Code fragments of helper.py

• import pandas as pd
• from numpy import dot
• import numpy as np
• from numpy.linalg import norm
• import json
• import itertools
• def normalize(data):
• ''' Normalizes input data to be between 0 and 1 '''
• min_val = np.nanmin(data)
• max_val = np.nanmax(data)
• range_val = max_val - min_val
• if range_val == 0:
• return data
• # min max are the same
• # normalize only non-NaN values
• normalized_data = np.where(np.isnan(data), data, (data - min_val) / range_val)
• return normalized_data
• def one_hot_encoding(df, enc_col):
• ''' One hot encodes specified columns and adds them to the input df '''
• enc_df = pd.get_dummies(df[enc_col])
• enc_df.reset_index(drop = True, inplace = True)
• return pd.concat([df, enc_df], axis = 1)
• def cosine_sim(v1,v2):
• ''' Calculates the cosine similarity between two vectors '''
• return dot(v1,v2)/(norm(v1)*norm(v2))
• def show_welcome_text():
• text = ''' ################################## # SUPPLIER RECOMMENDATION SYSTEM # ################################## '''
• print(text)
• def get_combinations():
• ''' Computes all possible combinations of params in parameters.json file '''
• with open('parameters.json', 'r') as file:
• params = json.load(file)
• combinations_values = list(itertools.product(*params.values()))
• combinations = [dict(zip(params.keys(), combination)) for combination in combinations_values]
• return combinations
• def preprocess(df): # encode categorical features as a one-hot numeric array
• print("\nDataFrame Merged with Input Order:\n", df)
• if 'volume' in df.columns:
• df = one_hot_encoding(df = df, enc_col = 'volume')
• if 'product_type' in df.columns:
• df = one_hot_encoding(df = df, enc_col = 'product_type')
• if 'CO2_report' in df.columns:
• df = one_hot_encoding(df = df, enc_col = 'CO2_report')
• if 'transport_mode' in df.columns:
• df = one_hot_encoding(df = df, enc_col = 'transport_mode')
• # drop columns that are no longer used
• cols = ['supplier_name', 'rating', 'product_type', 'volume', 'CO2_report','transport_mode']
• # filter out the columns that not exist
• cols_to_drop = [col for col in cols if col in df.columns]
• df.drop(columns = cols_to_drop, inplace = True)
• print("\nNormalized/One-Hot Encoded DataFrame:\n", df)
• return df

Graph