fix: typing_extensions only with python < 3.10 (#134)

Author: baggiponte

docs/code/feature_engineering.py

@@ -1,25 +1,29 @@
-import polars as pl
 import numpy as np
-from functime.feature_extractors import FeatureExtractor, binned_entropy
+import polars as pl
+
+from functime.feature_extractors import binned_entropy
 
 # Load commodities price data
-y = pl.read_parquet("https://github.com/TracecatHQ/functime/raw/main/data/commodities.parquet")
+y = pl.read_parquet(
+    "https://github.com/TracecatHQ/functime/raw/main/data/commodities.parquet"
+)
 
 # Get column names ("commodity_type", "time", "price")
 entity_col, time_col, value_col = y.columns
 
 # Extract a single feature from a single time-series
 binned_entropy = binned_entropy(
-    pl.Series(np.random.normal(0, 1, size=10)),
-    bin_count=10
+    pl.Series(np.random.normal(0, 1, size=10)), bin_count=10
 )
 
 # 🔥 Also works on LazyFrames with query optimization
 features = (
-    pl.LazyFrame({
-        "index": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
-        "value": np.random.normal(0, 1, size=10)
-    })
+    pl.LazyFrame(
+        {
+            "index": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
+            "value": np.random.normal(0, 1, size=10),
+        }
+    )
     .select(
         binned_entropy=pl.col("value").ts.binned_entropy(bin_count=10),
         lempel_ziv_complexity=pl.col("value").ts.lempel_ziv_complexity(threshold=3),
@@ -30,13 +34,10 @@
 
 # 🚄 Extract features blazingly fast on many
 # stacked time-series using `group_by`
-features = (
-    y.group_by(entity_col)
-    .agg(
-        binned_entropy=pl.col(value_col).ts.binned_entropy(bin_count=10),
-        lempel_ziv_complexity=pl.col(value_col).ts.lempel_ziv_complexity(threshold=3),
-        longest_streak_above_mean=pl.col(value_col).ts.longest_streak_above_mean(),
-    )
+features = y.group_by(entity_col).agg(
+    binned_entropy=pl.col(value_col).ts.binned_entropy(bin_count=10),
+    lempel_ziv_complexity=pl.col(value_col).ts.lempel_ziv_complexity(threshold=3),
+    longest_streak_above_mean=pl.col(value_col).ts.longest_streak_above_mean(),
 )
 
 # 🚄 Extract features blazingly fast on windows
@@ -47,8 +48,7 @@
         time_col,
         every="12mo",
         by=entity_col,
-    )
-    .agg(
+    ).agg(
         binned_entropy=pl.col(value_col).ts.binned_entropy(bin_count=10),
         lempel_ziv_complexity=pl.col(value_col).ts.lempel_ziv_complexity(threshold=3),
         longest_streak_above_mean=pl.col(value_col).ts.longest_streak_above_mean(),

docs/notebooks/benchmarks.ipynb

@@ -30,10 +30,12 @@
    "outputs": [],
    "source": [
     "from typing import Callable\n",
+    "\n",
     "import pandas as pd\n",
     "import perfplot\n",
     "import polars as pl\n",
     "from tsfresh.feature_extraction import feature_calculators as tsfresh\n",
+    "\n",
     "from functime import feature_extractors as fe"
    ]
   },
@@ -104,7 +106,12 @@
     "    (fe.absolute_energy, tsfresh.abs_energy, {}, {}),\n",
     "    (fe.absolute_maximum, tsfresh.absolute_maximum, {}, {}),\n",
     "    (fe.absolute_sum_of_changes, tsfresh.absolute_sum_of_changes, {}, {}),\n",
-    "    (fe.lempel_ziv_complexity, tsfresh.lempel_ziv_complexity, {\"threshold\": (pl.col(\"value\").max() - pl.col(\"value\").min()) / 2}, {\"bins\": 2}),\n",
+    "    (\n",
+    "        fe.lempel_ziv_complexity,\n",
+    "        tsfresh.lempel_ziv_complexity,\n",
+    "        {\"threshold\": (pl.col(\"value\").max() - pl.col(\"value\").min()) / 2},\n",
+    "        {\"bins\": 2},\n",
+    "    ),\n",
     "    (\n",
     "        fe.approximate_entropy,\n",
     "        tsfresh.approximate_entropy,\n",
@@ -279,7 +286,11 @@
     "):\n",
     "    if f_feat.__name__ == \"approximate_entropy\":\n",
     "        n_range = [10_000]\n",
-    "    elif f_feat.__name__ in (\"number_cwt_peaks\", \"sample_entropy\", \"lempel_ziv_complexity\"):\n",
+    "    elif f_feat.__name__ in (\n",
+    "        \"number_cwt_peaks\",\n",
+    "        \"sample_entropy\",\n",
+    "        \"lempel_ziv_complexity\",\n",
+    "    ):\n",
     "        n_range = [10_000, 100_000]\n",
     "    else:\n",
     "        n_range = [10_000, 100_000, 1_000_000, 9_000_000]\n",

functime/base/forecaster.py

@@ -1,14 +1,19 @@
 import logging
+import sys
 from dataclasses import dataclass
-from typing import Callable, List, Mapping, Optional, Tuple, TypeVar, Union
+from typing import Callable, List, Literal, Mapping, Optional, Tuple, TypeVar, Union
 
 import polars as pl
-from typing_extensions import Literal, ParamSpec
 
 from functime.base.model import Model, ModelState
 from functime.base.transformer import Transformer
 from functime.ranges import make_future_ranges
 
+if sys.version_info < (3, 10):
+    from typing_extensions import ParamSpec
+else:
+    from typing import ParamSpec
+
 # The parameters of the Model
 P = ParamSpec("P")
 # The return type of the esimator's curried function

functime/base/transformer.py

@@ -1,12 +1,18 @@
 import inspect
+import sys
 from functools import cached_property, wraps
 from typing import Callable, Tuple, TypeVar, Union
 
 import polars as pl
-from typing_extensions import ParamSpec
 
 from functime.base.model import ModelState
 
+if sys.version_info < (3, 10):
+    from typing_extensions import ParamSpec
+else:
+    from typing import ParamSpec
+
+
 P = ParamSpec("P")  # The parameters of the Model
 R = TypeVar("R")
 

functime/evaluation.py

@@ -1,10 +1,9 @@
 from functools import partial
-from typing import List, Optional
+from typing import List, Literal, Optional
 
 import polars as pl
 import polars.selectors as cs
 from scipy.stats import norm, normaltest
-from typing_extensions import Literal
 
 from functime.base.metric import METRIC_TYPE
 from functime.metrics import (

functime/forecasting/_ar.py

@@ -1,10 +1,9 @@
 import logging
-from typing import Any, Callable, List, Mapping, Optional, Union
+from typing import Any, Callable, List, Literal, Mapping, Optional, Union
 
 import numpy as np
 import polars as pl
 from tqdm import tqdm, trange
-from typing_extensions import Literal
 
 from functime.cross_validation import expanding_window_split
 from functime.forecasting._evaluate import evaluate

functime/forecasting/_regressors.py

@@ -2,12 +2,11 @@
 Fit-predict regressors with special needs.
 """
 
-from typing import Callable, Optional, Union
+from typing import Callable, Literal, Optional, Union
 
 import numpy as np
 import polars as pl
 import sklearn
-from typing_extensions import Literal
 
 from functime.conversion import X_to_numpy, y_to_numpy
 from functime.preprocessing import PL_NUMERIC_COLS

functime/forecasting/automl.py

@@ -1,10 +1,9 @@
 from abc import abstractmethod
 from functools import partial
-from typing import Any, Mapping, Optional, Union
+from typing import Any, Literal, Mapping, Optional, Union
 
 import polars as pl
 from flaml import tune
-from typing_extensions import Literal
 
 from functime.base.forecaster import FORECAST_STRATEGIES, SUPPORTED_FREQ, Forecaster
 from functime.base.transformer import Transformer

functime/forecasting/elite.py

@@ -1,11 +1,10 @@
 from functools import partial
-from typing import Any, Mapping, Optional, Union
+from typing import Any, Literal, Mapping, Optional, Union
 
 import polars as pl
 import polars.selectors as cs
 from sklearn.linear_model import LassoLarsIC
 from tqdm import tqdm
-from typing_extensions import Literal
 
 from functime.backtesting import backtest
 from functime.base.forecaster import Forecaster

functime/forecasting/lance.py

@@ -1,10 +1,9 @@
-from typing import Optional
+from typing import Literal, Optional
 
 import lance
 import numpy as np
 import polars as pl
 from tqdm import tqdm
-from typing_extensions import Literal
 
 from functime.base import Forecaster
 from functime.forecasting._ar import fit_autoreg