andrei-stoica
/
recipe-graph
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

added more units and instuctions to list

This commit is contained in:
Andrei Stoica 2022-07-24 15:21:57 -04:00
parent 3092f67899
commit 910597b3ff
1 changed files with 16 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
src/scrape.py
View File

@ -1,4 +1,5 @@
from ast import alias from ast import alias
from dis import Instruction
import db import db
import re import re
from sqlalchemy import select, desc, exists, not_, except_ from sqlalchemy import select, desc, exists, not_, except_
@ -13,9 +14,22 @@ from argparse import ArgumentParser
def parse_ingredient(ingredient_text): def parse_ingredient(ingredient_text):
units = ['teaspoon', 'tablespoon', 'gram', 'ounce', 'jar', 'cup', 'pinch', units = ['teaspoon', 'tablespoon', 'gram', 'ounce', 'jar', 'cup', 'pinch',
'container', 'slice', 'package', 'pound', 'can', 'dash', 'spear', 'container', 'slice', 'package', 'pound', 'can', 'dash', 'spear',
'bunch', 'quart', 'cube', 'envelope', 'square', 'sprig', 'bags', 'bunch', 'quart', 'cube', 'envelope', 'square', 'sprig', 'bag',
'box', 'drop', 'fluid ounce', 'gallon', 'head', 'link', 'loaf', 'box', 'drop', 'fluid ounce', 'gallon', 'head', 'link', 'loaf',
'pint', 'pod', 'sheet', 'stalk', 'whole', 'bar'] 'pint', 'pod', 'sheet', 'stalk', 'whole', 'bar', 'bottle', 'bulb',
'year', 'fillet', 'litter', 'packet', 'slices']
instructions = ['and', 'or', 'chopped', 'diced', 'brewed', 'chilled', 'chunky',
'small', 'medium', 'large', 'couarse(:?ly)?', 'cracked',
'crushed', 'ground', 'cooked', 'cubed', 'crumbled', 'cut',
'cold', 'hot', 'warm', 'day', 'old', 'drained', 'canned',
'dried', 'dry', 'fine(?:ly)', 'firmly', 'fresh', 'frozen',
'grated', 'grilled', 'hard', 'hot', 'juliened?', 'leftover',
'light', 'lite', 'mashed', 'melted', 'minced', 'packed',
'peeled', 'pitted', 'sliced', 'prepared', 'refrigerated',
'rehydrated', 'seedless', 'shaved', 'shredded', 'sifted',
'sieved', 'shucked', 'slivered', 'thick', 'sliced', 'thin',
'toasted', 'trimmed', 'unbaked', 'uncooked', 'unpeeled',
'unopened', 'unseasoned']
number_regex = '((?:[\d\\./\\u00BC-\\u00BE\\u2150-\\u215E]*\s?(?:\(.+\))?)*)' number_regex = '((?:[\d\\./\\u00BC-\\u00BE\\u2150-\\u215E]*\s?(?:\(.+\))?)*)'
ingredient_regex = '([a-zA-Z \'\-]+)' ingredient_regex = '([a-zA-Z \'\-]+)'
supplement_regex = ',?(.*)' supplement_regex = ',?(.*)'