Spaces:
Sleeping
Sleeping
File size: 6,618 Bytes
e8d4213 816790f e8d4213 816790f e8d4213 816790f e8d4213 816790f e8d4213 d0f2767 e8d4213 816790f e8d4213 816790f e8d4213 816790f e8d4213 816790f e8d4213 816790f e8d4213 816790f d0f2767 816790f e8d4213 816790f d0f2767 816790f e8d4213 d0f2767 e8d4213 d0f2767 816790f d0f2767 816790f e8d4213 d0f2767 e8d4213 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 |
import json
import joblib
import pandas as pd
from sklearn.preprocessing import StandardScaler
import numpy as np
class VAVPipeline:
"""
A class representing a Variable Air Volume (VAV) pipeline.
Attributes:
rtu_id (int): The ID of the RTU (Roof Top Unit).
scaler_path (str): The path to the scaler file.
window_size (int): The size of the sliding window.
Methods:
get_scaler(scaler_path): Loads the scaler from the given path.
get_window(df): Returns the sliding window of the given dataframe.
transform_window(df_window): Transforms the values of the dataframe using the scaler.
prepare_input(df_trans): Prepares the input for the model.
get_input_output(df): Extracts the input and output column names from the dataframe.
extract_data_from_message(message): Extracts data from the message payload and returns a dataframe.
fit(message): Fits the model with the extracted data and returns the prepared input and transformed data.
"""
def __init__(self, rtu_id, scaler_path=None, window_size=30):
"""
Initializes a VAVPipeline object.
Args:
rtu_id (int): The ID of the RTU (Roof Top Unit).
scaler_path (str, optional): The path to the scaler file. Defaults to None.
window_size (int, optional): The size of the sliding window. Defaults to 30.
"""
self.get_cols = True
self.window_size = window_size
self.rtu_id = rtu_id
if rtu_id == 1:
self.zones = [69, 68, 67, 66, 65, 64, 42, 41, 40, 39, 38, 37, 36]
if rtu_id == 2:
self.zones = [72, 71, 63, 62, 60, 59, 58,57, 50, 49, 44, 43, 35, 34, 33, 32, 31, 30, 29, 28]
self.output_col_names = []
self.input_col_names = [
f"rtu_00{rtu_id}_fltrd_sa_flow_tn",
f"rtu_00{rtu_id}_sa_temp",
"air_temp_set_1",
"air_temp_set_2",
"dew_point_temperature_set_1d",
"relative_humidity_set_1",
"solar_radiation_set_1",
]
self.column_names = self.output_col_names + self.input_col_names
self.num_inputs = len(self.input_col_names)
self.num_outputs = len(self.output_col_names)
if scaler_path:
self.scaler = self.get_scaler(scaler_path)
def get_scaler(self, scaler_path):
"""
Loads the scaler from the given path.
Args:
scaler_path (str): The path to the scaler file.
Returns:
StandardScaler: The loaded scaler object.
"""
return joblib.load(scaler_path)
def get_window(self, df):
"""
Returns the sliding window of the given dataframe.
Args:
df (pd.DataFrame): The dataframe.
Returns:
pd.DataFrame: The sliding window dataframe.
"""
len_df = len(df)
if len_df > self.window_size:
return df[len_df - (self.window_size + 1) : len_df].astype("float32")
else:
return None
def transform_window(self, df_window):
"""
Transforms the values of the dataframe using the scaler.
Args:
df_window (pd.DataFrame): The dataframe.
Returns:
np.ndarray: The transformed values.
"""
return self.scaler.transform(df_window.values)
def prepare_input(self, df_trans):
"""
Prepares the input for the model.
Args:
df_trans (np.ndarray): The transformed values.
Returns:
np.ndarray: The prepared input.
"""
return df_trans[: self.window_size, :].reshape(
(1, self.window_size, len(self.column_names))
)
def get_input_output(self, df: pd.DataFrame):
"""
Extracts the input and output column names from the dataframe.
Args:
df (pd.DataFrame): The dataframe.
"""
for zone in self.zones:
for column in df.columns:
if (
f"zone_0{zone}" in column
and "co2" not in column
and "hw_valve" not in column
and "cooling_sp" not in column
and "heating_sp" not in column
):
self.output_col_names.append(column)
self.input_col_names = [
f"rtu_00{self.rtu_id}_fltrd_sa_flow_tn",
f"rtu_00{self.rtu_id}_sa_temp",
"air_temp_set_1",
"air_temp_set_2",
"dew_point_temperature_set_1d",
"relative_humidity_set_1",
"solar_radiation_set_1",
]
for zone in self.zones:
for column in df.columns:
if f"zone_0{zone}" in column:
if "cooling_sp" in column or "heating_sp" in column:
self.input_col_names.append(column)
self.column_names = self.output_col_names + self.input_col_names
self.num_inputs = len(self.input_col_names)
self.num_outputs = len(self.output_col_names)
self.df = pd.DataFrame(columns=self.column_names)
def extract_data_from_message(self, df: pd.DataFrame):
"""
Extracts data from the message payload and returns a dataframe.
Args:
message: The message containing the payload.
Returns:
pd.DataFrame: The extracted data as a dataframe.
"""
if self.get_cols == True:
self.get_input_output(df)
self.get_cols = False
df = df[self.column_names]
len_df = len(self.df)
if len_df != 0:
self.df = pd.concat([self.df, df], axis=0)
else:
self.df = df
if len_df > 31:
self.df = self.df.iloc[len_df - 31 : len_df]
self.df.loc[len_df] = self.df.mean()
return self.df
else:
return None
def fit(self, df: pd.DataFrame):
"""
Fits the model with the extracted data and returns the prepared input and transformed data.
Args:
message: The message containing the data.
Returns:
tuple: A tuple containing the prepared input and transformed data.
"""
df_window = self.extract_data_from_message(df)
if df_window is not None:
df_trans = self.transform_window(df_window)
df_new = self.prepare_input(df_trans)
else:
df_new = None
df_trans = None
return df_new, df_trans
|