* Refactor into automl subpackage
Moved some of the packages into an automl subpackage to tidy before the
task-based refactor. This is in response to discussions with the group
and a comment on the first task-based PR.
Only changes here are moving subpackages and modules into the new
automl, fixing imports to work with this structure and fixing some
dependencies in setup.py.
* Fix doc building post automl subpackage refactor
* Fix broken links in website post automl subpackage refactor
* Fix broken links in website post automl subpackage refactor
* Remove vw from test deps as this is breaking the build
* Move default back to the top-level
I'd moved this to automl as that's where it's used internally, but had
missed that this is actually part of the public interface so makes sense
to live where it was.
* Re-add top level modules with deprecation warnings
flaml.data, flaml.ml and flaml.model are re-added to the top level,
being re-exported from flaml.automl for backwards compatability. Adding
a deprecation warning so that we can have a planned removal later.
* Fix model.py line-endings
* WIP
* WIP - Notes below
Got to the point where the methods from AutoML are pulled to
GenericTask. Started removing private markers and removing the passing
of automl to these methods. Done with decide_split_type, started on
prepare_data. Need to do the others after
* Re-add generic_task
* Most of the merge done, test_forecast_automl fit succeeds, fails at predict()
* Remaining fixes - test_forecast.py passes
* Comment out holidays-related code as it's not currently used
* Further holidays cleanup
* Fix imports in a test
* tidy up validate_data in time series task
* Test fixes
* Fix tests: add Task.__str__
* Fix tests: test for ray.ObjectRef
* Hotwire TS_Sklearn wrapper to fix test fail
* Attempt at test fix
* Fix test where val_pred_y is a list
* Attempt to fix remaining tests
* Push to retrigger tests
* Push to retrigger tests
* Push to retrigger tests
* Push to retrigger tests
* Remove plots from automl/test_forecast
* Remove unused data size field from Task
* Fix import for CLASSIFICATION in notebook
* Monkey patch TFT to avoid plotting, to fix tests on MacOS
* Monkey patch TFT to avoid plotting v2, to fix tests on MacOS
* Monkey patch TFT to avoid plotting v2, to fix tests on MacOS
* Fix circular import
* remove redundant code in task.py post-merge
* Fix test: set svd_solver="full" in PCA
* Update flaml/automl/data.py
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
* Fix review comments
* Fix task -> str in custom learner constructor
* Remove unused CLASSIFICATION imports
* Hotwire TS_Sklearn wrapper to fix test fail by setting
optimizer_for_horizon == False
* Revert changes to the automl_classification and pin FLAML version
* Fix imports in reverted notebook
* Fix FLAML version in automl notebooks
* Fix ml.py line endings
* Fix CLASSIFICATION task import in automl_classification notebook
* Uncomment pip install in notebook and revert import
Not convinced this will work because of installing an older version of
the package into the environment in which we're running the tests, but
let's see.
* Revert c6a5dd1a0
* Fix get_classification_objective import in suggest.py
* Remove hcrystallball docs reference in TS_Sklearn
* Merge markharley:extract-task-class-from-automl into this
* Fix import, remove smooth.py
* Fix dependencies to fix TFT fail on Windows Python 3.8 and 3.9
* Add tensorboardX dependency to fix TFT fail on Windows Python 3.8 and 3.9
* Set pytorch-lightning==1.9.0 to fix TFT fail on Windows Python 3.8 and 3.9
* Set pytorch-lightning==1.9.0 to fix TFT fail on Windows Python 3.8 and 3.9
* Disable PCA reduction of lagged features for now, to fix svd convervence fail
* Merge flaml/main into time_series_task
* Attempt to fix formatting
* Attempt to fix formatting
* tentatively implement holt-winters-no covariates
* fix forecast method, clean class
* checking external regressors too
* update test forecast
* remove duplicated test file, re-add sarimax, search space cleanup
* Update flaml/automl/model.py
removed links. Most important one probably was: https://robjhyndman.com/hyndsight/ets-regressors/
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
* prevent short series
* add docs
* First attempt at merging Holt-Winters
* Linter fix
* Add holt-winters to TimeSeriesTask.estimators
* Fix spark test fail
* Attempt to fix another spark test fail
* Attempt to fix another spark test fail
* Change Black max line length to 127
* Change Black max line length to 120
* Add logging for ARIMA params, clean up time series models inheritance
* Add more logging for missing ARIMA params
* Remove a meaningless test causing a fail, add stricter check on ARIMA params
* Fix a bug in HoltWinters
* A pointless change to hopefully trigger the on and off KeyError in ARIMA.fit()
* Fix formatting
* Attempt to fix formatting
* Attempt to fix formatting
* Attempt to fix formatting
* Attempt to fix formatting
* Add type annotations to _train_with_config() in state.py
* Add type annotations to prepare_sample_train_data() in state.py
* Add docstring for time_col argument of AutoML.fit()
* Address @sonichi's comments on PR
* Fix formatting
* Fix formatting
* Reduce test time budget
* Reduce test time budget
* Increase time budget for the test to pass
* Remove redundant imports
* Remove more redundant imports
* Minor fixes of points raised by Qingyun
* Try to fix pandas import fail
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Try to fix pandas import fail, again
* Formatting fixes
* More formatting fixes
* Added test that loops over TS models to ensure coverage
* Fix formatting issues
* Fix more formatting issues
* Fix random fail in check
* Put back in tests for ARIMA predict without fit
* Put back in tests for lgbm
* Update test/test_model.py
cover dedup
* Match target length to X length in missing test
---------
Co-authored-by: Mark Harley <mark.harley@transferwise.com>
Co-authored-by: Mark Harley <mharley.code@gmail.com>
Co-authored-by: Qingyun Wu <qingyun.wu@psu.edu>
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
Co-authored-by: Andrea W <a.ruggerini@ammagamma.com>
Co-authored-by: Andrea Ruggerini <nescio.adv@gmail.com>
Co-authored-by: Egor Kraev <Egor.Kraev@tw.com>
Co-authored-by: Li Jiang <bnujli@gmail.com>
* update openai model support
* new gpt3.5
* docstr
* function_call and content may co-exist
* test function call
---------
Co-authored-by: Qingyun Wu <qingyun.wu@psu.edu>
* version update post release v1.2.2
* automl option
* import pandas
* remove automl.utils
* default
* test
* type hint and version update
* dependency update
* link to open in colab
* use packging.version to close#725
---------
Co-authored-by: Li Jiang <lijiang1@microsoft.com>
Co-authored-by: Li Jiang <bnujli@gmail.com>
* Mark experimental classes
* template
* multi model
* test
* multi-config doc
* doc
* doc
* test
---------
Co-authored-by: Li Jiang <bnujli@gmail.com>
* math utils in autogen
* cleanup
* code utils
* remove check function from code response
* comment out test
* GPT-4
* increase request timeout
* name
* logging and error handling
* better doc
* doc
* codegen optimized
* GPT series
* text
* no demo example
* math
* import openai
* import openai
* azure model name
* azure model name
* openai version
* generate assertion if necessary
* condition to generate assertions
* init region key
* rename
* comments about budget
* prompt
---------
Co-authored-by: Susan Xueqing Liu <liususan091219@users.noreply.github.com>
* add basic support to Spark dataframe
add support to SynapseML LightGBM model
update to pyspark>=3.2.0 to leverage pandas_on_Spark API
* clean code, add TODOs
* add sample_train_data for pyspark.pandas dataframe, fix bugs
* improve some functions, fix bugs
* fix dict change size during iteration
* update model predict
* update LightGBM model, update test
* update SynapseML LightGBM params
* update synapseML and tests
* update TODOs
* Added support to roc_auc for spark models
* Added support to score of spark estimator
* Added test for automl score of spark estimator
* Added cv support to pyspark.pandas dataframe
* Update test, fix bugs
* Added tests
* Updated docs, tests, added a notebook
* Fix bugs in non-spark env
* Fix bugs and improve tests
* Fix uninstall pyspark
* Fix tests error
* Fix java.lang.OutOfMemoryError: Java heap space
* Fix test_performance
* Update test_sparkml to test_0sparkml to use the expected spark conf
* Remove unnecessary widgets in notebook
* Fix iloc java.lang.StackOverflowError
* fix pre-commit
* Added params check for spark dataframes
* Refactor code for train_test_split to a function
* Update train_test_split_pyspark
* Refactor if-else, remove unnecessary code
* Remove y from predict, remove mem control from n_iter compute
* Update workflow
* Improve _split_pyspark
* Fix test failure of too short training time
* Fix typos, improve docstrings
* Fix index errors of pandas_on_spark, add spark loss metric
* Fix typo of ndcgAtK
* Update NDCG metrics and tests
* Remove unuseful logger
* Use cache and count to ensure consistent indexes
* refactor for merge maain
* fix errors of refactor
* Updated SparkLightGBMEstimator and cache
* Updated config2params
* Remove unused import
* Fix unknown parameters
* Update default_estimator_list
* Add unit tests for spark metrics
* improve max_valid_n and doc
* Update README.md
Co-authored-by: Li Jiang <lijiang1@microsoft.com>
* add support for chatgpt
* notebook
* newline at end of file
* chatgpt notebook
* ChatGPT in Azure
* doc
* math
* warning, timeout, log file name
* handle import error
* doc update; default value
* paper
* doc
* docstr
* eval_func
* prompt and messages
* remove confusing words
* notebook name
---------
Co-authored-by: Li Jiang <lijiang1@microsoft.com>
Co-authored-by: Susan Xueqing Liu <liususan091219@users.noreply.github.com>
* improve max_valid_n and doc
* Update README.md
Co-authored-by: Li Jiang <lijiang1@microsoft.com>
* newline at end of file
* doc
---------
Co-authored-by: Li Jiang <lijiang1@microsoft.com>
Co-authored-by: Susan Xueqing Liu <liususan091219@users.noreply.github.com>
Co-authored-by: Qingyun Wu <qingyun.wu@psu.edu>
* add cost budget; move loc of make_dir
* support openai completion
* install pytest in workflow
* skip openai test
* test openai
* path for docs rebuild
* install datasets
* signal
* notebook
* notebook in workflow
* optional arguments and special params
* key -> k
* improve readability
* assumption
* optimize for model selection
* larger range of max_tokens
* notebook
* python package workflow
* skip on win
* notebook test
* add ipykernel, remove except
* only create dir if not empty
* Stop sequential tuning when result is None
* fix reproducibility of global search
* save gs seed
* use get to avoid KeyError
* test
* make performance test reproducible
* fix test error
* Doc update and disable logging
* document random_state and version
* remove hardcoded budget
* fix test error and dependency; close#777
* iloc
* install editable package in codespace
* fix test error in test_forecast
* fix test error in test_space
* openml version
* break tests; pre-commit
* skip on py10+win32
* install mlflow in test
* install mlflow in [test]
* skip test in windows
* import
* handle PermissionError
* skip test in windows
* skip test in windows
* skip test in windows
* skip test in windows
* remove ts_forecast_panel from doc
* add vw version requirement
* vw version
* version range
* add documentation
* vw version range
* skip test on py3.10
* vw version
* rephrase
* don't install vw on py 3.10
* move import location
* remove inherit
* 3.10 in version
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
* time series forecasting with panel datasets
- integrate Temporal Fusion Transformer as a learner based on pytorchforecasting
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update setup.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update test_forecast.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update setup.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update setup.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update model.py and test_forecast.py
- remove blank lines
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update model.py to prevent errors
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update automl.py and data.py
- change forecast task name
- update documentation for fit() method
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update test_forecast.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update test_forecast.py
- add performance test
- use 'fit_kwargs_by_estimator'
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* add time index function
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update test_forecast.py performance test
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update data.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update automl.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update data.py to prevent type error
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update setup.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update for pytorch forecasting tft on panel datasets
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update automl.py documentations
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* - rename estimator
- add 'gpu_per_trial' for tft estimator
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update test_forecast.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* include ts panel forecasting as an example
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update model.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update documentations
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update automl_time_series_forecast.ipynb
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update documentations
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* "weights_summary" argument deprecated and removed for pl.Trainer()
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update model.py tft estimator prediction method
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update model.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update `fit_kwargs` documentation
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* update automl.py
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
* support latest xgboost version
* Update test_classification.py
* Update
Exists problems when installing xgb1.6.1 in py3.6
* cleanup
* xgboost version
* remove time_budget_s in test
* remove redundancy
* stop support of python 3.6
Co-authored-by: zsk <shaokunzhang529@gmail.com>
Co-authored-by: Qingyun Wu <qingyun.wu@psu.edu>
Issue I encountered:
#542 run test_restore.py and got _pickle.UnpicklingError: state is not a dictionary
I observed:
1. numpy version
i. When numpy==1.16*, np.random.RandomState.__getstate__() returns a tuple, not a dict.
_pickle.UnpicklingError occurs
ii. When numpy>1.17.0rc1, it returns a dict;
_pickle.UnpicklingError does not occur
iii. When numpy>1.17.0rc1, flaml uses np_random_generator = np.random.Generator,
_pickle.UnpicklingError does not occur
2. class _BackwardsCompatibleNumpyRng
When I remove func _BackwardsCompatibleNumpyRng.__getattr__() , _pickle.UnpicklingError doesn't occur (regardless of numpy version == 1.16* or 1.17*)
To sum up,
I think making modifications to class _BackwardsCompatibleNumpyRng is not a good choice (_BackwardsCompatibleNumpyRng came from ray）and we still need to learn more about the operation mechanism of pickle.
So I upgraded the numpy version that flaml requires：
setup.py："NumPy>=1.17.0rc1"
* add sklearn regressors as learners for ts_forecast task
* add direct forecasting strategy
warnings and errors for duplicate rows and missing values
- add preprocess for sklearn time series forecast
update automl.py
update test/test_forecast.py
* update model.py and test_forecast.py for cv eval_method
* add "hcrystalball" dependency in setup.py
* update automl.py
- add _validate_ts_data function for abstraction
- include xgb_limitdepth as a learner
* update model.py
- update search space for sklearn ts regressors
* update automl.py and test_forecast.py for numpy array inputs
* add documentations to model.py
* add documentation for removing catboost regressor
* update automl.py
- _validate_ts_data() function
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* limit time and memory
* separate tests
* lrl1 can't be limited by limit_resource
* free memory when possible
* passthrough=False when ensemble fails;
retrain when trained_estimator is None
* use callback to for resource limit
* handle lower version of xgb with no callback
* free mem ratio
* reduce verbosity
* retrain_final when max_iter==1
* remove trained_estimator from result
* model_history
* wheel
* retrain time as best_config_train_time
* ci: libomp version for xgboost on macos
* limit_resource not working in windows
* test pickle load
* mute forecaster
* notebook update
* check hard
* preventive callback
* add use_ray
* Integrate multivariate time series forecasting, now supports
continuous and categorical variables
- update data.py to transform time series data
- update search space
- update documentations to reflect changes
- update test_forecast.py
- rename 'forecast' task to 'ts_forecast' task
* update automl.py and test_forecast.py
* update forecast notebook
* update README.md and setup.py
* update ml.py and test_forecast.py
- make "ds" and "y" constant variables
* replace constants with constant variables
* bump version to 0.7.0
* update setup.py
- support 'forecast' and 'ts_forecast'
* update automl.py and data.py
- support 'forecast' and 'ts_forecast' tasks
* warning -> info for low cost partial config
#195, #110
* when n_estimators < 0, use trained_estimator's
* log debug info
* test random seed
* remove "objective"; avoid ZeroDivisionError
* hp config to estimator params
* check type of searcher
* default n_jobs
* try import
* Update searchalgo_auto.py
* CLASSIFICATION
* auto_augment flag
* min_sample_size
* make catboost optional
* config in result
* value can be float
* pytorch notebook example
* docker, pre-commit
* max_failure (#192); early_stop
* extend starting_points (#196)
Co-authored-by: Chi Wang (MSR) <wang.chi@microsoft.com>
Co-authored-by: Qingyun Wu <qw2ky@virginia.edu>
* remove catboost training dir
* close#48
* bs for hierarchical space. close#85
* retrain for hierarchical space
* clean ml (#180)
Co-authored-by: Qingyun Wu <qxw5138@psu.edu>
* support ranking task
* examples
* cv shuffle
* forecast api and implementation cleaner
* period constraints
* delete groups after fit
* added 'forecast' task with estimators ['fbprophet', 'arima', 'sarimax']
* update setup.py
* add TimeSeriesSplit to 'regression' and 'classification' task
* add 'time' split_type for 'classification' and 'regression' task
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* feature importance
* variable name
* Update test/test_split.py
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
* Update test/test_forecast.py
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
* prophet installation fail in windows
* upload flaml_forecast.ipynb
Signed-off-by: Kevin Chen <chenkevin.8787@gmail.com>
* subspace in flow2
* search space and trainable from AutoML
* experimental features: multivariate TPE, grouping, add_evaluated_points
* test experimental features
* readme
* define by run
* set time_budget_s for bs
Co-authored-by: liususan091219 <Xqq630517>
* version
* acl
* test define_by_run_func
* size
* constraints
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
* pickle the AutoML object
* get best model per estimator
* test deberta
* stateless API
* pickle the AutoML object
* get best model per estimator
* test deberta
* stateless API
* prevent divide by zero
* test roberta
* BlendSearchTuner
* sync
* version number
* update gitignore
* delta time
* reindex columns when dropping int-indexed columns
* add seed
* add seed in Args
* merge
* init upload of ChaCha
* remove redundancy
* add back catboost
* improve AutoVW API
* set min_resource_lease in VWOnlineTrial
* docstr
* rename
* docstr
* add docstr
* improve API and documentation
* fix name
* docstr
* naming
* remove max_resource in scheduler
* add TODO in flow2
* remove redundancy in rearcher
* add input type
* adapt code from ray.tune
* move files
* naming
* documentation
* fix import error
* fix format issues
* remove cb in worse than test
* improve _generate_all_comb
* remove ray tune
* naming
* VowpalWabbitTrial
* import error
* import error
* merge test code
* scheduler import
* fix import
* remove
* import, minor bug and version
* Float or Categorical
* fix default
* add test_autovw.py
* add vowpalwabbit and openml
* lint
* reorg
* lint
* indent
* add autovw notebook
* update notebook
* update log msg and autovw notebook
* update autovw notebook
* update autovw notebook
* add available strings for model_select_policy
* string for metric
* Update vw format in flaml/onlineml/trial.py
Co-authored-by: olgavrou <olgavrou@gmail.com>
* make init_config optional
* add _setup_trial_runner and update notebook
* space
Co-authored-by: Chi Wang (MSR) <chiw@microsoft.com>
Co-authored-by: Chi Wang <wang.chi@microsoft.com>
Co-authored-by: Qingyun Wu <qiw@microsoft.com>
Co-authored-by: olgavrou <olgavrou@gmail.com>
* pickle the AutoML object
* get best model per estimator
* test deberta
* stateless API
* Add Gitter badge (#41)
* prevent divide by zero
* test roberta
* BlendSearchTuner
Co-authored-by: Chi Wang (MSR) <chiw@microsoft.com>
Co-authored-by: The Gitter Badger <badger@gitter.im>
* xgboost notebook
* finetuning notebook
* finetuning test
* experimental nni support
* support nested search space
* log file name
* record training_iteration
* eps
* reset times
* std set to default step size if 0
* v0.2.2
separate the HPO part into the module flaml.tune
enhanced implementation of FLOW^2, CFO and BlendSearch
support parallel tuning using ray tune
add support for sample_weight and generic fit arguments
enable mlflow logging
Co-authored-by: Chi Wang (MSR) <chiw@microsoft.com>
Co-authored-by: qingyun-wu <qw2ky@virginia.edu>
* set default logging level to INFO
* remove unnecessary import
* API future compatibility
* add test for customized learner
* test dependency
Co-authored-by: Chi Wang (MSR) <chiw@microsoft.com>
Shaokun