orbit.constants package

Submodules

orbit.constants.constants module

class orbit.constants.constants.BacktestAnalyzeKeys(value)

Bases: enum.Enum

hash table keys for the dictionary of back-test aggregation analysis result

METRIC_GEO = 'metric_geo'
METRIC_NAME = 'metric_name'
METRIC_PER_BTMOD = 'metric_per_btmod'
METRIC_PER_HORIZON = 'metric_per_horizon'
class orbit.constants.constants.BacktestFitColumnNames(value)

Bases: enum.Enum

column names for the data frame of back-test fitting result

ACTUAL = 'actual'
FORECAST_DATES = 'forecast_dates'
PRED = 'pred'
PRED_HORIZON = 'pred_horizon'
TRAIN_END_DATE = 'train_end_date'
TRAIN_START_DATE = 'train_start_date'
class orbit.constants.constants.DateInfo(value)

Bases: enum.Enum

date_column: the data column name of the training/prediction data frame; starting_date: the date of first day of training data; format: yyyy-mm-dd date_interval: ‘day’, ‘week’, ‘month’

DATE_COLUMN = 'date_column'
DATE_COLUMN_NAME = 'date_column_name'
DATE_INTERVAL = 'date_interval'
END_DATE = 'end_date'
START_DATE = 'start_date'
class orbit.constants.constants.EstimatorOptionsMapper(value)

Bases: enum.Enum

Mapper for available options of a downstream input given an input upstream (within some other set of options)

ENGINE_TO_SAMPLE = {'pyro': ['map', 'vi'], 'stan': ['map', 'vi', 'mcmc']}
SAMPLE_TO_PREDICT = {'map': ['map'], 'mcmc': ['mean', 'median', 'full'], 'vi': ['mean', 'median', 'full']}
class orbit.constants.constants.InferMethod(value)

Bases: enum.Enum

The predict method for all of the stan models. Often used are mean and median.

MAP = 'map'
MARKOV_CHAIN_MONTE_CARLO = 'mcmc'
VARIATIONAL_INFERENCE = 'vi'
class orbit.constants.constants.PlotLabels(value)

Bases: enum.Enum

An enumeration.

ACTUAL_RESPONSE = 'actual_response'
PREDICTED_RESPONSE = 'predicted_response'
TRAINING_ACTUAL_RESPONSE = 'training_actual_response'
class orbit.constants.constants.PredictMethod(value)

Bases: enum.Enum

The predict method for all of the stan models. Often used are mean and median.

FULL_SAMPLING = 'full'
MAP = 'map'
MEAN = 'mean'
MEDIAN = 'median'
class orbit.constants.constants.PredictedComponents(value)

Bases: enum.Enum

column names for the data frame of predicted result with decomposed components

REGRESSION = 'regression'
SEASONALITY = 'seasonality'
TREND = 'trend'
class orbit.constants.constants.PredictionColumnNames(value)

Bases: enum.Enum

In the output of SLGTModel.transform() and SLGT.predict(), the column names if ‘return_decomposed_components’ = True.

ACTUAL_RESPONSE = 'actual'
LEVEL = 'level'
PREDICTED_RESPONSE = 'predicted'
REGRESSOR = 'regressor'
SEASONALITY = 'seasonality'
class orbit.constants.constants.StanModelKeys(value)

Bases: enum.Enum

All of the keys in the trained stan model from uTS. For example, for LGT/SLGT, the model is the output of SLGT.fit() and input of SLGTModel.

DATE_INFO = 'date_info'
MODELS = 'models'
REGRESSOR_COLUMNS = 'regressor_columns'
RESPONSE_COLUMN = 'response_column'
STAN_INPUTS = 'stan_inputs'
class orbit.constants.constants.TimeSeriesSplitSchemeNames(value)

Bases: enum.Enum

hash table keys for the dictionary of back-test meta data

MODEL = 'model'
TEST_IDX = 'test_idx'
TRAIN_END_DATE = 'train_end_date'
TRAIN_IDX = 'train_idx'
TRAIN_START_DATE = 'train_start_date'

orbit.constants.dlt module

class orbit.constants.dlt.BaseSamplingParameters(value)

Bases: enum.Enum

base parameters in posteriors sampling

LEVEL_SMOOTHING_FACTOR = 'lev_sm'
LOCAL_TREND = 'lt_sum'
LOCAL_TREND_LEVELS = 'l'
LOCAL_TREND_SLOPES = 'b'
RESIDUAL_DEGREE_OF_FREEDOM = 'nu'
RESIDUAL_SIGMA = 'obs_sigma'
SLOPE_SMOOTHING_FACTOR = 'slp_sm'
class orbit.constants.dlt.DataInputMapper(value)

Bases: enum.Enum

mapping from object input to stan file

AUTO_RIDGE_SCALE = 'AUTO_RIDGE_SCALE'
DAMPED_FACTOR = 'DAMPED_FACTOR'
LASSO_SCALE = 'LASSO_SCALE'
class orbit.constants.dlt.GlobalTrendOption(value)

Bases: enum.Enum

An enumeration.

flat = 3
linear = 0
logistic = 2
loglinear = 1
class orbit.constants.dlt.GlobalTrendSamplingParameters(value)

Bases: enum.Enum

An enumeration.

GLOBAL_TREND = 'gt_sum'
GLOBAL_TREND_LEVEL = 'gl'
GLOBAL_TREND_SLOPE = 'gb'
class orbit.constants.dlt.LatentSamplingParameters(value)

Bases: enum.Enum

latent variables to be sampled

INITIAL_SEASONALITY = 'init_sea'
REGRESSION_NEGATIVE_COEFFICIENTS = 'nr_beta'
REGRESSION_POSITIVE_COEFFICIENTS = 'pr_beta'
REGRESSION_REGULAR_COEFFICIENTS = 'rr_beta'
class orbit.constants.dlt.RegressionPenalty(value)

Bases: enum.Enum

An enumeration.

auto_ridge = 2
fixed_ridge = 0
lasso = 1
class orbit.constants.dlt.RegressionSamplingParameters(value)

Bases: enum.Enum

regression component related parameters in posteriors sampling

REGRESSION_COEFFICIENTS = 'beta'
class orbit.constants.dlt.SeasonalitySamplingParameters(value)

Bases: enum.Enum

seasonality component related parameters in posteriors sampling

SEASONALITY_LEVELS = 's'
SEASONALITY_SMOOTHING_FACTOR = 'sea_sm'

orbit.constants.lgt module

class orbit.constants.lgt.BaseSamplingParameters(value)

Bases: enum.Enum

base parameters in posteriors sampling

GLOBAL_TREND_COEF = 'gt_coef'
GLOBAL_TREND_POWER = 'gt_pow'
LEVEL_SMOOTHING_FACTOR = 'lev_sm'
LOCAL_GLOBAL_TREND_SUMS = 'lgt_sum'
LOCAL_TREND_COEF = 'lt_coef'
LOCAL_TREND_LEVELS = 'l'
LOCAL_TREND_SLOPES = 'b'
RESIDUAL_DEGREE_OF_FREEDOM = 'nu'
RESIDUAL_SIGMA = 'obs_sigma'
SLOPE_SMOOTHING_FACTOR = 'slp_sm'
class orbit.constants.lgt.DataInputMapper(value)

Bases: enum.Enum

mapping from object input to stan file

AUTO_RIDGE_SCALE = 'AUTO_RIDGE_SCALE'
LASSO_SCALE = 'LASSO_SCALE'
class orbit.constants.lgt.LatentSamplingParameters(value)

Bases: enum.Enum

latent variables to be sampled

INITIAL_SEASONALITY = 'init_sea'
REGRESSION_NEGATIVE_COEFFICIENTS = 'nr_beta'
REGRESSION_POSITIVE_COEFFICIENTS = 'pr_beta'
REGRESSION_REGULAR_COEFFICIENTS = 'rr_beta'
class orbit.constants.lgt.RegressionPenalty(value)

Bases: enum.Enum

An enumeration.

auto_ridge = 2
fixed_ridge = 0
lasso = 1
class orbit.constants.lgt.RegressionSamplingParameters(value)

Bases: enum.Enum

regression component related parameters in posteriors sampling

REGRESSION_COEFFICIENTS = 'beta'
class orbit.constants.lgt.SeasonalitySamplingParameters(value)

Bases: enum.Enum

seasonality component related parameters in posteriors sampling

SEASONALITY_LEVELS = 's'
SEASONALITY_SMOOTHING_FACTOR = 'sea_sm'

orbit.constants.palette module

class orbit.constants.palette.QualitativePalette(value)

Bases: enum.Enum

Palette for visualizing discrete categorical data

Bar5 = ['#ef476fff', '#ffd166ff', '#06d6a0ff', '#118ab2ff', '#073b4cff']
Line4 = ['#e6c72b', '#2be669', '#2b4ae6', '#e62ba8']
PostQ = ['#1fc600', '#ff4500']
Rainbow8 = ['#ffadadff', '#ffd6a5ff', '#fdffb6ff', '#caffbfff', '#9bf6ffff', '#a0c4ffff', '#bdb2ffff', '#ffc6ffff']
Stack = ['#12939A', '#F15C17', '#DDB27C', '#88572C', '#FF991F', '#DA70BF', '#125C77', '#4DC19C', '#776E57', '#17B8BE', '#F6D18A', '#B7885E', '#FFCB99', '#F89570', '#829AE3', '#E79FD5', '#1E96BE', '#89DAC1', '#B3AD9E']

Module contents