phanipavank/IIITB_Menu
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add omelete and fixes for receing format changes

Browse Source
This commit is contained in:
Phani Pavan Kambhampati
2024-03-26 09:43:06 +05:30
committed by GitHub
parent 16ce2d888c
commit d01de9dedb
1 changed files with 12 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
scrapeToJson.py
View File

@@ -1,6 +1,5 @@
import pandas import pandas
import numpy import numpy
import numpy
import datetime import datetime
import json import json
import hashlib import hashlib
@@ -18,20 +17,15 @@ ALLDAYS = [
if __name__ == "__main__": if __name__ == "__main__":
#! TODO: Find a better XLSX import process. XLS IS ASS #! TODO: Find a better XLSX import process. XLS IS ASS
# mainData = pandas.read_excel("menu.ods") # mainData = pandas.read_excel("menu.ods")
mainData = pandas.read_csv("menu.csv") mainData = pandas.read_csv("menu.csv", skiprows=1)
mainData = mainData.drop(" ", axis="columns") mainData = mainData.replace("\xa0", numpy.nan)
mainData[" .1"] = mainData[" .1"].replace("\xa0", numpy.nan) breaksLoc = numpy.where(pandas.isna(mainData["\xa0"]).to_numpy() == False)[0]
breaksLoc = numpy.where( breaksLoc = numpy.append(breaksLoc, [len(mainData)])
pandas.isna(mainData[" .1"]).to_numpy()
== True
)[0][2:]
print(breaksLoc) print(breaksLoc)
rowNames = mainData[" .1"].to_numpy() rowNames = mainData["\xa0.1"].to_numpy()
# Item names: x[x["Unnamed: 1"]=="BREAKFAST 1"]["MONDAY"].item()
# print(breaksLoc)
finalData = {"menu": {}, "dates": {}, "items": {}} finalData = {"menu": {}, "dates": {}, "items": {}}
for sesNum, ses in enumerate(["bf", "ln", "sk", "dn"]): for sesNum, ses in enumerate(["bf", "ln", "sk", "dn"]):
finalData["items"][ses] = mainData[" .1"].to_list()[ finalData["items"][ses] = mainData["\xa0.1"].to_list()[
breaksLoc[sesNum] + 1 : breaksLoc[sesNum + 1] breaksLoc[sesNum] + 1 : breaksLoc[sesNum + 1]
] ]
for dayNumber, eachDay in enumerate(ALLDAYS): for dayNumber, eachDay in enumerate(ALLDAYS):
@@ -69,11 +63,15 @@ if __name__ == "__main__":
) )
if isinstance(part, float) and (part is numpy.nan): if isinstance(part, float) and (part is numpy.nan):
part = "MT" part = "MT"
if part.strip()=="": if part.strip() == "":
part = "MT" part = "MT"
# print(f"{item}: {part}") # print(f"{item}: {part}")
if "egg" in part.strip().lower() or "omlet" in part.strip().lower(): if (
"egg" in part.strip().lower()
or "omlet" in part.strip().lower()
or "omelet" in part.strip().lower()
):
print(f"EGG Found, {part} at {ses}") print(f"EGG Found, {part} at {ses}")
eggy = "EGG" eggy = "EGG"
elif "chicken" in part.strip().lower(): elif "chicken" in part.strip().lower():