File size: 6,290 Bytes
67b1c6c |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 |
import pandas as pd
import ujson as json
import gc
import numpy as np
from concurrent.futures import ProcessPoolExecutor
import multiprocessing as mp
from pymongo import MongoClient
from collections import defaultdict
from pathlib import Path
# def read_json_parallel(file_path, num_workers=None):
# """Read JSON file using parallel processing"""
# if num_workers is None:
# num_workers = max(1, mp.cpu_count() - 1)
# print(f"Reading {file_path}...")
# # Read chunks and concatenate them into a single DataFrame
# df = pd.read_json(file_path, lines=True, dtype_backend="pyarrow", chunksize=100000)
# return next(df)
def read_data_mongo(file_path, num_workers=None):
"""Read JSON file using parallel processing"""
if num_workers is None:
num_workers = max(1, mp.cpu_count() - 1)
print(f"Reading {file_path}...")
conn_str = "mongodb://Mtalha:[email protected]/"
client = MongoClient(conn_str)
databases = client.list_database_names()
db_client=client["Yelp"]
# Read the entire file at once since chunksize isn't needed for parallel reading here
# Use 'records' orient if your JSON was saved with this format
try:
collection = db_client[file_path]
documents = collection.find({}, {"_id": 0})
data = list(documents)
final_dict=defaultdict(list)
for dictt in data:
for k,v in dictt.items():
final_dict[k].append(v)
df=pd.DataFrame(final_dict)
# df = pd.read_json(file_path, orient='records', dtype_backend="pyarrow")
except Exception as e:
# If 'records' doesn't work, try without specifying orient or with 'split'
# This is a fallback for different JSON structures
# df = pd.read_json(file_path, dtype_backend="pyarrow")
print("ERROR WHILE READING FILES FORM MONGODB AS : ",e)
print(f"Finished reading. DataFrame shape: {df.shape}")
return df
def process_datasets(output_path,filename):
# File paths
file_paths = {
'business': "yelp_academic_dataset_business",
'checkin': "yelp_academic_dataset_checkin",
'review': "yelp_academic_dataset_review",
'tip': "yelp_academic_dataset_tip",
'user': "yelp_academic_dataset_user",
'google': "google_review_dataset"
}
# Read datasets with progress tracking
print("Reading datasets...")
dfs = {}
for name, path in file_paths.items():
print(f"Processing {name} dataset...")
dfs[name] = read_data_mongo(path)
print(f"Finished reading {name} dataset. Shape: {dfs[name].shape}")
print("All files read. Starting column renaming...")
# Rename columns to avoid conflicts
# Reviews
dfs['review'] = dfs['review'].rename(columns={
'date': 'review_date',
'stars': 'review_stars',
'text': 'review_text',
'useful': 'review_useful',
'funny': 'review_funny',
'cool': 'review_cool'
})
# print("COLUMNS IN REVIEW DAFRA)
# Tips
dfs['tip'] = dfs['tip'].rename(columns={
'date': 'tip_date',
'text': 'tip_text',
'compliment_count': 'tip_compliment_count'
})
# Checkins
dfs['checkin'] = dfs['checkin'].rename(columns={
'date': 'checkin_date'
})
# Users
dfs['user'] = dfs['user'].rename(columns={
'name': 'user_name',
'review_count': 'user_review_count',
'useful': 'user_useful',
'funny': 'user_funny',
'cool': 'user_cool'
})
# Business
dfs['business'] = dfs['business'].rename(columns={
'name': 'business_name',
'stars': 'business_stars',
'review_count': 'business_review_count'
})
dfs['google'] = dfs['google'].rename(columns={
'name': 'business_name',
'stars': 'business_stars',
'review_count': 'business_review_count'
})
df_business_final= dfs['business']
df_google_final=dfs['google']
df_review_final=dfs['review']
df_tip_final=dfs['tip']
df_checkin_final=dfs['checkin']
df_user_final=dfs['user']
df_business_final=pd.concat([df_business_final,df_google_final],axis=0)
df_business_final.reset_index(drop=True,inplace=True)
print("Starting merge process...")
# Merge process with memory management
print("Step 1: Starting with reviews...")
merged_df = df_review_final
print("Step 2: Merging with business data...")
merged_df = merged_df.merge(
df_business_final,
on='business_id',
how='left'
)
print("Step 3: Merging with user data...")
merged_df = merged_df.merge(
df_user_final,
on='user_id',
how='left'
)
print("Step 4: Merging with checkin data...")
merged_df = merged_df.merge(
df_checkin_final,
on='business_id',
how='left'
)
print("Step 5: Aggregating and merging tip data...")
tip_agg = df_tip_final.groupby('business_id').agg({
'tip_compliment_count': 'sum',
'tip_text': 'count'
}).rename(columns={'tip_text': 'tip_count'})
merged_df = merged_df.merge(
tip_agg,
on='business_id',
how='left'
)
print("Filling NaN values...")
merged_df['tip_count'] = merged_df['tip_count'].fillna(0)
merged_df['tip_compliment_count'] = merged_df['tip_compliment_count'].fillna(0)
merged_df['checkin_date'] = merged_df['checkin_date'].fillna('')
merged_df["friends"].fillna(0,inplace=True)
for col in merged_df.columns:
if merged_df[col].isnull().sum()>0:
print(f" {col} has {merged_df[col].isnull().sum()} null values")
print("Shape of Merged Dataset is : ",merged_df.shape)
output_file = Path(output_path) / filename
print("COLUMNS BEFORE PREPROCESING")
print()
print(merged_df.info())
for col in merged_df.columns:
for v in merged_df[col]:
print(f"Type of values in {col} is {type(v)} and values are like : {v}")
break
merged_df.to_csv(output_file,index=False)
return merged_df
# if __name__ == "__main__":
# process_datasets()
|