interactionsMatrixGenerator.py

from scipy import sparse
from collections import defaultdict

import glob
import csv
import sys  
import io
import numpy

# later fix this to look at all browse files
# customer_id,contact_id,product_id,event_date,url
# path = "fordocker/browse_production_siteId_=35569/*.csv"
# map_customerId_productIdList= {}
# for fname in glob.glob(path):

# allUsers = {}
allProducts = set()

map_customerId_productIdList = defaultdict(list)


# if True:
#     fname = "fordocker/browse_production_siteId_=35569/part-r-00016-55b1cd2d-d2c7-43dc-ac2a-da953f82d47b.csv"
path = "fordocker/browse_production_siteId_=35569/*.csv"
for fname in glob.glob(path):
    print(fname)
    with open(fname, mode='r') as infile:
        reader = csv.reader(infile)
        next(reader, None)
        for row in reader:
#             print(row)
            customer_id = row[0]
            product_id = row[2]
            allProducts.add(product_id)
            map_customerId_productIdList[customer_id].append(product_id)

#ok now i have lists of products per user ... now what .... get set of users and set of products.  for amounts.

allUsers = list(set(map_customerId_productIdList.keys()))
allProducts = list(allProducts)

map_userId_index = {}
map_index_userId = {}

for i, userId in enumerate(allUsers):
    map_userId_index[userId] = i
    map_index_userId[i] = userId

map_productId_index = {}
map_index_productId = {}

for i, productId in enumerate(allProducts):
    map_productId_index[productId] = i
    map_index_productId[i] = productId
    

interactions = numpy.zeros(shape=(len(allUsers),len(allProducts)))

for userIndex, userId in enumerate(allUsers):
    productsList = map_customerId_productIdList[userId]
#     print(userIndex,  " ", userId) 
#     print(productsList)
    for productId in productsList:
#         print(productId)
        productIndex = map_productId_index[productId]
        interactions[userIndex][productIndex] += 1

#now convert to coo sparse matrix

interactions_sparse_coo_matrix = sparse.csr_matrix(interactions)