Skip to content

Commit

Permalink
Merge pull request #72 from Nixtla/feat/add_date_features
Browse files Browse the repository at this point in the history 
[FEAT] Add date features argument
  • Loading branch information
@AzulGarza
AzulGarza authored Aug 10, 2023
2 parents 846979b + 7626202 commit fe40b7b
Show file tree
Hide file tree
Showing 3 changed files with 784 additions and 124 deletions.
737 changes: 613 additions & 124 deletions nbs/timegpt.ipynb
View file

Large diffs are not rendered by default.

6 changes: 6 additions & 0 deletions nixtlats/_modidx.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,13 +7,19 @@
'lib_path': 'nixtlats'},
'syms': { 'nixtlats.timegpt': { 'nixtlats.timegpt.TimeGPT': ('timegpt.html#timegpt', 'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT.__init__': ('timegpt.html#timegpt.__init__', 'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._add_date_features': ( 'timegpt.html#timegpt._add_date_features',
'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._compute_date_feature': ( 'timegpt.html#timegpt._compute_date_feature',
'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._get_to_dict_args': ( 'timegpt.html#timegpt._get_to_dict_args',
'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._hit_multi_series_endpoint': ( 'timegpt.html#timegpt._hit_multi_series_endpoint',
'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._hit_multi_series_historic_endpoint': ( 'timegpt.html#timegpt._hit_multi_series_historic_endpoint',
'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._infer_freq': ('timegpt.html#timegpt._infer_freq', 'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._make_future_dataframe': ( 'timegpt.html#timegpt._make_future_dataframe',
'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._model_params': ('timegpt.html#timegpt._model_params', 'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._multi_series': ('timegpt.html#timegpt._multi_series', 'nixtlats/timegpt.py'),
'nixtlats.timegpt.TimeGPT._parse_response': ( 'timegpt.html#timegpt._parse_response',
Expand Down
165 changes: 165 additions & 0 deletions nixtlats/timegpt.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,13 +8,64 @@
import inspect
import json
import requests
import warnings
from typing import Dict, List, Optional, Union

import numpy as np
import pandas as pd

logger = logging.getLogger()

# %% ../nbs/timegpt.ipynb 7
date_features_by_freq = {
# Daily frequencies
"B": ["year", "month", "day", "weekday"],
"C": ["year", "month", "day", "weekday"],
"D": ["year", "month", "day", "weekday"],
# Weekly
"W": ["year", "week", "weekday"],
# Monthly
"M": ["year", "month"],
"SM": ["year", "month", "day"],
"BM": ["year", "month"],
"CBM": ["year", "month"],
"MS": ["year", "month"],
"SMS": ["year", "month", "day"],
"BMS": ["year", "month"],
"CBMS": ["year", "month"],
# Quarterly
"Q": ["year", "quarter"],
"BQ": ["year", "quarter"],
"QS": ["year", "quarter"],
"BQS": ["year", "quarter"],
# Yearly
"A": ["year"],
"Y": ["year"],
"BA": ["year"],
"BY": ["year"],
"AS": ["year"],
"YS": ["year"],
"BAS": ["year"],
"BYS": ["year"],
# Hourly
"BH": ["year", "month", "day", "hour", "weekday"],
"H": ["year", "month", "day", "hour"],
# Minutely
"T": ["year", "month", "day", "hour", "minute"],
"min": ["year", "month", "day", "hour", "minute"],
# Secondly
"S": ["year", "month", "day", "hour", "minute", "second"],
# Milliseconds
"L": ["year", "month", "day", "hour", "minute", "second", "millisecond"],
"ms": ["year", "month", "day", "hour", "minute", "second", "millisecond"],
# Microseconds
"U": ["year", "month", "day", "hour", "minute", "second", "microsecond"],
"us": ["year", "month", "day", "hour", "minute", "second", "microsecond"],
# Nanoseconds
"N": [],
}

# %% ../nbs/timegpt.ipynb 8
class TimeGPT:
"""
A class used to interact with the TimeGPT API.
Expand Down Expand Up @@ -144,6 +195,79 @@ def _resample_dataframe(
resampled_df["ds"] = resampled_df["ds"].astype(str)
return resampled_df

def _compute_date_feature(self, dates, feature):
if callable(feature):
feat_name = feature.__name__
feat_vals = feature(dates)
else:
feat_name = feature
if feature in ("week", "weekofyear"):
dates = dates.isocalendar()
feat_vals = getattr(dates, feature)
vals = np.asarray(feat_vals)
return feat_name, vals

def _make_future_dataframe(
self, df: pd.DataFrame, h: int, freq: str, reconvert: bool = True
):
last_dates = df.groupby("unique_id")["ds"].max()

def _future_date_range(last_date):
future_dates = pd.date_range(last_date, freq=freq, periods=h + 1)[-h:]
return future_dates

future_df = last_dates.apply(_future_date_range).reset_index()
future_df = future_df.explode("ds").reset_index(drop=True)
if reconvert and df.dtypes["ds"] == "object":
# avoid date 000
future_df["ds"] = future_df["ds"].astype(str)
return future_df

def _add_date_features(
self,
df: pd.DataFrame,
X_df: Optional[pd.DataFrame],
h: int,
freq: str,
date_features: List[str],
date_features_to_one_hot: Optional[List[str]],
):
# df contains exogenous variables
# X_df are the future values of the exogenous variables
# construct dates
train_dates = df["ds"].unique().tolist()
# if we dont have future exogenos variables
# we need to compute the future dates
if X_df is None:
X_df = self._make_future_dataframe(df=df, h=h, freq=freq)
future_dates = X_df["ds"].unique().tolist()
dates = pd.DatetimeIndex(train_dates + future_dates)
date_features_df = pd.DataFrame({"ds": dates})
for feature in date_features:
feat_name, feat_vals = self._compute_date_feature(dates, feature)
date_features_df[feat_name] = feat_vals
if df.dtypes["ds"] == "object":
date_features_df["ds"] = date_features_df["ds"].astype(str)
if date_features_to_one_hot is not None:
date_features_df = pd.get_dummies(
date_features_df,
columns=date_features_to_one_hot,
dtype=int,
)
# remove duplicated columns if any
date_features_df = date_features_df.drop(
columns=[
col
for col in date_features_df.columns
if col in df.columns and col not in ["unique_id", "ds"]
]
)
# add date features to df
df = df.merge(date_features_df, on="ds", how="left")
# add date features to X_df
X_df = X_df.merge(date_features_df, on="ds", how="left")
return df, X_df

def _preprocess_dataframes(
self,
df: pd.DataFrame,
Expand Down Expand Up @@ -285,9 +409,37 @@ def _multi_series(
finetune_steps: int,
clean_ex_first: bool,
add_history: bool,
date_features: Union[bool, List[str]],
date_features_to_one_hot: Union[bool, List[str]],
):
if freq is None:
freq = self._infer_freq(df)
# add date features logic
if isinstance(date_features, bool):
if date_features:
date_features = date_features_by_freq.get(freq)
if date_features is None:
warnings.warn(
f"Non default date features for {freq} "
"please pass a list of date features"
)
else:
date_features = None

if date_features is not None:
if isinstance(date_features_to_one_hot, bool):
if date_features_to_one_hot:
date_features_to_one_hot = date_features
else:
date_features_to_one_hot = None
df, X_df = self._add_date_features(
df=df,
X_df=X_df,
h=h,
freq=freq,
date_features=date_features,
date_features_to_one_hot=date_features_to_one_hot,
)
Y_df, X_df, x_cols = self._preprocess_dataframes(
df=df,
h=h,
Expand Down Expand Up @@ -328,6 +480,8 @@ def forecast(
clean_ex_first: bool = True,
validate_token: bool = False,
add_history: bool = False,
date_features: Union[bool, List[str]] = False,
date_features_to_one_hot: Union[bool, List[str]] = True,
):
"""Forecast your time series using TimeGPT.
Expand Down Expand Up @@ -371,6 +525,15 @@ def forecast(
sending requests.
add_history : bool (default=False)
Return fitted values of the model.
date_features : bool or list of str or callable, optional (default=False)
Features computed from the dates.
Can be pandas date attributes or functions that will take the dates as input.
If True automatically adds most used date features for the
frequency of `df`.
date_features_to_one_hot : bool or list of str (default=True)
Apply one-hot encoding to these date features.
If `date_features=True`, then all date features are
one-hot encoded by default.
Returns
-------
Expand Down Expand Up @@ -399,6 +562,8 @@ def forecast(
finetune_steps=finetune_steps,
clean_ex_first=clean_ex_first,
add_history=add_history,
date_features=date_features,
date_features_to_one_hot=date_features_to_one_hot,
)
fcst_df = self._validate_outputs(
fcst_df=fcst_df,
Expand Down

0 comments on commit fe40b7b

Please sign in to comment.