openai api generated code for ETNA shortcuts

[martins0n]

🚀 Feature Request

As a newbie you can just prompt feature names and model just from general knowledge and get proper pipeline or forecast.

Proposal

import os
import ast
import openai
from hydra_slayer import get_from_params

from etna.auto.pool.utils import fill_template

openai.api_key = os.getenv("OPENAI_API_KEY")


SYSTEM_CONTEXT = """

Our task to generate pipeline with model and transforms for time series forecasting.
All possible models and transforms you is here in ```PROTOCOLS AND DESCRIPTIONS: ...``` section:

PROTOCOLS AND DESCRIPTIONS:
```
class PerSegmentModelMixin(ModelForecastingMixin):def __init__(self, base_model: Any):
class MultiSegmentModelMixin(ModelForecastingMixin):def __init__(self, base_model: Any):
class DeadlineMovingAverageModel(NonPredictionIntervalContextRequiredAbstractModel,):def __init__(self, window: int = 3, seasonality: str = "month"):
class SeasonalMovingAverageModel(NonPredictionIntervalContextRequiredAbstractModel,):def __init__(self, window: int = 5, seasonality: int = 7):
class SimpleExpSmoothingModel(PerSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self,initialization_method: str = "estimated",initial_level: Optional[float] = None,smoothing_level: Optional[float] = None,**fit_kwargs,):
class HoltWintersModel(PerSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self,trend: Optional[str] = None,damped_trend: bool = False,seasonal: Optional[str] = None,seasonal_periods: Optional[int] = None,initialization_method: str = "estimated",initial_level: Optional[float] = None,initial_trend: Optional[float] = None,initial_seasonal: Optional[Sequence[float]] = None,use_boxcox: Union[bool, str, float] = False,bounds: Optional[Dict[str, Tuple[float, float]]] = None,dates: Optional[Sequence[datetime]] = None,freq: Optional[str] = None,missing: str = "none",smoothing_level: Optional[float] = None,smoothing_trend: Optional[float] = None,smoothing_seasonal: Optional[float] = None,damping_trend: Optional[float] = None,**fit_kwargs,):
class HoltModel(PerSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self,exponential: bool = False,damped_trend: bool = False,initialization_method: str = "estimated",initial_level: Optional[float] = None,initial_trend: Optional[float] = None,smoothing_level: Optional[float] = None,smoothing_trend: Optional[float] = None,damping_trend: Optional[float] = None,**fit_kwargs,):
class LinearPerSegmentModel(PerSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self, fit_intercept: bool = True, **kwargs):
class LinearMultiSegmentModel(MultiSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self, fit_intercept: bool = True, **kwargs):
class ElasticMultiSegmentModel(MultiSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self, alpha: float = 1.0, l1_ratio: float = 0.5, fit_intercept: bool = True, **kwargs):
class ElasticPerSegmentModel(PerSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self, alpha: float = 1.0, l1_ratio: float = 0.5, fit_intercept: bool = True, **kwargs):
class CatBoostPerSegmentModel(PerSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self,iterations: Optional[int] = None,depth: Optional[int] = None,learning_rate: Optional[float] = None,logging_level: Optional[str] = "Silent",l2_leaf_reg: Optional[float] = None,thread_count: Optional[int] = None,**kwargs,):
class CatBoostMultiSegmentModel(MultiSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self,iterations: Optional[int] = None,depth: Optional[int] = None,learning_rate: Optional[float] = None,logging_level: Optional[str] = "Silent",l2_leaf_reg: Optional[float] = None,thread_count: Optional[int] = None,**kwargs,):
class SARIMAXModel(PerSegmentModelMixin, PredictionIntervalContextIgnorantModelMixin, PredictionIntervalContextIgnorantAbstractModel):def __init__(self,order: Tuple[int, int, int] = (1, 0, 0),seasonal_order: Tuple[int, int, int, int] = (0, 0, 0, 0),trend: Optional[str] = None,measurement_error: bool = False,time_varying_regression: bool = False,mle_regression: bool = True,simple_differencing: bool = False,enforce_stationarity: bool = True,enforce_invertibility: bool = True,hamilton_representation: bool = False,concentrate_scale: bool = False,trend_offset: float = 1,use_exact_diffuse: bool = False,dates: Optional[List[datetime]] = None,freq: Optional[str] = None,missing: str = "none",validate_specification: bool = True,**kwargs,):
class MovingAverageModel(SeasonalMovingAverageModel):def __init__(self, window: int = 5):
class NaiveModel(SeasonalMovingAverageModel):def __init__(self, lag: int = 1):
class ProphetModel(PerSegmentModelMixin, PredictionIntervalContextIgnorantModelMixin, PredictionIntervalContextIgnorantAbstractModel):def __init__(self,growth: str = "linear",changepoints: Optional[List[datetime]] = None,n_changepoints: int = 25,changepoint_range: float = 0.8,yearly_seasonality: Union[str, bool] = "auto",weekly_seasonality: Union[str, bool] = "auto",daily_seasonality: Union[str, bool] = "auto",holidays: Optional[pd.DataFrame] = None,seasonality_mode: str = "additive",seasonality_prior_scale: float = 10.0,holidays_prior_scale: float = 10.0,changepoint_prior_scale: float = 0.05,mcmc_samples: int = 0,interval_width: float = 0.8,uncertainty_samples: Union[int, bool] = 1000,stan_backend: Optional[str] = None,additional_seasonality_params: Iterable[Dict[str, Union[str, float, int]]] = (),):
class BATSModel(PerSegmentModelMixin, PredictionIntervalContextIgnorantModelMixin, PredictionIntervalContextIgnorantAbstractModel):def __init__(self,use_box_cox: Optional[bool] = None,box_cox_bounds: Tuple[int, int] = (0, 1),use_trend: Optional[bool] = None,use_damped_trend: Optional[bool] = None,seasonal_periods: Optional[Iterable[int]] = None,use_arma_errors: bool = True,show_warnings: bool = True,n_jobs: Optional[int] = None,multiprocessing_start_method: str = "spawn",context: Optional[ContextInterface] = None,):
class TBATSModel(PerSegmentModelMixin, PredictionIntervalContextIgnorantModelMixin, PredictionIntervalContextIgnorantAbstractModel):def __init__(self,use_box_cox: Optional[bool] = None,box_cox_bounds: Tuple[int, int] = (0, 1),use_trend: Optional[bool] = None,use_damped_trend: Optional[bool] = None,seasonal_periods: Optional[Iterable[int]] = None,use_arma_errors: bool = True,show_warnings: bool = True,n_jobs: Optional[int] = None,multiprocessing_start_method: str = "spawn",context: Optional[ContextInterface] = None,):
class SklearnPerSegmentModel(PerSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self, regressor: RegressorMixin):
class SklearnMultiSegmentModel(MultiSegmentModelMixin,NonPredictionIntervalContextIgnorantModelMixin,NonPredictionIntervalContextIgnorantAbstractModel,):def __init__(self, regressor: RegressorMixin):
class AutoARIMAModel(PerSegmentModelMixin, PredictionIntervalContextIgnorantModelMixin, PredictionIntervalContextIgnorantAbstractModel):def __init__(self,**kwargs,):
class DeepBaseNet(DeepAbstractNet, LightningModule):def __init__(self):
class DeepBaseModel(DeepBaseAbstractModel, SaveNNMixin, NonPredictionIntervalContextRequiredAbstractModel):def __init__(self,*,net: DeepBaseNet,encoder_length: int,decoder_length: int,train_batch_size: int,test_batch_size: int,trainer_params: Optional[dict],train_dataloader_params: Optional[dict],test_dataloader_params: Optional[dict],val_dataloader_params: Optional[dict],split_params: Optional[dict],):
class MLPNet(DeepBaseNet):def __init__(self,input_size: int,hidden_size: List[int],lr: float,loss: "torch.nn.Module",optimizer_params: Optional[dict],) -> None:super().__init__()self.input_size = input_sizeself.hidden_size = hidden_sizeself.lr = lrself.loss = lossself.optimizer_params = {} if optimizer_params is None else optimizer_paramslayers = [nn.Linear(in_features=input_size, out_features=hidden_size[0]), nn.ReLU()]for i in range(1, len(hidden_size)):
class MLPModel(DeepBaseModel):def __init__(self,input_size: int,decoder_length: int,hidden_size: List,encoder_length: int = 0,lr: float = 1e-3,loss: Optional["torch.nn.Module"] = None,train_batch_size: int = 16,test_batch_size: int = 16,optimizer_params: Optional[dict] = None,trainer_params: Optional[dict] = None,train_dataloader_params: Optional[dict] = None,test_dataloader_params: Optional[dict] = None,val_dataloader_params: Optional[dict] = None,split_params: Optional[dict] = None,):
class PytorchForecastingDatasetBuilder(BaseMixin):def __init__(self,max_encoder_length: int = 30,min_encoder_length: Optional[int] = None,min_prediction_idx: Optional[int] = None,min_prediction_length: Optional[int] = None,max_prediction_length: int = 1,static_categoricals: Optional[List[str]] = None,static_reals: Optional[List[str]] = None,time_varying_known_categoricals: Optional[List[str]] = None,time_varying_known_reals: Optional[List[str]] = None,time_varying_unknown_categoricals: Optional[List[str]] = None,time_varying_unknown_reals: Optional[List[str]] = None,variable_groups: Optional[Dict[str, List[int]]] = None,constant_fill_strategy: Optional[Dict[str, Union[str, float, int, bool]]] = None,allow_missing_timesteps: bool = True,lags: Optional[Dict[str, List[int]]] = None,add_relative_time_idx: bool = True,add_target_scales: bool = True,add_encoder_length: Union[bool, str] = True,target_normalizer: Union[NORMALIZER, str, List[NORMALIZER], Tuple[NORMALIZER]] = "auto",categorical_encoders: Optional[Dict[str, NaNLabelEncoder]] = None,scalers: Optional[Dict[str, Union[StandardScaler, RobustScaler, TorchNormalizer, EncoderNormalizer]]] = None,):
class TFTModel(_DeepCopyMixin, PytorchForecastingMixin, SaveNNMixin, PredictionIntervalContextRequiredAbstractModel):def __init__(self,decoder_length: Optional[int] = None,encoder_length: Optional[int] = None,dataset_builder: Optional[PytorchForecastingDatasetBuilder] = None,train_batch_size: int = 64,test_batch_size: int = 64,lr: float = 1e-3,hidden_size: int = 16,lstm_layers: int = 1,attention_head_size: int = 4,dropout: float = 0.1,hidden_continuous_size: int = 8,loss: "MultiHorizonMetric" = None,trainer_params: Optional[Dict[str, Any]] = None,quantiles_kwargs: Optional[Dict[str, Any]] = None,**kwargs,):
class RNNNet(DeepBaseNet):def __init__(self,input_size: int,num_layers: int,hidden_size: int,lr: float,loss: "torch.nn.Module",optimizer_params: Optional[dict],) -> None:super().__init__()self.num_layers = num_layersself.input_size = input_sizeself.hidden_size = hidden_sizeself.loss = torch.nn.MSELoss() if loss is None else lossself.rnn = nn.LSTM(num_layers=self.num_layers, hidden_size=self.hidden_size, input_size=self.input_size, batch_first=True)self.projection = nn.Linear(in_features=self.hidden_size, out_features=1)self.lr = lrself.optimizer_params = {} if optimizer_params is None else optimizer_paramsdef forward(self, x: RNNBatch, *args, **kwargs):
class RNNModel(DeepBaseModel):def __init__(self,input_size: int,decoder_length: int,encoder_length: int,num_layers: int = 2,hidden_size: int = 16,lr: float = 1e-3,loss: Optional["torch.nn.Module"] = None,train_batch_size: int = 16,test_batch_size: int = 16,optimizer_params: Optional[dict] = None,trainer_params: Optional[dict] = None,train_dataloader_params: Optional[dict] = None,test_dataloader_params: Optional[dict] = None,val_dataloader_params: Optional[dict] = None,split_params: Optional[dict] = None,):
class DeepARModel(_DeepCopyMixin, PytorchForecastingMixin, SaveNNMixin, PredictionIntervalContextRequiredAbstractModel):def __init__(self,decoder_length: Optional[int] = None,encoder_length: Optional[int] = None,dataset_builder: Optional[PytorchForecastingDatasetBuilder] = None,train_batch_size: int = 64,test_batch_size: int = 64,lr: float = 1e-3,cell_type: str = "LSTM",hidden_size: int = 10,rnn_layers: int = 2,dropout: float = 0.1,loss: Optional["DistributionLoss"] = None,trainer_params: Optional[Dict[str, Any]] = None,quantiles_kwargs: Optional[Dict[str, Any]] = None,):
class LDS(BaseMixin):def __init__(self,emission_coeff: Tensor,# (batch_size, seq_length, latent_dim)transition_coeff: Tensor,# (latent_dim, latent_dim)innovation_coeff: Tensor,# (batch_size, seq_length, latent_dim)noise_std: Tensor,# (batch_size, seq_length, 1)prior_mean: Tensor,# (batch_size, latent_dim)prior_cov: Tensor,# (batch_size, latent_dim, latent_dim)offset: Tensor,# (batch_size, seq_length, 1)seq_length: int,latent_dim: int,):
class CompositeSSM(SSM):def __init__(self, seasonal_ssms: List[SeasonalitySSM], nonseasonal_ssm: Optional[Union[LevelSSM, LevelTrendSSM]] = None):
class DaylySeasonalitySSM(SeasonalitySSM):def __init__(self):
class WeeklySeasonalitySSM(SeasonalitySSM):def __init__(self):
class YearlySeasonalitySSM(SeasonalitySSM):def __init__(self):
class SeasonalitySSM(LevelSSM):def __init__(self, num_seasons: int, timestamp_transform: Callable[[pd.Timestamp], int]):
class DeepStateModel(DeepBaseModel):def __init__(self,ssm: CompositeSSM,input_size: int,encoder_length: int,decoder_length: int,num_layers: int = 1,n_samples: int = 5,lr: float = 1e-3,train_batch_size: int = 16,test_batch_size: int = 16,optimizer_params: Optional[dict] = None,trainer_params: Optional[dict] = None,train_dataloader_params: Optional[dict] = None,test_dataloader_params: Optional[dict] = None,val_dataloader_params: Optional[dict] = None,split_params: Optional[dict] = None,):
class DeepStateNet(DeepBaseNet):def __init__(self,ssm: CompositeSSM,input_size: int,num_layers: int,n_samples: int,lr: float,optimizer_params: Optional[dict],):
class ReversiblePerSegmentWrapper(PerSegmentWrapper, ReversibleTransform):def __init__(self, transform: OneSegmentTransform, required_features: Union[Literal["all"], List[str]]):
class Transform(SaveMixin, BaseMixin):def __init__(self, required_features: Union[Literal["all"], List[str]]):
class IrreversibleTransform(Transform):def __init__(self, required_features: Union[Literal["all"], List[str]]):
class PerSegmentWrapper(Transform):def __init__(self, transform: OneSegmentTransform, required_features: Union[Literal["all"], List[str]]):
class ReversibleTransform(Transform):def __init__(self, required_features: Union[Literal["all"], List[str]]):
class IrreversiblePerSegmentWrapper(PerSegmentWrapper, IrreversibleTransform):def __init__(self, transform: OneSegmentTransform, required_features: Union[Literal["all"], List[str]]):
class ResampleWithDistributionTransform(IrreversiblePerSegmentWrapper):def __init__(self, in_column: str, distribution_column: str, inplace: bool = True, out_column: Optional[str] = None):
class TimeSeriesImputerTransform(ReversibleTransform):def __init__(self,in_column: str = "target",strategy: str = ImputerMode.constant,window: int = -1,seasonality: int = 1,default_value: Optional[float] = None,constant_value: float = 0,):
class FilterFeaturesTransform(ReversibleTransform):def __init__(self,include: Optional[Sequence[str]] = None,exclude: Optional[Sequence[str]] = None,return_features: bool = False,):
class MRMRFeatureSelectionTransform(BaseFeatureSelectionTransform):def __init__(self,relevance_table: RelevanceTable,top_k: int,features_to_use: Union[List[str], Literal["all"]] = "all",fast_redundancy: bool = False,relevance_aggregation_mode: str = AggregationMode.mean,redundancy_aggregation_mode: str = AggregationMode.mean,atol: float = 1e-10,return_features: bool = False,**relevance_params,):
class TreeFeatureSelectionTransform(BaseFeatureSelectionTransform):def __init__(self,model: Union[Literal["catboost"], Literal["random_forest"], TreeBasedRegressor],top_k: int,features_to_use: Union[List[str], Literal["all"]] = "all",return_features: bool = False,):
class SegmentGaleShapley(BaseGaleShapley):def __init__(self, name: str, ranked_candidates: List[str]):
class GaleShapleyFeatureSelectionTransform(BaseFeatureSelectionTransform):def __init__(self,relevance_table: RelevanceTable,top_k: int,features_to_use: Union[List[str], Literal["all"]] = "all",use_rank: bool = False,return_features: bool = False,**relevance_params,):
class BaseGaleShapley(BaseMixin):def __init__(self, name: str, ranked_candidates: List[str]):
class GaleShapleyMatcher(BaseMixin):def __init__(self, segments: List[SegmentGaleShapley], features: List[FeatureGaleShapley]):
class BaseFeatureSelectionTransform(ReversibleTransform, ABC):def __init__(self, features_to_use: Union[List[str], Literal["all"]] = "all", return_features: bool = False):
class LogTransform(ReversibleTransform):def __init__(self, in_column: str, base: int = 10, inplace: bool = True, out_column: Optional[str] = None):
class LagTransform(IrreversibleTransform, FutureMixin):def __init__(self, in_column: str, lags: Union[List[int], int], out_column: Optional[str] = None):
class ExogShiftTransform(IrreversibleTransform, FutureMixin):def __init__(self, lag: Union[int, Literal["auto"]], horizon: Optional[int] = None):
class StandardScalerTransform(SklearnTransform):def __init__(self,in_column: Optional[Union[str, List[str]]] = None,inplace: bool = True,out_column: Optional[str] = None,with_mean: bool = True,with_std: bool = True,mode: Union[TransformMode, str] = "per-segment",):
class RobustScalerTransform(SklearnTransform):def __init__(self,in_column: Optional[Union[str, List[str]]] = None,inplace: bool = True,out_column: Optional[str] = None,with_centering: bool = True,with_scaling: bool = True,quantile_range: Tuple[float, float] = (25, 75),unit_variance: bool = False,mode: Union[TransformMode, str] = "per-segment",):
class MaxAbsScalerTransform(SklearnTransform):def __init__(self,in_column: Optional[Union[str, List[str]]] = None,inplace: bool = True,out_column: Optional[str] = None,mode: Union[TransformMode, str] = "per-segment",):
class MinMaxScalerTransform(SklearnTransform):def __init__(self,in_column: Optional[Union[str, List[str]]] = None,inplace: bool = True,out_column: Optional[str] = None,feature_range: Tuple[float, float] = (0, 1),clip: bool = True,mode: Union[TransformMode, str] = "per-segment",):
class DifferencingTransform(ReversibleTransform):def __init__(self,in_column: str,period: int = 1,order: int = 1,inplace: bool = True,out_column: Optional[str] = None,):
class LambdaTransform(ReversibleTransform):def __init__(self,in_column: str,transform_func: Callable[[pd.DataFrame], pd.DataFrame],inplace: bool = True,out_column: Optional[str] = None,inverse_transform_func: Optional[Callable[[pd.DataFrame], pd.DataFrame]] = None,):
class MADTransform(WindowStatisticsTransform):def __init__(self,in_column: str,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,):
class MeanTransform(WindowStatisticsTransform):def __init__(self,in_column: str,window: int,seasonality: int = 1,alpha: float = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,):
class StdTransform(WindowStatisticsTransform):def __init__(self,in_column: str,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,ddof: int = 1,):
class WindowStatisticsTransform(IrreversibleTransform, ABC):def __init__(self,in_column: str,out_column: str,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,**kwargs,):
class SumTransform(WindowStatisticsTransform):def __init__(self,in_column: str,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,):
class MinTransform(WindowStatisticsTransform):def __init__(self,in_column: str,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,):
class MinMaxDifferenceTransform(WindowStatisticsTransform):def __init__(self,in_column: str,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,):
class MaxTransform(WindowStatisticsTransform):def __init__(self,in_column: str,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,):
class MedianTransform(WindowStatisticsTransform):def __init__(self,in_column: str,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,):
class QuantileTransform(WindowStatisticsTransform):def __init__(self,in_column: str,quantile: float,window: int,seasonality: int = 1,min_periods: int = 1,fillna: float = 0,out_column: Optional[str] = None,):
class SklearnTransform(ReversibleTransform):def __init__(self,in_column: Optional[Union[str, List[str]]],out_column: Optional[str],transformer: TransformerMixin,inplace: bool = True,mode: Union[TransformMode, str] = "per-segment",):
class YeoJohnsonTransform(SklearnTransform):def __init__(self,in_column: Optional[Union[str, List[str]]] = None,inplace: bool = True,out_column: Optional[str] = None,standardize: bool = True,mode: Union[TransformMode, str] = "per-segment",):
class BoxCoxTransform(SklearnTransform):def __init__(self,in_column: Optional[Union[str, List[str]]] = None,inplace: bool = True,out_column: Optional[str] = None,standardize: bool = True,mode: Union[TransformMode, str] = "per-segment",):
class AddConstTransform(ReversibleTransform):def __init__(self, in_column: str, value: float, inplace: bool = True, out_column: Optional[str] = None):
class OutliersTransform(ReversibleTransform, ABC):def __init__(self, in_column: str):
class DensityOutliersTransform(OutliersTransform):def __init__(self,in_column: str,window_size: int = 15,distance_coef: float = 3,n_neighbors: int = 3,distance_func: Callable[[float, float], float] = absolute_difference_distance,):
class PredictionIntervalOutliersTransform(OutliersTransform):def __init__(self,in_column: str,model: Union[Literal["prophet"], Literal["sarimax"], Type["ProphetModel"], Type["SARIMAXModel"]],interval_width: float = 0.95,**model_kwargs,):
class MedianOutliersTransform(OutliersTransform):def __init__(self, in_column: str, window_size: int = 10, alpha: float = 3):
class LabelEncoderTransform(IrreversibleTransform):def __init__(self, in_column: str, out_column: Optional[str] = None, strategy: str = ImputerMode.mean):
class OneHotEncoderTransform(IrreversibleTransform):def __init__(self, in_column: str, out_column: Optional[str] = None):
class DateFlagsTransform(IrreversibleTransform, FutureMixin):def __init__(self,day_number_in_week: Optional[bool] = True,day_number_in_month: Optional[bool] = True,day_number_in_year: Optional[bool] = False,week_number_in_month: Optional[bool] = False,week_number_in_year: Optional[bool] = False,month_number_in_year: Optional[bool] = False,season_number: Optional[bool] = False,year_number: Optional[bool] = False,is_weekend: Optional[bool] = True,special_days_in_week: Sequence[int] = (),special_days_in_month: Sequence[int] = (),out_column: Optional[str] = None,):
class TimeFlagsTransform(IrreversibleTransform, FutureMixin):def __init__(self,minute_in_hour_number: bool = True,fifteen_minutes_in_hour_number: bool = False,hour_number: bool = True,half_hour_number: bool = False,half_day_number: bool = False,one_third_day_number: bool = False,out_column: Optional[str] = None,):
class FourierTransform(IrreversibleTransform, FutureMixin):def __init__(self,period: float,order: Optional[int] = None,mods: Optional[Sequence[int]] = None,out_column: Optional[str] = None,):
class SpecialDaysTransform(IrreversiblePerSegmentWrapper, FutureMixin):def __init__(self, find_special_weekday: bool = True, find_special_month_day: bool = True):
class HolidayTransform(IrreversibleTransform, FutureMixin):def __init__(self, iso_code: str = "RUS", out_column: Optional[str] = None):
class TheilSenTrendTransform(ReversiblePerSegmentWrapper):def __init__(self, in_column: str, poly_degree: int = 1, **regression_params):
class LinearTrendTransform(ReversiblePerSegmentWrapper):def __init__(self, in_column: str, poly_degree: int = 1, **regression_params):
class STLTransform(ReversiblePerSegmentWrapper):def __init__(self,in_column: str,period: int,model: Union[str, TimeSeriesModel] = "arima",robust: bool = False,model_kwargs: Optional[Dict[str, Any]] = None,stl_kwargs: Optional[Dict[str, Any]] = None,):
class ChangePointsTrendTransform(ReversibleChangePointsTransform):def __init__(self,in_column: str,change_points_model: Optional[BaseChangePointsModelAdapter] = None,per_interval_model: Optional[PerIntervalModel] = None,):
class TrendTransform(IrreversibleChangePointsTransform):def __init__(self,in_column: str,change_points_model: Optional[BaseChangePointsModelAdapter] = None,per_interval_model: Optional[PerIntervalModel] = None,out_column: Optional[str] = None,):
class ChangePointsSegmentationTransform(IrreversibleChangePointsTransform):def __init__(self,in_column: str,change_points_model: Optional[BaseChangePointsModelAdapter] = None,out_column: Optional[str] = None,):
class MedianPerIntervalModel(StatisticsPerIntervalModel):def __init__(self):
class MeanPerIntervalModel(StatisticsPerIntervalModel):def __init__(self):
class StatisticsPerIntervalModel(PerIntervalModel):def __init__(self, statistics_function: Callable[[np.ndarray], float]):
class ConstantPerIntervalModel(PerIntervalModel):def __init__(self):
class SklearnPreprocessingPerIntervalModel(PerIntervalModel):def __init__(self, preprocessing: TransformerMixin):
class SklearnRegressionPerIntervalModel(PerIntervalModel):def __init__(self, model: Optional[RegressorMixin] = None):
class RupturesChangePointsModel(BaseChangePointsModelAdapter):def __init__(self, change_points_model: BaseEstimator, **change_points_model_predict_params):
```


I'll give you some examples of prompts and expected results you should generate in ```EXAMPLES: ...``` section:

```
EXAMPLES:


Example 1:

given prompt:
a need simple model
expected result:
{'_target_': 'etna.pipeline.Pipeline', 'horizon': '${__aux__.horizon}', 'model': {'_target_': 'etna.models.NaiveModel', 'lag': 1}}

Example 2:

given prompt:
let's try gradient boosting with lag features, segment encoding and some date flags
expected result:
{'_target_': 'etna.pipeline.Pipeline', 'horizon': '${__aux__.horizon}', 'model': {'_target_': 'etna.models.CatBoostMultiSegmentModel'}, 'transforms': [{'_target_': 'etna.transforms.LagTransform', 'in_column': 'target', 'lags': '${shift:${horizon},[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]}'}, {'_target_': 'etna.transforms.SegmentEncoderTransform'}, {'_target_': 'etna.transforms.DateFlagsTransform', 'day_number_in_week': True, 'is_weekend': True, 'week_number_in_year': True}]}

Example 3:

given prompt:
boosting with data drift detection
expected result:
{'model': {'_target_': 'etna.models.catboost.CatBoostMultiSegmentModel'}, 'transforms': [{'in_column': 'target', 'change_points_model': {'change_points_model': {'_target_': 'ruptures.detection.binseg.Binseg'}, 'change_points_model_predict_params': {'n_bkps': 5}, '_target_': 'etna.transforms.decomposition.change_points_based.change_points_models.ruptures_based.RupturesChangePointsModel'}, '_target_': 'etna.transforms.decomposition.change_points_based.segmentation.ChangePointsSegmentationTransform'}], 'horizon': '${__aux__.horizon}', '_target_': 'etna.pipeline.pipeline.Pipeline'}

```

RULES:

* We can use only and only models and transforms from ```PROTOCOLS: ...``` section.
* I insist on using models and transorms from ```PROTOCOLS: ...``` section!!!!!
* It's all case sensitive, so be careful.
* In 'model' field classes from etna.models
* In 'transforms' field classes from etna.transforms
* You should respect protocol of each model and transform
* We need to use __aux__.horizon variable in 'horizon' field and shifts for lagged features
"""

def etnaGPT(prompt: str, horizon: int):
    completion = openai.ChatCompletion.create(
        model="gpt-3.5-turbo-16k",
        messages=[
            {"role": "system", "content": SYSTEM_CONTEXT},
            {"role": "user", "content": prompt},
        ],
        temperature=0,
    )
    message = completion.choices[0].message["content"] 
    message = ast.literal_eval(message)
        
    message = fill_template(message, {"horizon": horizon})
        
    pipe = get_from_params(**message)
    
    return pipe


print(etnaGPT("Regression with lags, data drift detection and date features", 14))

Test cases

No response

Additional context

No response