamps * update from_xarray * improve from_group_dataframe() * remove test time series * remove old time series * add option to drop group columns from from_group_dataframe * update changelog * apply suggestions from PR review | 02 February 2024, 10:12:29 UTC |
0b4dcf0 | Dennis Bader | 29 January 2024, 15:12:28 UTC | update readme (#2197) | 29 January 2024, 15:12:28 UTC |
c52fe32 | Jirka Borovec | 29 January 2024, 15:12:04 UTC | move pytest config to `pyproject.toml` (#2191) | 29 January 2024, 15:12:04 UTC |
20ee5ec | dennisbader | 21 January 2024, 16:05:18 UTC | Release 0.27.2 | 21 January 2024, 16:05:18 UTC |
3ee3efd | Dennis Bader | 21 January 2024, 14:23:58 UTC | bump u8darts 0.27.1 to 0.27.2 and update changelog (#2179) | 21 January 2024, 14:23:58 UTC |
0520c43 | Dennis Bader | 21 January 2024, 14:11:39 UTC | fix wrong pandas freqstr for py38 (#2178) | 21 January 2024, 14:11:39 UTC |
9cb1a56 | Dennis Bader | 21 January 2024, 12:58:50 UTC | fixes for pandas >= 2.2.0 (#2177) | 21 January 2024, 12:58:50 UTC |
ccecc8a | dependabot[bot] | 19 January 2024, 22:17:11 UTC | Bump jupyterlab from 4.0.3 to 4.0.11 in /requirements (#2173) | 19 January 2024, 22:17:11 UTC |
315cb6f | Dennis Bader | 19 January 2024, 14:29:38 UTC | fix torch covariates support tables in user guide (#2172) | 19 January 2024, 14:29:38 UTC |
8b77a69 | Dennis Bader | 19 January 2024, 11:40:37 UTC | Feat/ccf (#2122) * added statistics function plot_ccf * add unit test * update changelog * add reference to ccf | 19 January 2024, 11:40:37 UTC |
7fe7128 | Felix Divo | 19 January 2024, 11:33:00 UTC | Update .gitignore to ignore intermediate coverage files (#2158) * Update .gitignore to ignore intermediate coverage files Ignores files such as the following, which get created by coverage measurement tools: .coverage.68a13e4de0d0.43514.XMNYGdrx .coverage.68a13e4de0d0.43515.XjbvzNqx * Update .gitignore Get rid of one line --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 19 January 2024, 11:33:00 UTC |
68f72a7 | Dennis Bader | 19 January 2024, 11:32:12 UTC | fix deprecated method import from sklearn==1.4.0 (#2170) | 19 January 2024, 11:32:12 UTC |
962fd78 | DavidKleindienst | 15 January 2024, 16:06:32 UTC | Fix: Improve TimeSeries.__getitem__ frequency inference (#2152) * Improve TimeSeries.__getitem__ frequency inference * adapt getitem handling * update changelog --------- Co-authored-by: David Kleindienst <kleindienst@ximes.com> Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 15 January 2024, 16:06:32 UTC |
de4afd1 | Dennis Bader | 14 January 2024, 12:33:43 UTC | sphinx sphinx release workflow issue (#2165) | 14 January 2024, 12:33:43 UTC |
cb724d1 | FourierMourier | 13 January 2024, 11:58:42 UTC | Fix: removed input re-normalization by rin inside `io_processor` (#2160) * prevented input re-normalization by rin using .clone() inside `io_processor` * Update CHANGELOG.md --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 13 January 2024, 11:58:42 UTC |
ea79679 | dependabot[bot] | 12 January 2024, 14:31:05 UTC | Bump jinja2 from 3.0.3 to 3.1.3 in /requirements (#2155) Bumps [jinja2](https://github.com/pallets/jinja) from 3.0.3 to 3.1.3. - [Release notes](https://github.com/pallets/jinja/releases) - [Changelog](https://github.com/pallets/jinja/blob/main/CHANGES.rst) - [Commits](https://github.com/pallets/jinja/compare/3.0.3...3.1.3) --- updated-dependencies: - dependency-name: jinja2 dependency-type: direct:production ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> | 12 January 2024, 14:31:05 UTC |
6082f62 | Dennis Bader | 29 December 2023, 15:02:53 UTC | upper cap pytorch lightning for ckpt symlink fix (#2136) | 29 December 2023, 15:02:53 UTC |
ab3d77a | ryfactor | 21 December 2023, 09:38:00 UTC | Fix typo in examples Ensemble Model (#2127) | 21 December 2023, 09:38:00 UTC |
4362df2 | dennisbader | 10 December 2023, 17:43:39 UTC | Release 0.27.1 | 10 December 2023, 17:43:39 UTC |
acbd488 | Dennis Bader | 10 December 2023, 16:13:05 UTC | bump u8darts 0.27.0 and update changelog (#2116) | 10 December 2023, 16:13:05 UTC |
8897e93 | Francisco Merino-Casallo | 10 December 2023, 16:02:00 UTC | fix TimeSeries.pd_series not retaining original name (#2102) * fix TimeSeries.pd_series not retaining original name * Improve access to TimeSeries's original name By directly accessing to the `components.array` attribute from the TimeSeries instance, I do not have to convert the `components` attribute to a list anymore. * Apply suggestions from code review As suggested by @dennisbader, we can directly access the items in the `components` attribute from the TimeSeries instance, no need to use the `components.array` attribute. Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * Reformat code to pass lint check * fix failing unit test --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 10 December 2023, 16:02:00 UTC |
70752ce | Dennis Bader | 10 December 2023, 15:37:06 UTC | exclude PLForecastingModule docs from custom rnn and blockrnn modules (#2115) | 10 December 2023, 15:37:06 UTC |
4170093 | Dennis Bader | 10 December 2023, 12:45:31 UTC | fix custom module for RNNModel and add tests (#2088) * fix custom module for RNNModel and add tests * update changelog * update changelog * make custom rnn module * update rnn docs * add custom block rnn module * update changelog * fix docs for BlockRNNModel | 10 December 2023, 12:45:31 UTC |
b68f64d | Dennis Bader | 01 December 2023, 12:30:59 UTC | Fix/hist fc predict kwargs (#2103) * add explicit predict params to torch models * avoid ignoring fit/predict_kwargs when args or kwargs are in model method signature * refactor fit/predict wrappers * update changelog | 01 December 2023, 12:30:59 UTC |
b20a1f2 | eschibli | 01 December 2023, 10:31:53 UTC | Bump version of pytorch-lightning to 2.0.0 (#2087) * Bump version of pytorch-lightning to 2.0.0 * relax PL lower version and adapt progress bar import instead * update changelog * revert some changes to changelog --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 01 December 2023, 10:31:53 UTC |
7cfdf62 | madtoinou | 29 November 2023, 08:18:47 UTC | Fix/update doc (#2091) * fix: link to models table * fix: VARIMA supports probabilistic forecast * fix: naive model support multivariate series * fix: removed link to unexisting section of the install.md, removed typo * fix: use split_after to make example more coherent * fix: adressing review comment, link to static covariate notebook and model table * fix: more specific comment in example notebook | 29 November 2023, 08:18:47 UTC |
b29317d | Ondřej Holub | 27 November 2023, 13:44:43 UTC | Fix formula in single-shot forecast example (#2095) | 27 November 2023, 13:44:43 UTC |
dd08413 | Felix Divo | 22 November 2023, 10:54:31 UTC | Link docker profile in README.md (#2079) | 22 November 2023, 10:54:31 UTC |
603b57d | dennisbader | 18 November 2023, 18:16:53 UTC | Release 0.27.0 | 18 November 2023, 18:16:53 UTC |
679b4d0 | Dennis Bader | 18 November 2023, 16:01:48 UTC | Release 0.27.0 (#2073) * bump u8darts 0.26.0 * update changelog * update unreleased section | 18 November 2023, 16:01:48 UTC |
d153c95 | madtoinou | 18 November 2023, 15:20:58 UTC | Feat/MIDAS transformer (#1820) * MIDASTransformer now outputs a low sample variant of the high sample series through 'ts_transform()' * MIDASTransformer now outputs a low sample variant of the high sample series through 'ts_transform()' * extracted '_create_midas_df' from 'ts_transform' * extracted '_create_midas_df' from 'ts_transform' * Added some comments to helper functions * changed some variable names * changed some variable names * added warning if target frequency and input frequency don't match up like they should in case of a MIDAS transformation. * add _transform_iterator like the one in window_transformer * docstring finished, including example * added comments and description to 'ts_transform()' * more robust up and downsampling in order to get for example 3 months for every quarter instead of there being some missing months, also changed the docstring a bit such that it works with the debugger * tests are coming along, but still giving some errors * tests work * small comment change * adapted to 'params['fixed']['variable_name']' way of dealing with args * feat: attemp to make midas invertible * feat: multivariate ts are supported by the inverse_transform * fix: updated changelog * feat: make the transformer fittable to improve inversability * fix: when using anchored low freq, try to adjust the time index when inverse transform either the fitted or predicted ts * fix: addressed reviewer comments * fix: bug in TimeSeries constructor when the length of the pd.Series containing the static covariates was equal to the number of components * fix: bug in static_covariates, component-specific static covariates were not supported * fix: new argument in pandas 2.0 * fix: improved finite row detection * fix: revert changes to timeseries.py * fix: tests properly account for static covariates representation depending on the number of components * test: added test for strip=True * feat: adding how argument to TimeSeries.strip(), updated tests * fix: MIDAS properly leverage the TimeSeries.strip method * fix: MIDAS properly leverage the TimeSeries.strip method * fix: fixed bug when series was sliced prior to inverse transform * fix: for anchored low freq, the first value of the row correspond to the anchor value. * fix: splitting aligned/shifted unit-tests * work with moving windows for midas transform * fix all tests * update midas transform * fix transformation * refactor transform * update midas feature separator * remove old create midas df * probabilistic support * update changelog --------- Co-authored-by: Beerstabr <boydbiersteker@gmail.com> Co-authored-by: Boyd Biersteker <108391625+Beerstabr@users.noreply.github.com> Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 18 November 2023, 15:20:58 UTC |
72ea486 | madtoinou | 18 November 2023, 15:13:18 UTC | Fix/dlinear and nlinear use_static_cov. with multivariate series (#2070) * feat: added test to check that use_static_covariates covers all possible static covariates representations * fix: properly account for the two possible static covariates representation in multivariates series * fix: typo in the warning message * feat: added type hint, reordered docstring to match argument order * feat: added type hint, reordered docstring to match argument order (dlinear) * feat: updated changelog --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 18 November 2023, 15:13:18 UTC |
f196665 | madtoinou | 18 November 2023, 15:03:10 UTC | Fix/nlinear normalization for multivariate series (#2072) * fix: properly slice x when normalizing the target features * feat: updated changelog | 18 November 2023, 15:03:10 UTC |
36a1c09 | madtoinou | 16 November 2023, 13:16:24 UTC | Feat/auto-regression and future values of past covariates documentation (#2049) * feat: added warning about usage of past covariates during auto-regression * feat: slightly changed the docstring for output_chunk_length * fix: forgot to add the new argument to some wrapper * feat: display autoregression and past cov related warning only once when using historical forecast with a global model * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * feat: updated docstring for output_chunk_length in both regression and torch models, updated docstring about past/future covariates requirements in the torch models that were missing it * update docs --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 16 November 2023, 13:16:24 UTC |
c8f5948 | madtoinou | 16 November 2023, 10:44:14 UTC | Fix/add fit/predict_kwargs argument to historical_forecasts (#2050) * fix: num_loader_workers can be passed to historical_forecasts, only relevant for TorchForecastingModels * feat: added fit/predict_kwargs to historical_forecasts, backtest and gridsearch * fix: default value None for dict * feat: increased the number of parameters handled by GlobalForecastingModels._fit_wrapper * fix: removed obsolete arg/docstring * fix: updated docstring * fix: only pass the supported argument to GlobalForecastingModel.predict() * fix: simplify the logic of the fit/predict wrapper and hist fc sanity checks * fix: same signature for all _optimized... * fix: changed the exception into warning * fix: missing arg * fix: harmonized signatures of optimized_hist, improved kwargs checks * feat: added warning when fit_kwargs is set with retrain=False, possibility to deactivate the warnings * feat: improve fit/predict_kwargs handling, add tests * feat: parametrize the tests to check both optimized and unoptimized methods * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * feat: updated changelog * feat: added tests when fit_kwargs contains invalid arguments and retrain=False * fix: set self._uses_future_covariates to True in FutureCovariatesLocalForecastingModel when fitted with a future covariate series * fix: predict_kwargs[trainer] is properly passed to fit_from_dataset * feat: added exception when unsupported covariates are passed to the fit/predict throught the wrapper * feat: added tests checking that the exception are raised when expected * fix: ensemble model pass covariates only to forecasting models supporting them * update changelog * update hist fc tests * fix failing bt test p2 * uddate fit/predict wrappers * update docs * shorten electricity dataset to avoid source data updates --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 16 November 2023, 10:44:14 UTC |
a7f5d09 | madtoinou | 16 November 2023, 07:37:57 UTC | Fix: broken dataset links in transfer learning notebook (#2067) * fix: broken dataset links, updated observation accordingly since results sligthly changed * feat: updated changelog * fix: addressed review comments (updated observation) * change darts v26 to v27 --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 16 November 2023, 07:37:57 UTC |
d206055 | Dennis Bader | 11 November 2023, 13:31:48 UTC | fix dataset for python38 (#2065) | 11 November 2023, 13:31:48 UTC |
09300d9 | madtoinou | 11 November 2023, 12:00:13 UTC | Feat/Example notebook for the regression models (#2039) * feat: example notebook for the regression models and new dataset (energy consumption and weather in Zurich, between 2015 and 2022) * fix: tests, some datasets width were missing * feat: udpated changelog * fix: to keep the API uniform, Zurich energy consumption and weather was split into two datasets. Energy consumption was added to the darts repo * fix: changed the way datasets are loaded, added an illustration for multi_models=True * fix: tweaked notebook * feat: grouped dataset and their width into a single variable to improve readibility * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * fix: simplified API to load the EnergyConsumptionZurich dataset, updated notebook accordingly * fix: remove the obsolete dataset from the tests * blabla * update dataset * update notebook p1 * update regression model notebook * notebook last fixes * fix: typo * add regression model example test to merge workflow --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 11 November 2023, 12:00:13 UTC |
da049e5 | madtoinou | 08 November 2023, 07:51:25 UTC | Fix/exp smooth constructor args (#2059) * feat: adding support for constructor kwargs * feat: adding tests * fix: udpated representation test for ExponentialSmoothing model * update changelog.md --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 08 November 2023, 07:51:25 UTC |
a5a4306 | Dennis Bader | 08 November 2023, 07:39:35 UTC | fix num_samles not being passed to optimized hist fc rountine of torch models (#2060) | 08 November 2023, 07:39:35 UTC |
af5b141 | Freddie Hsin-Fu Huang | 06 November 2023, 15:39:05 UTC | fix: torch_forecasting_model load_weights with float16 and float32 (#2046) | 06 November 2023, 15:39:05 UTC |
772d705 | Dennis Bader | 06 November 2023, 10:06:52 UTC | Feat/tz aware dta (#2054) * add tests for datetime_attribute_timeseries and holidays * add tz to encoders * add tests * add timezone to some timeseries methods * update changelog and model docs * apply suggestions from PR review | 06 November 2023, 10:06:52 UTC |
2d0233f | Dennis Bader | 03 November 2023, 13:29:17 UTC | Native quantile regression for xgb 2.0.0 and above (#2051) * use native quantile regression for xgb 2.0.0 and above * update changelog * revert change to objective --------- Co-authored-by: madtoinou <32447896+madtoinou@users.noreply.github.com> | 03 November 2023, 13:29:17 UTC |
3fb1080 | madtoinou | 03 November 2023, 12:25:00 UTC | Feat/summary_plot() returns the shap values (#2048) * feat: summary_plot returns the shap explanations * feat: returning the whole dict * feat: adding a test * fix:typo in type hint | 03 November 2023, 12:25:00 UTC |
f6e994e | Felix Divo | 01 November 2023, 14:30:02 UTC | Fix doc rendering in statistics.py (#2044) * Fix doc rendering in statistics.py * fix other things in statistics docs --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 01 November 2023, 14:30:02 UTC |
ae5caa6 | Felix Divo | 01 November 2023, 14:19:59 UTC | Fix doc formatting of LinearRegressionModel (#2045) * For doc formatting of LinearRegressionModel * fix other model docs --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 01 November 2023, 14:19:59 UTC |
ea37dc9 | Dennis Bader | 28 October 2023, 15:03:49 UTC | Feat/optimized hfc torch (#2013) * move hfc optimization checks to GlobalForecastingModel * setup optimization files for hfc with torch models * adapt torch infrerence datasets to work with stride and bounds * first working version * adapt for overlap_end=True * fix test * make tests for integer indexed series * make multiple ts work * update documentation * fix issue with regression model optim hfc * fix basic sample comparison * allow autoregression in optim hfc for torch models * remove some unnecessary lines * update changelog * refactor hist fc forecastable index * add unit test for exact end * update docs * apply suggestions from PR review | 28 October 2023, 15:03:49 UTC |
c777193 | Samriddhi Singh | 28 October 2023, 14:26:11 UTC | Fix/plot alpha kwarg (#2011) * rm data leakage in example notebooks * fix: linting * parameter alpha in plots * fix: reverting changes * parameter alpha changes --------- Co-authored-by: simsalabim1 <63498227+simsalabim1@users.noreply.github.com> Co-authored-by: madtoinou <32447896+madtoinou@users.noreply.github.com> Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 28 October 2023, 14:26:11 UTC |
039d898 | madtoinou | 28 October 2023, 14:25:31 UTC | Fix/optimized historical forecast with component specific lags (#2040) * fix: properly call model._get_lags instead of model.lags.get to account for component specific lags * updated changelog * add unit tests --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 28 October 2023, 14:25:31 UTC |
e6f2208 | madtoinou | 28 October 2023, 13:55:19 UTC | Fix/operand error with encoders (#2034) * fix: create a temporary Datetime index when series frequency represents a ambiguous timedelta value to extract the start time index * feat: updated changelog * fix: fixed corner case, generate the shortest temporary datetimeindex possible * feat: added tests to cover the cases where the series freq cannot be converted to Timedelta | 28 October 2023, 13:55:19 UTC |
f3bdbcf | Samriddhi Singh | 13 October 2023, 06:29:55 UTC | Doc/data leak (#2020) * changes * deleating logs * update notebooks * add progress bar callback * update first notebook with new progress bar * update remaining notebooks --------- Co-authored-by: simsalabim1 <63498227+simsalabim1@users.noreply.github.com> Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 13 October 2023, 06:29:55 UTC |
2f7c9db | dennisbader | 16 September 2023, 17:34:06 UTC | Release 0.26.0 | 16 September 2023, 17:34:06 UTC |
9982a50 | Dennis Bader | 16 September 2023, 17:01:23 UTC | udpate ensemble example notebook (#1996) | 16 September 2023, 17:01:23 UTC |
74bd524 | Dennis Bader | 16 September 2023, 15:53:33 UTC | remove torch import (#1995) | 16 September 2023, 15:53:33 UTC |
98c3ff4 | Dennis Bader | 16 September 2023, 15:43:45 UTC | Release 0.26.0 (#1994) * bump u8darts 0.25.0 * update changelog | 16 September 2023, 15:43:45 UTC |
1929d9f | madtoinou | 16 September 2023, 15:28:46 UTC | feat: EnsembleModel accepts pretrained global models (#1815) * feat: EnsembleModel can accept pretrained global models * fix: moved retrain_forecasting_models to the constructor (default is True), renamed models to forecasting_models in all ensemble, improved the logic around retraining * fix: updated argument name in notebook * updated changelog * addressing review comments * fix: updare training_series attribute of models in NaiveEnsemble so that predict() behave as expected * fix: bug for regression_train_n_points=-1 * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * fix: revert changes in type hinting * fix: changes models to forecasting_models in all the exception messages * fix: linting in docstring * fix: type hinting * fix: errors caused by merge with master * fix: error message when ensemble contains local models and historical forecast is called with retrain=False * doc: updated changelog * fix: EnsembleModel support_lkl_params_predict was not taking into account probabilistic models fitted without likelihood such as ExponentialSmoothing * doc: new notebook for the ensemble models * feat: possible to use historical forecasts to train the regression model of RegressionEnsemble * doc: updating the notebook * fix: conversion unittest to pytest * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * fix: addressed review comments, added test for train_using_historical_forecast * doc: updated the notebook based on the review comments * doc: updated the notebook based on the review comments * doc: removed uneccessary plot() and imports * doc: fir formatting * feat: adjust start of historical forecast depending on the ocl of forecasting models * fix: simplied a bit the series2seq ans seq2series logic * doc: updating the CI/CD and reference to the notebook in the doccumentation * fix: addressing review comments * feat: when regression_train_n_points=-1 and training with several series, the number of points is optimised for eash series independently * test: also check the length of the series used to train the regression model of the ensemble * fix: addressing reviewer comments, identified a bug in historical forecasts * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * update start_format and train_using_historical_forecasts checks * fix issue with too short historical forecasts * revert changes to fitted model check with train_using_historical_forecast * update tests * fix last predictable time step issues for historical forecasting * add new unit test for optimized hist fc for regression models * extend test with stride * fix overlap end issue with optimized hist fc for regression models * fix last issues * apply suggestions from pr review * reactivate commented tests --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 16 September 2023, 15:28:46 UTC |
fca3993 | Dennis Bader | 15 September 2023, 15:57:00 UTC | add feature projection for past covariates to TiDEModel (#1993) | 15 September 2023, 15:57:00 UTC |
a9b6fbc | Eliot Zubkoff | 14 September 2023, 10:18:49 UTC | feature/cumsum (#1988) * added cumsum function to time series * typo in unit test function naming * update changelog.md * apply suggestions from pr review and extend tests --------- Co-authored-by: eliot <eliot@eliotz.dev> Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 14 September 2023, 10:18:49 UTC |
b3498bf | madtoinou | 14 September 2023, 09:54:35 UTC | Feat/specify lags per component for RegressionModel (#1962) * feat: updated lags sanity checks to accept dictionnary * fix: better management of corner cases during lags checks * fix: improved modularity * fix: simplified the logic a bit * feat: when generating lagged data, the values can be extracted using component-specific lags * feat: raise error if all the ts in target/past/future don't have the same number of components * feat: added support for component-specific lags in fit() and predict() * test: added tests and fix some bug accordingly * feat: component-wise lags support encoders, improved sanity checks * feat: possibility to declare default lags for all the not specified components, updated changelog * test: adding a test for the lagged data creation * fix: typo * fix: adressing review comments * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * refactor: lags argument are converted to dict before running the type check and processing of the values * refactor: lags argument are converted to dict before running the type check and processing of the values * doc: improved documentation of the component-specific lags in tabularization * test: adding a test for the multivariate scenario * test: checking the appriopriate lags are extracted by the shap explainer * fix: shapexplainer extract the appropriate lags, updated the type hints * fix: passing covariates when trained on multiple series * fix: moved the series components consistency to create_lagged_data to limit iteration of the series * fix: improved the error message for components inconsistency, improve tests parametrization * fix: addressing reviewer comments * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * test: checking that the name of the features is correctly generated when using dict to define the lags * fix: linting * fix: updating the error msg * fix: bug when the number of lags is different across components * fix: future lags in test --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 14 September 2023, 09:54:35 UTC |
a6ceb5d | Dennis Bader | 14 September 2023, 08:56:39 UTC | remove warning for missing dependencies for main darts package (#1986) | 14 September 2023, 08:56:39 UTC |
59937f3 | madtoinou | 14 September 2023, 08:43:14 UTC | Doc/Adding basic usage example in each model docstring (#1956) * doc: adding very basic usage example for each model * doc: addressing review comments * fix: revert changes, will be included in a separate PR * updated changelog * fix: addressed review comments * added review suggestion * update leftover lambda mentions in model docs * update statsforecast model examples --------- Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 14 September 2023, 08:43:14 UTC |
626ac36 | madtoinou | 12 September 2023, 15:41:34 UTC | fix: force tree_method argument since xgboost changed the default value to hist (#1990) | 12 September 2023, 15:41:34 UTC |
74ed2bb | Eliot Zubkoff | 07 September 2023, 14:36:18 UTC | Fix NLinear normalization to support past covariates (#1873) * adding test for past covariates, need to check with darts about past covariates needing to end in the future * slicing typo * fix slicing * add explicit comment from @felixdivo Co-authored-by: Felix Divo <4403130+felixdivo@users.noreply.github.com> * update test to use future covariates in predict function due to autoregressive predictions * linting * fix tests * making changes proposed by @dennisbader * update comment * remove double denormalization * update CHANGELOG.md --------- Co-authored-by: eliot <eliot@eliotz.dev> Co-authored-by: Felix Divo <4403130+felixdivo@users.noreply.github.com> Co-authored-by: dennisbader <dennis.bader@gmx.ch> | 07 September 2023, 14:36:18 UTC |
f5259b9 | Vladimir Fokow | 06 September 2023, 07:02:07 UTC | fix typo in documentation (#1979) * fix typo * fix another typo (grammar) in docs | 06 September 2023, 07:02:07 UTC |
fecb99d | Dennis Bader | 02 September 2023, 10:21:02 UTC | add rin to all torch models (#1969) * add rin to all torch models * fix failint transformer test * update CHANGELOG.md * update model docs * ignore RIN for RNNModel * apply suggestions from PR review * fix model links in readme * remove statforecast utils from generated API Reference * Update CHANGELOG.md * add RIN reference link to PLForecastingModule | 02 September 2023, 10:21:02 UTC |
52ac181 | Alex Colpitts | 31 August 2023, 12:58:58 UTC | Feat/Dockerfile test environment (#1878) * changed dockerfile for testing * Update Dockerfile Co-authored-by: madtoinou <32447896+madtoinou@users.noreply.github.com> * update INSTALL.md * update CHANGELOG.md * Apply suggestions from code review Co-authored-by: madtoinou <32447896+madtoinou@users.noreply.github.com> --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> Co-authored-by: madtoinou <32447896+madtoinou@users.noreply.github.com> | 31 August 2023, 12:58:58 UTC |
bbb6aa6 | madtoinou | 31 August 2023, 12:24:41 UTC | Fix/Updated tests with patch release 2.0.8 of PL (#1976) * fix: release 2.0.8 of pytorch lightning slightly changed the performance of models, updated the error value for the TiDE model which was failing, separated the univariate and multivariate tests, made the values closer to the expected values to detect similar changed in the future * feat: parametrize the deterministic test using pytest | 31 August 2023, 12:24:41 UTC |
e1abf51 | madtoinou | 31 August 2023, 07:22:38 UTC | Fix/no lambda function in add_encoders (#1957) * feat: prevent lambda functions in the encoders as they can cause problem when saving the model (pickling) * doc: updated notebook * fix: updated test of shap explainer * fix: pickable encoder in shap tests * doc: updating the docstring of all the models supporting * doc: updated changelog | 31 August 2023, 07:22:38 UTC |
3d6e07b | twobitunicorn | 31 August 2023, 06:58:59 UTC | Update timeseries.py (#1975) In a command line the missing space gives us the word "contiguousin" instead of "contiguous in". Added said space. Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 31 August 2023, 06:58:59 UTC |
76d3634 | madtoinou | 31 August 2023, 06:55:16 UTC | Fix/stasforecast 1 6 release (#1974) * fix: updated error value after upstream bugfix (AAZ could become AAM, a forbidden combination, which was occuring in darts unittest) * doc: updated docstring, fixed link to statsforecast documentation, added note about conformal prediction * fix: typo * doc: removing argument documentation, refering to statsforecast documentation only * fix: typo | 31 August 2023, 06:55:16 UTC |
27278c1 | Biebrya | 25 August 2023, 06:34:41 UTC | Update tft_explainer.py (#1967) tft_explainer has a shap_explainer doc in its readme, changing it be TFTExplainer for more copy/pasteability | 25 August 2023, 06:34:41 UTC |
6096968 | Dennis Bader | 18 August 2023, 07:00:53 UTC | fix failing core requirement imports (#1960) | 18 August 2023, 07:00:53 UTC |
7c6e04a | Dennis Bader | 17 August 2023, 13:04:37 UTC | Refactor/unittests (#1950) * replace unittest base class with pytest * add pytest conftest for migrating from unittest to pytest * update test_logging.py * update test_timeseries.py * update test_timeseries_multivariate.py * update test_timeseries.py * refactored tests from unittest to pytest * fix failing unittests * parametrize slow tests * remove some unittest decorators * fix failing unit tests * Update CHANGELOG.md * apply suggestions from code review * remove checkpoints from git history * fix unit test | 17 August 2023, 13:04:37 UTC |
b1290cb | madtoinou | 15 August 2023, 09:02:12 UTC | Feat/historical_forecasts accept negative integer as start value (#1866) * feat: historical_foreacst accept negative integer as start value * fix: improved the negative start unit test * fix: simplified the logic around exception raising * fix: instead of adding capabilities to get_index_at_point, use a new argument in historical_forecasts. Updated exception accordingly * test: udpated tests accordingly * doc: updated changelog * test: added test for historical forecast on ts using a rangeindex starting with a negative value * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * fix: changed the literal to 'positional_index' and 'value_index' * feat: making the error messages more informative, adapted the tests accordingly * feat: extending the new argument to backtest and gridsearch * fix: import of Literal for python 3.8 * doc: updated changelog * fix: shortened the literal for start_format, updated tests accordingly * doc: updated start docstring * test: limited the dependency on unittest in anticipation of the refactoring * doc: updated changelog * fix: fixed typo * fix: fixed typo * doc: copy start and start_format docstring from hist_fct to backtest and gridsearch * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 15 August 2023, 09:02:12 UTC |
b69b8ca | madtoinou | 15 August 2023, 07:48:36 UTC | Fix/Robuster parameters check when loading weights (#1952) * fix: comparing the parameters stored in .model_params (saved and loading models) to make it more robust. this check can be skipped (not recommended). * feat: nicer message, all the hp discrepancies are listed at once * feat: repr method for LikelihoodModel * fix: removed unused param from LikelihoodModel.__repr__ * fix: better sanity check of the kwargs during weights loading from ckpt * fix: removed copy-paste leftovers from unittests * fix: removing redudant helper code * test: adding tests, made the assertion more specific to distinguish incorrect/missing * Apply suggestions from code review Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> * fix: linting * Update CHANGELOG.md --------- Co-authored-by: Dennis Bader <dennis.bader@gmx.ch> | 15 August 2023, 07:48:36 UTC |