def __init__(self):
if Predictor.__instance is not None:
raise Exception("I'm Singleton")
else:
Predictor.__instance = self
self.cfg = Config.get_instance()
self.expenses = Expenses.get_instance()
def predict():
expenses = Expenses.get_instance()
if expenses.data is None:
predictor = Predictor.get_instance()
predictor.read_data()
filteredData = filter_predict_window(expenses.data)
result = filteredData[['Description', 'DueDate', 'Where']]
return curate_result(result, processType='predict')
def usage():
expenses = Expenses.get_instance()
if expenses.data is None:
predictor = Predictor.get_instance()
predictor.read_data()
# shopData.sort_values(by=['iDate'], ascending=False, inplace=True)
result = expenses.data[['Date', 'Description', 'Category', 'Amount', 'Where']]
return curate_result(result, processType='usage')
def predict_item(item):
expenses = Expenses.get_instance()
if expenses.data is None:
predictor = Predictor.get_instance()
predictor.read_data()
shopData = expenses.data[(expenses.data['Description'] == item)]
filteredData = filter_predict_window(shopData)
result = filteredData[['Description', 'DueDate']]
return curate_result(result, item=item, processType='predict')
def usage_item(item):
expenses = Expenses.get_instance()
if expenses.data is None:
predictor = Predictor.get_instance()
predictor.read_data()
shopData = expenses.data[(expenses.data['Description'] == item)]
result = shopData[['Date', 'Description', 'Where', 'Category', 'Amount']]
# result.sort_values(by='Date', ascending=False)
return curate_result(result, item=item, processType='usage')
def predict_shop(shop):
expenses = Expenses.get_instance()
if expenses.data is None:
predictor = Predictor.get_instance()
predictor.read_data()
shopData = expenses.data[(expenses.data['Where'] == shop)]
filteredData = filter_predict_window(shopData)
# print(filteredData)
result = filteredData[['Description', 'DueDate']]
return curate_result(result, shop=shop, processType='predict')
def usage_shop_item(shop, item):
expenses = Expenses.get_instance()
if expenses.data is None:
predictor = Predictor.get_instance()
predictor.read_data()
shopData = expenses.data[(expenses.data['Where'] == shop) & (expenses.data['Description'] == item)]
# print(shopData)
# shopData.sort_values(by=['iDate'], ascending=False, inplace=True)
result = shopData[['Date', 'Description', 'Category', 'Amount']]
return curate_result(result, shop=shop, item=item, processType='usage')
class Predictor:
log = Logger.get_instance()
log.log(LogLevel.VERBOSE, 'Started.')
expenses = Expenses.get_instance()
log.log(LogLevel.VERBOSE, 'Config Instance fetched.')
__instance = None
def __init__(self):
if Predictor.__instance is not None:
raise Exception("I'm Singleton")
else:
Predictor.__instance = self
self.cfg = Config.get_instance()
self.expenses = Expenses.get_instance()
@staticmethod
def get_instance():
if Predictor.__instance is None:
Predictor.__instance = Predictor()
return Predictor.__instance
def read_data(self):
gsa = GoogleSheetAdaptor(self.cfg.credFileName, 'Expenses_2020')
curatedData = gsa.get_spreadsheet_data()
# print(len(curatedData))
# Sort the data by Dates, so that we can run through the usage.
curatedData = curatedData.sort_values(by=['iDate'])
# RowIndex = A sequence running through the dataset
curatedData['RowIndex'] = np.arange(len(curatedData))
# StitchIndex = Connecting links to the next usage.
curatedData['StitchIndex'] = curatedData.apply(lambda row: curatedData[
(curatedData['RowIndex'] <= row['RowIndex']) & (curatedData[
'Description'] == row['Description'])]['Description'].count(),
axis=1)
minDate = str(min(curatedData['iDate'])) if os.environ.get('PREDICT_FROM') is None \
else os.environ.get('PREDICT_FROM')
# DateIndex = A sequence for Dates.
# 0 = minDate, for example 01-Jan-2020,
# 1 = minDate + 1, for example 02-Jan-2020, even if purchase is not made.
curatedData['DateIndex'] = curatedData.apply(lambda row: (dt.strptime(
str(row['Date']), '%Y%m%d') - dt.strptime(minDate, '%Y%m%d')).days,
axis=1)
# UsageIndex = Distance between current purchase and the next purchase, in days
curatedData['UsageIndex'] = curatedData.apply(
lambda row: -999 if len(curatedData[
(curatedData['Description'] == row['Description']) &
(curatedData['StitchIndex'] == row['StitchIndex'] + 1)][
'DateIndex'].values - row['DateIndex']) == 0 else
(curatedData[(curatedData['Description'] == row['Description']) &
(curatedData['StitchIndex'] == row['StitchIndex'] + 1
)]['DateIndex'].values - row['DateIndex'])[0],
axis=1)
descriptionList = curatedData['Description'].unique()
# Split the data based on the descriptions.
descriptionGroupMap = {
description: curatedData[curatedData['Description'] == description]
for description in descriptionList
}
# for each of these descriptionGroups, fetch y-Intercept and Slope using LinearRegression
modelValues = {
description: fit_model(descriptionGroupMap[description])
for description in descriptionGroupMap
}
# merge the y-Intercept and slope into the dataframe.
# curatedData['Intercept'] = curatedData.apply(lambda row:
# None if (modelValues[row['Description']] is None
# # or
# # row['UsageIndex'] == -999
# )
# else modelValues[row['Description']][0], axis=1)
# curatedData['Slope'] = curatedData.apply(lambda row:
# None if (modelValues[row['Description']] is None
# # or
# # row['UsageIndex'] == -999
# )
# else modelValues[row['Description']][1], axis=1)
# curatedData['PredictedUsageIndex'] = curatedData.apply(lambda row:
# None if (modelValues[row['Description']] is None
# # or
# # row['UsageIndex'] == -999
# )
# else modelValues[row['Description']][2], axis=1)
curatedData['DueDate'] = curatedData.apply(
lambda row: None if (modelValues[row['Description']] is None
# or
# row['UsageIndex'] == -999
) else
(dt.strptime(str(row['iDate']), "%Y%m%d") + td(days=modelValues[
row['Description']][2])).strftime('%Y%m%d'),
axis=1)
self.expenses.data = curatedData
def run_prediction_shop(self, shopName, fromDate=None):
if self.expenses.data is None:
self.read_data()
if fromDate is not None:
subset = self.expenses.data[
(self.expenses.data['Where'] == shopName)
& (self.expenses.data['iDate'] >= fromDate)
& (self.expenses.data['Description'] == 'Rice - Ponni Boiled')]
else:
subset = self.expenses.data[
(self.expenses.data['Where'] == shopName)
& (self.expenses.data['Description'] == 'Rice - Ponni Boiled')]
subset.sort_values(by=['DateIndex'], inplace=True)
def read_data():
predictor = Predictor.get_instance()
predictor.read_data()
expenses = Expenses.get_instance()
expenses.wordCloud = None
return redirect('/wc/')
from wordcloud import (WordCloud, get_single_color_func)
import matplotlib
import matplotlib.pyplot as plt
from InventoryPredictor.src.config import Config
from InventoryPredictor.src.predictor import Predictor
from InventoryPredictor.src.expenses import Expenses
from Authenticator.src.decorator import login_required, is_valid_user
import io
import os
import base64
cfg = Config.get_instance()
ex = Expenses.get_instance()
htmlTemplate = '<!DOCTYPE html>' \
'<html lang="en">' \
' <head>' \
' <title>Help me Shop!</title>' \
' <meta charset="UTF-8"> ' \
' <meta name="viewport" content="width=device-width, initial-scale=1, maximum-scale=1"> ' \
' <style>' \
' body, html {' \
' height: 100%;' \
' width: 100%;' \
' overflow-x: hidden;' \
' margin: 0;' \
' -webkit-perspective: 1000;' \
' font-family: "Poppins", sans-serif;' \
' background: #303030;' \
' -webkit-font-smoothing: antialiased;' \
' }' \
' .overlay {' \