Merge branch 'master' into release/1.10.0

CHANGELOG.md

@@ -10,7 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Added
 - 
 - 
-- 
+- Jupyter extension for black ([#742](https://github.com/tinkoff-ai/etna/pull/742))
 - 
 - 
 - 

Makefile

@@ -1,4 +1,4 @@
-lint: isort-check black-check flake8-check mypy-check spell-check imported-deps-check
+lint: isort-check black-check flake8-check mypy-check spell-check imported-deps-check notebooks-check
 
 isort-check:
 	isort --skip etna/libs --sl -c etna/
@@ -22,11 +22,15 @@ spell-check:
 imported-deps-check:
 	python -m scripts.check_imported_dependencies
 
+notebooks-check:
+	black --check examples/*.ipynb
+
 format:
 	isort --skip etna/libs --sl etna/
 	isort --skip etna/libs --sl tests/
 	black etna/
 	black tests/
+	black examples/*.ipynb
 	flake8 --exclude etna/libs etna/
 	flake8 --exclude etna/libs tests/ --select E,W,C,F401,N
 	mypy

etna/analysis/outliers/hist_outliers.py

@@ -34,9 +34,9 @@ def optimal_sse(left: int, right: int, p: np.ndarray, pp: np.ndarray) -> float:
     """
     if left == 0:
         avg = p[right]
-        return pp[right] - avg ** 2 / (right - left + 1)
+        return pp[right] - avg**2 / (right - left + 1)
     avg = p[right] - p[left - 1]
-    return pp[right] - pp[left - 1] - avg ** 2 / (right - left + 1)
+    return pp[right] - pp[left - 1] - avg**2 / (right - left + 1)
 
 
 @numba.jit(nopython=True)

etna/clustering/distances/base.py

@@ -18,7 +18,7 @@
 class Distance(ABC, BaseMixin):
     """Base class for distances between series."""
 
-    def __init__(self, trim_series: bool = False, inf_value: float = sys.float_info.max // 10 ** 200):
+    def __init__(self, trim_series: bool = False, inf_value: float = sys.float_info.max // 10**200):
         """Init Distance.
 
         Parameters

etna/transforms/math/statistics.py

@@ -164,7 +164,7 @@ def transform(self, df: pd.DataFrame) -> pd.DataFrame:
             dataframe with results
         """
         window = self.window if self.window != -1 else len(df)
-        self._alpha_range = np.array([self.alpha ** i for i in range(window)])
+        self._alpha_range = np.array([self.alpha**i for i in range(window)])
         self._alpha_range = np.expand_dims(self._alpha_range, axis=0)  # (1, window)
         return super().transform(df)
 

examples/EDA.ipynb

@@ -43,6 +43,7 @@
    "outputs": [],
    "source": [
     "import warnings\n",
+    "\n",
     "warnings.filterwarnings(\"ignore\")"
    ]
   },
@@ -609,7 +610,7 @@
     "    distribution_plot,\n",
     "    sample_acf_plot,\n",
     "    sample_pacf_plot,\n",
-    "    plot_correlation_matrix\n",
+    "    plot_correlation_matrix,\n",
     ")"
    ]
   },
@@ -773,7 +774,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "lags = LagTransform(in_column=\"target\", lags=[1,7], out_column=\"lag\")\n",
+    "lags = LagTransform(in_column=\"target\", lags=[1, 7], out_column=\"lag\")\n",
     "ts.fit_transform([lags])"
    ]
   },
@@ -797,7 +798,7 @@
     }
    ],
    "source": [
-    "plot_correlation_matrix(ts, segments=[\"segment_a\",\"segment_b\"], method=\"spearman\", vmin=0.5, vmax=1)"
+    "plot_correlation_matrix(ts, segments=[\"segment_a\", \"segment_b\"], method=\"spearman\", vmin=0.5, vmax=1)"
    ]
   },
   {

examples/NN_examples.ipynb

@@ -52,6 +52,7 @@
     "from etna.models import SeasonalMovingAverageModel\n",
     "\n",
     "import warnings\n",
+    "\n",
     "warnings.filterwarnings(\"ignore\")"
    ]
   },
@@ -449,13 +450,17 @@
     "\n",
     "transform_date = DateFlagsTransform(day_number_in_week=True, day_number_in_month=False, out_column=\"dateflag\")\n",
     "num_lags = 10\n",
-    "transform_lag = LagTransform(in_column=\"target\", lags=[HORIZON+i for i in range(num_lags)], out_column=\"target_lag\")\n",
+    "transform_lag = LagTransform(\n",
+    "    in_column=\"target\",\n",
+    "    lags=[HORIZON + i for i in range(num_lags)],\n",
+    "    out_column=\"target_lag\",\n",
+    ")\n",
     "lag_columns = [f\"target_lag_{HORIZON+i}\" for i in range(num_lags)]\n",
     "\n",
     "transform_deepar = PytorchForecastingTransform(\n",
     "    max_encoder_length=HORIZON,\n",
     "    max_prediction_length=HORIZON,\n",
-    "    time_varying_known_reals=[\"time_idx\"]+lag_columns,\n",
+    "    time_varying_known_reals=[\"time_idx\"] + lag_columns,\n",
     "    time_varying_unknown_reals=[\"target\"],\n",
     "    time_varying_known_categoricals=[\"dateflag_day_number_in_week\"],\n",
     "    target_normalizer=GroupNormalizer(groups=[\"segment\"]),\n",
@@ -483,9 +488,11 @@
     "model_deepar = DeepARModel(max_epochs=150, learning_rate=[0.01], gpus=0, batch_size=64)\n",
     "metrics = [SMAPE(), MAPE(), MAE()]\n",
     "\n",
-    "pipeline_deepar = Pipeline(model=model_deepar,\n",
-    "                           horizon=HORIZON,\n",
-    "                           transforms=[transform_lag, transform_date, transform_deepar])"
+    "pipeline_deepar = Pipeline(\n",
+    "    model=model_deepar,\n",
+    "    horizon=HORIZON,\n",
+    "    transforms=[transform_lag, transform_date, transform_deepar],\n",
+    ")"
    ]
   },
   {
@@ -884,7 +891,11 @@
    "source": [
     "transform_date = DateFlagsTransform(day_number_in_week=True, day_number_in_month=False, out_column=\"dateflag\")\n",
     "num_lags = 10\n",
-    "transform_lag = LagTransform(in_column=\"target\", lags=[HORIZON+i for i in range(num_lags)], out_column=\"target_lag\")\n",
+    "transform_lag = LagTransform(\n",
+    "    in_column=\"target\",\n",
+    "    lags=[HORIZON + i for i in range(num_lags)],\n",
+    "    out_column=\"target_lag\",\n",
+    ")\n",
     "lag_columns = [f\"target_lag_{HORIZON+i}\" for i in range(num_lags)]\n",
     "\n",
     "transform_tft = PytorchForecastingTransform(\n",
@@ -909,9 +920,11 @@
     "\n",
     "model_tft = TFTModel(max_epochs=200, learning_rate=[0.01], gpus=0, batch_size=64)\n",
     "\n",
-    "pipeline_tft = Pipeline(model=model_tft,\n",
-    "                        horizon=HORIZON,\n",
-    "                        transforms=[transform_lag, transform_date, transform_tft])"
+    "pipeline_tft = Pipeline(\n",
+    "    model=model_tft,\n",
+    "    horizon=HORIZON,\n",
+    "    transforms=[transform_lag, transform_date, transform_tft],\n",
+    ")"
    ]
   },
   {
@@ -1326,11 +1339,9 @@
    "outputs": [],
    "source": [
     "model_sma = SeasonalMovingAverageModel(window=5, seasonality=7)\n",
-    "linear_trend_transform = LinearTrendTransform(in_column='target')\n",
+    "linear_trend_transform = LinearTrendTransform(in_column=\"target\")\n",
     "\n",
-    "pipeline_sma = Pipeline(model=model_sma,\n",
-    "                        horizon=HORIZON,\n",
-    "                        transforms=[linear_trend_transform])"
+    "pipeline_sma = Pipeline(model=model_sma, horizon=HORIZON, transforms=[linear_trend_transform])"
    ]
   },
   {

examples/backtest.ipynb

@@ -102,7 +102,7 @@
    "source": [
     "img = plt.imread(\"./assets/backtest/backtest.jpg\")\n",
     "plt.figure(figsize=(15, 10))\n",
-    "plt.axis('off')\n",
+    "plt.axis(\"off\")\n",
     "_ = plt.imshow(img)"
    ]
   },
@@ -198,7 +198,7 @@
    "source": [
     "df = pd.read_csv(\"./data/example_dataset.csv\")\n",
     "df[\"timestamp\"] = pd.to_datetime(df[\"timestamp\"])\n",
-    "df = df.loc[df.segment == 'segment_a']\n",
+    "df = df.loc[df.segment == \"segment_a\"]\n",
     "df.head()"
    ]
   },
@@ -218,7 +218,7 @@
    "outputs": [],
    "source": [
     "df = TSDataset.to_dataset(df)\n",
-    "ts = TSDataset(df, freq='D')"
+    "ts = TSDataset(df, freq=\"D\")"
    ]
   },
   {
@@ -249,7 +249,7 @@
     }
    ],
    "source": [
-    "ts.plot(segments=['segment_a'])"
+    "ts.plot(segments=[\"segment_a\"])"
    ]
   },
   {
@@ -275,9 +275,9 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "horizon = 31 # Set the horizon for predictions\n",
-    "model = ProphetModel() # Create a model\n",
-    "transforms = [] #A list of transforms -  we will not use any of them"
+    "horizon = 31  # Set the horizon for predictions\n",
+    "model = ProphetModel()  # Create a model\n",
+    "transforms = []  # A list of transforms -  we will not use any of them"
    ]
   },
   {
@@ -463,9 +463,7 @@
     }
    ],
    "source": [
-    "metrics_df, forecast_df, fold_info_df = pipeline.backtest(\n",
-    "    ts=ts, metrics=[MAE(), MSE(), SMAPE()]\n",
-    ")"
+    "metrics_df, forecast_df, fold_info_df = pipeline.backtest(ts=ts, metrics=[MAE(), MSE(), SMAPE()])"
    ]
   },
   {
@@ -931,9 +929,7 @@
    ],
    "source": [
     "metrics_df, forecast_df, fold_info_df = pipeline.backtest(\n",
-    "    ts=ts,\n",
-    "    metrics=[MAE(), MSE(), SMAPE()],\n",
-    "    aggregate_metrics=True\n",
+    "    ts=ts, metrics=[MAE(), MSE(), SMAPE()], aggregate_metrics=True\n",
     ")"
    ]
   },