autogluon.timeseries 1.0.0b20231124__py3-none-any.whl → 1.0.0b20231125__py3-none-any.whl

autogluon/timeseries/dataset/ts_dataframe.py

@@ -5,6 +5,7 @@ import itertools
 import logging
 import reprlib
 from collections.abc import Iterable
+from itertools import islice
 from pathlib import Path
 from typing import Any, List, Optional, Tuple, Type, Union
 
@@ -849,6 +850,8 @@ class TimeSeriesDataFrame(pd.DataFrame):
         freq: Union[str, pd.DateOffset],
         agg_numeric: str = "mean",
         agg_categorical: str = "first",
+        num_cpus: int = -1,
+        chunk_size: int = 100,
         **kwargs,
     ) -> TimeSeriesDataFrame:
         """Convert each time series in the data frame to the given frequency.
@@ -858,6 +861,10 @@ class TimeSeriesDataFrame(pd.DataFrame):
         1. Converting an irregularly-sampled time series to a regular time index.
         2. Aggregating time series data by downsampling (e.g., convert daily sales into weekly sales)
 
+        Standard ``df.groupby(...).resample(...)`` can be extremely slow for large datasets, so we parallelize this
+        operation across multiple CPU cores.
+
+
         Parameters
         ----------
         freq : Union[str, pd.DateOffset]
@@ -867,6 +874,10 @@ class TimeSeriesDataFrame(pd.DataFrame):
             Aggregation method applied to numeric columns.
         agg_categorical : {"first", "last"}, default = "first"
             Aggregation method applied to categorical columns.
+        num_cpus : int, default = -1
+            Number of CPU cores used when resampling in parallel. Set to -1 to use all cores.
+        chunk_size : int, default = 100
+            Number of time series in a chunk assigned to each parallel worker.
         **kwargs
             Additional keywords arguments that will be passed to ``pandas.DataFrameGroupBy.resample``.
 
@@ -928,7 +939,8 @@ class TimeSeriesDataFrame(pd.DataFrame):
         0       2020-12-31    10.0
                 2021-12-31    26.0
         """
-        if self.freq == pd.tseries.frequencies.to_offset(freq).freqstr:
+        offset = pd.tseries.frequencies.to_offset(freq)
+        if self.freq == offset.freqstr:
             return self
 
         # We need to aggregate categorical columns separately because .agg("mean") deletes all non-numeric columns
@@ -939,9 +951,23 @@ class TimeSeriesDataFrame(pd.DataFrame):
             else:
                 aggregation[col] = agg_categorical
 
-        resampled_df = TimeSeriesDataFrame(
-            self.groupby(level=ITEMID, sort=False).resample(freq, level=TIMESTAMP, **kwargs).agg(aggregation)
-        )
+        def split_into_chunks(iterable: Iterable, size: int) -> Iterable[Iterable]:
+            # Based on https://stackoverflow.com/a/22045226/5497447
+            iterable = iter(iterable)
+            return iter(lambda: tuple(islice(iterable, size)), ())
+
+        def resample_chunk(chunk: Iterable[Tuple[str, pd.DataFrame]]) -> pd.DataFrame:
+            resampled_dfs = []
+            for item_id, df in chunk:
+                resampled_df = df.resample(offset, level=TIMESTAMP, **kwargs).agg(aggregation)
+                resampled_dfs.append(pd.concat({item_id: resampled_df}, names=[ITEMID]))
+            return pd.concat(resampled_dfs)
+
+        # Resampling time for 1 item < overhead time for a single parallel job. Therefore, we group items into chunks
+        # so that the speedup from parallelization isn't dominated by the communication costs.
+        chunks = split_into_chunks(pd.DataFrame(self).groupby(level=ITEMID, sort=False), chunk_size)
+        resampled_chunks = Parallel(n_jobs=num_cpus)(delayed(resample_chunk)(chunk) for chunk in chunks)
+        resampled_df = TimeSeriesDataFrame(pd.concat(resampled_chunks))
         resampled_df.static_features = self.static_features
         return resampled_df
 

autogluon/timeseries/models/gluonts/abstract_gluonts.py

@@ -283,6 +283,7 @@ class AbstractGluonTSModel(AbstractTimeSeriesModel):
         init_args = self._get_estimator_init_args()
 
         default_trainer_kwargs = {
+            "limit_val_batches": 3,
             "max_epochs": init_args["max_epochs"],
             "callbacks": init_args["callbacks"],
             "enable_progress_bar": False,

autogluon/timeseries/version.py

@@ -1,3 +1,3 @@
 """This is the autogluon version file."""
-__version__ = '1.0.0b20231124'
+__version__ = '1.0.0b20231125'
 __lite__ = False

{autogluon.timeseries-1.0.0b20231124.dist-info → autogluon.timeseries-1.0.0b20231125.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: autogluon.timeseries
-Version: 1.0.0b20231124
+Version: 1.0.0b20231125
 Summary: AutoML for Image, Text, and Tabular Data
 Home-page: https://github.com/autogluon/autogluon
 Author: AutoGluon Community
@@ -50,9 +50,9 @@ Requires-Dist: utilsforecast <0.0.11,>=0.0.10
 Requires-Dist: tqdm <5,>=4.38
 Requires-Dist: orjson ~=3.9
 Requires-Dist: tensorboard <3,>=2.9
-Requires-Dist: autogluon.core[raytune] ==1.0.0b20231124
-Requires-Dist: autogluon.common ==1.0.0b20231124
-Requires-Dist: autogluon.tabular[catboost,lightgbm,xgboost] ==1.0.0b20231124
+Requires-Dist: autogluon.core[raytune] ==1.0.0b20231125
+Requires-Dist: autogluon.common ==1.0.0b20231125
+Requires-Dist: autogluon.tabular[catboost,lightgbm,xgboost] ==1.0.0b20231125
 Provides-Extra: all
 Provides-Extra: tests
 Requires-Dist: pytest ; extra == 'tests'

{autogluon.timeseries-1.0.0b20231124.dist-info → autogluon.timeseries-1.0.0b20231125.dist-info}/RECORD

@@ -1,14 +1,14 @@
-autogluon.timeseries-1.0.0b20231124-py3.8-nspkg.pth,sha256=cQGwpuGPqg1GXscIwt-7PmME1OnSpD-7ixkikJ31WAY,554
+autogluon.timeseries-1.0.0b20231125-py3.8-nspkg.pth,sha256=cQGwpuGPqg1GXscIwt-7PmME1OnSpD-7ixkikJ31WAY,554
 autogluon/timeseries/__init__.py,sha256=_CrLLc1fkjen7UzWoO0Os8WZoHOgvZbHKy46I8v_4k4,304
 autogluon/timeseries/evaluator.py,sha256=l642tYfTHsl8WVIq_vV6qhgAFVFr9UuZD7gLra3A_Kc,250
 autogluon/timeseries/learner.py,sha256=HVfsoWTG3dXBCc7JbPfHCCYCMwL3zlrqHwLBG33MTJ8,9633
 autogluon/timeseries/predictor.py,sha256=sohEmnK0Z-sf7zhQRR6i7zTtuTigs0QXQrzhxKx8v9o,59016
 autogluon/timeseries/splitter.py,sha256=eghGwAAN2_cxGk5aJBILgjGWtLzjxJcytMy49gg_q18,3061
-autogluon/timeseries/version.py,sha256=-k59F7BtYG5KzVCW8NlMl325YMkn2027VY6iivBRmI4,90
+autogluon/timeseries/version.py,sha256=d9yJ5IbELS1blBNVeuaYKDWdfnjJtkgwPmPd9R_0wec,90
 autogluon/timeseries/configs/__init__.py,sha256=BTtHIPCYeGjqgOcvqb8qPD4VNX-ICKOg6wnkew1cPOE,98
 autogluon/timeseries/configs/presets_configs.py,sha256=1u6tbOKJdIRULYDu41dlJwXRNswWsjBDF0aR2YhyMQs,479
 autogluon/timeseries/dataset/__init__.py,sha256=UvnhAN5tjgxXTHoZMQDy64YMDj4Xxa68yY7NP4vAw0o,81
-autogluon/timeseries/dataset/ts_dataframe.py,sha256=gbYz6kwA6DRIPw2ijuWV4CneDPKQO_Zx6ildCSMfV2E,42929
+autogluon/timeseries/dataset/ts_dataframe.py,sha256=PgOz-88hbxNnhbpp0DMJbGBdtM6wIB32YpPWdyROB1c,44424
 autogluon/timeseries/metrics/__init__.py,sha256=gzvHptT-UdvB26CLOoFIznaKT-5FDwuVO37gaYPp88o,1835
 autogluon/timeseries/metrics/abstract.py,sha256=-muJuc30zSqHYXNBYyGocL-4zT7bt4SRjW9ddWcCq9w,8069
 autogluon/timeseries/metrics/point.py,sha256=WdhUrKB0ilO_N9-jHljQBQOj8mDvlNCfwMAD0RO61kI,11277
@@ -26,7 +26,7 @@ autogluon/timeseries/models/ensemble/__init__.py,sha256=kFr11Gmt7lQJu9Rr8HuIPphQ
 autogluon/timeseries/models/ensemble/abstract_timeseries_ensemble.py,sha256=tifETwmiEGt-YtQ9eNK7ojJ3fBvtFMUJvisbfkIJ7gw,3393
 autogluon/timeseries/models/ensemble/greedy_ensemble.py,sha256=3xYzg0CIe0U4l-HScVThb-q8wfKCmNB8SwRjRBMkCMU,7369
 autogluon/timeseries/models/gluonts/__init__.py,sha256=M8PV9ZE4WpteScMobXM6RH1Udb1AZiHHtj2g5GQL3TU,329
-autogluon/timeseries/models/gluonts/abstract_gluonts.py,sha256=t6nyLTcvkLYh_xYhHQGu4UK-c7fqdYguQrqzJT2j9Oo,25563
+autogluon/timeseries/models/gluonts/abstract_gluonts.py,sha256=cdzWbJ36vnSIg5TxzRYaOedvtUipbvQLQbsUSfj43ZA,25599
 autogluon/timeseries/models/gluonts/torch/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autogluon/timeseries/models/gluonts/torch/models.py,sha256=7ktOy6MxEzD0ykhUwcVEufSjdQNwYadtInLN6cms4Ig,18322
 autogluon/timeseries/models/local/__init__.py,sha256=JyckWWgMG1BTIWJqFTW6e1O-eb0LPPOwtXwmb1ErohQ,756
@@ -48,11 +48,11 @@ autogluon/timeseries/utils/datetime/base.py,sha256=MsqIHY14m3QMjSwwtE7Uo1oNwepWU
 autogluon/timeseries/utils/datetime/lags.py,sha256=kcU4liKbHj7KP2ajNU-KLZ8OYSU35EgT4kJjZNSw0Zg,5875
 autogluon/timeseries/utils/datetime/seasonality.py,sha256=kgK_ukw2wCviEB7CZXRVC5HZpBJZu9IsRrvCJ9E_rOE,755
 autogluon/timeseries/utils/datetime/time_features.py,sha256=pROkYyxETQ8rHKfPGhf2paB73C7rWJ2Ui0cCswLqbBg,2562
-autogluon.timeseries-1.0.0b20231124.dist-info/LICENSE,sha256=CeipvOyAZxBGUsFoaFqwkx54aPnIKEtm9a5u2uXxEws,10142
-autogluon.timeseries-1.0.0b20231124.dist-info/METADATA,sha256=S1f5aKQr741Y-QzUCMNAH6W1PEp5e7LNRSvz3g6SKbQ,13324
-autogluon.timeseries-1.0.0b20231124.dist-info/NOTICE,sha256=7nPQuj8Kp-uXsU0S5so3-2dNU5EctS5hDXvvzzehd7E,114
-autogluon.timeseries-1.0.0b20231124.dist-info/WHEEL,sha256=Xo9-1PvkuimrydujYJAjF7pCkriuXBpUPEjma1nZyJ0,92
-autogluon.timeseries-1.0.0b20231124.dist-info/namespace_packages.txt,sha256=giERA4R78OkJf2ijn5slgjURlhRPzfLr7waIcGkzYAo,10
-autogluon.timeseries-1.0.0b20231124.dist-info/top_level.txt,sha256=giERA4R78OkJf2ijn5slgjURlhRPzfLr7waIcGkzYAo,10
-autogluon.timeseries-1.0.0b20231124.dist-info/zip-safe,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
-autogluon.timeseries-1.0.0b20231124.dist-info/RECORD,,
+autogluon.timeseries-1.0.0b20231125.dist-info/LICENSE,sha256=CeipvOyAZxBGUsFoaFqwkx54aPnIKEtm9a5u2uXxEws,10142
+autogluon.timeseries-1.0.0b20231125.dist-info/METADATA,sha256=eFmR7JN0SI0xEjL_5c6lQRxNuhHNz8ClyNr0F46Od-w,13324
+autogluon.timeseries-1.0.0b20231125.dist-info/NOTICE,sha256=7nPQuj8Kp-uXsU0S5so3-2dNU5EctS5hDXvvzzehd7E,114
+autogluon.timeseries-1.0.0b20231125.dist-info/WHEEL,sha256=Xo9-1PvkuimrydujYJAjF7pCkriuXBpUPEjma1nZyJ0,92
+autogluon.timeseries-1.0.0b20231125.dist-info/namespace_packages.txt,sha256=giERA4R78OkJf2ijn5slgjURlhRPzfLr7waIcGkzYAo,10
+autogluon.timeseries-1.0.0b20231125.dist-info/top_level.txt,sha256=giERA4R78OkJf2ijn5slgjURlhRPzfLr7waIcGkzYAo,10
+autogluon.timeseries-1.0.0b20231125.dist-info/zip-safe,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+autogluon.timeseries-1.0.0b20231125.dist-info/RECORD,,