cereniyim

def convert_target_to_array(target):
    # function to convert target df
    # to an array
    target_array = (np
                    .array(target)
                    .reshape((-1, )))
    return target_array

cereniyim

def convert_features_to_array(features):
    # function to convert feature df
    # to an array
    num_rows = len(features)
    num_cols = len(features.columns)
    
    features_array = (np
                      .array(features)
                      .reshape((num_rows, 
                                num_cols)))

cereniyim

# manipulation libraries
import pandas as pd
import numpy as np

# visualization libraries
import matplotlib.pyplot as plt
import seaborn as sns

%matplotlib inline
# to display visuals in the notebook

cereniyim

def ImputeMissingValues(train_df, test_df):
    # separete non-NA cols
    is_features = [col for col in train_df.columns
                   if col.find("is_") != -1]
    interim_train_1 = train_df[is_features]
    interim_test_1 = test_df[is_features]

    # impute taster_name NA with 0 as "Unknown"
    constant_impute = ImputeWithConstant(train_df,
                                         test_df)

cereniyim

def ImputeWithMostFrequent(train_df, test_df, 
                           cols=["country", "province", 
                                 "region_1", "variety"]):
  # function to impute country, province, region_1, variety
  # columns with the most_frequent value of each feature
  # most_frequent imputer is fitted on train dataset
  # transformation done on the train and test set

  train_df = pd.DataFrame(train_df[cols])
  test_df = pd.DataFrame(test_df[cols]) 

cereniyim

def ImputeWithMedian(train_df, test_df, cols=["price", "year"]):
  # function to impute price and year
  # columns with the median value of each 
  # median imputer is fitted on train dataset
  # transformation done on the train and test set

  train_df = pd.DataFrame(train_df[cols])
  test_df = pd.DataFrame(test_df[cols]) 
    
  median_imputer = SimpleImputer(strategy="median")

cereniyim

def ImputeWithConstant(train_df, test_df, cols=["taster_name"]):
  # function to impute taster_name
  # with 0 stands for "Unknown value"

  train_df = pd.DataFrame(train_df[cols])
  test_df = pd.DataFrame(test_df[cols])

  constant_imputer = SimpleImputer(strategy="constant", fill_value=0)

  constant_imputer.fit(train_df)

cereniyim

def EncodeCategoricalData(train_df, 
                          test_df, 
                          cols=ordinal_encode_cols):
  # funtion to create ordinal encoder object and
  # assigns -1 to the unseen labels of the test set
  # returns ordinal encoded train and test datasets
  # respectively
  ordinal_encoder = OrdinalEncoder(cols=ordinal_encode_cols,
                                   return_df=True,
                                   handle_unknown="value",

cereniyim

def ExtractFeatures(df):
  # adds is_red, is_white, is_rose, is_rose
  # is_sparkling, is_dry, is_sweet
  for key, value in desc_extracting_dict.items():
    interim_df = extract_features_from_description(df, "description", key, value)

  # adds year
  interim_df["title_numlist"] = (interim_df
                                 .title
                                 .str

cereniyim

def extract_blend_from_variety(variety):
  # function to find if a variety is a blend
  # ASSUMPTION: There is no NA values
  # in the variety feature
    
  if (variety.find("-") != -1) | (variety.find("Blend") != -1):
      if variety == "Xarel-lo":
          return 0
      else:
          return 1