moved scripts to module

This commit is contained in:
2022-09-18 11:39:54 -04:00
parent 04200b41ce
commit 4f61fb5ccc
4 changed files with 0 additions and 0 deletions

219
src/db.py
View File

@@ -1,219 +0,0 @@
import os
import logging
from types import NoneType
from dotenv import load_dotenv
from xmlrpc.client import Boolean
from sqlalchemy import create_engine, Column, Integer, String, Boolean, \
ForeignKey, UniqueConstraint, func, select, and_, or_, \
not_
from sqlalchemy.types import ARRAY
from sqlalchemy.engine import URL
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import Session
Base = declarative_base()
class Ingredient(Base):
__tablename__ = 'Ingredient'
id = Column(Integer, primary_key = True)
name = Column(String, nullable = False)
class RecipeSite(Base):
__tablename__ = 'RecipeSite'
id = Column(Integer, primary_key = True)
name = Column(String, nullable = False, unique = True)
ingredient_class = Column(String, nullable = False)
name_class = Column(String, nullable = False)
base_url = Column(String, nullable = False, unique = True)
class Recipe(Base):
__tablename__ = 'Recipe'
id = Column(Integer, primary_key = True)
name = Column(String)
identifier = Column(String, nullable = False)
recipe_site_id = Column(Integer, ForeignKey('RecipeSite.id'))
UniqueConstraint(identifier, recipe_site_id)
class RecipeIngredient(Base):
__tablename__ = 'RecipeIngredient'
id = Column(Integer, primary_key = True)
text = Column(String, nullable = False)
recipe_id = Column(Integer, ForeignKey('Recipe.id'))
ingredient_id = Column(Integer, ForeignKey("Ingredient.id"))
class RecipeIngredientParts(Base):
__tablename__ = 'RecipeIngredientParts'
id = Column(Integer, ForeignKey("RecipeIngredient.id"), primary_key=True)
quantity = Column(String)
unit = Column(String)
instruction = Column(String)
ingredient = Column(String)
supplement = Column(String)
class IngredientConnection(Base):
__tablename__ = 'IngredientConnection'
ingredient_a = Column(String,
ForeignKey("RecipeIngredientParts.ingredient"),
primary_key = True)
ingredient_b = Column(String,
ForeignKey("RecipeIngredientParts.ingredient"),
primary_key = True)
recipe_count = Column(Integer)
UniqueConstraint(ingredient_a, ingredient_b)
class RecipeConnection(Base):
__tablename__ = 'RecipeConnection'
recipe_a = Column(Integer,
ForeignKey("Recipe.id"),
primary_key = True)
recipe_b = Column(Integer,
ForeignKey("Recipe.id"),
primary_key = True)
ingredient_count = Column(Integer)
class RecipeGraphed(Base):
__tablename__ = "RecipeGraphed"
recipe_id = Column(Integer, ForeignKey("Recipe.id"), primary_key = True)
status = Column(Boolean, nullable = False, default = False)
def get_engine(use_dotenv = True, **kargs):
if use_dotenv:
load_dotenv()
DB_URL = os.getenv("POSTGRES_URL")
DB_USER = os.getenv("POSTGRES_USER")
DB_PASSWORD = os.getenv("POSTGRES_PASSWORD")
DB_NAME = os.getenv("POSTGRES_DB")
eng_url = URL.create('postgresql',
username=DB_USER,
password=DB_PASSWORD,
host=DB_URL,
database=DB_NAME)
return create_engine(eng_url)
def create_tables(eng):
logging.info(f"Createing DB Tables: {eng.url}")
Base.metadata.create_all(eng, checkfirst=True)
def pair_query(pairable, groupable, recipe_ids = None, pair_type = String):
pair_func= func.text_pairs
if pair_type == Integer:
pair_func=func.int_pairs
new_pairs = select(groupable,
pair_func(func.array_agg(pairable.distinct()),
type_=ARRAY(pair_type)).label("pair"))\
.join(RecipeIngredientParts)
if not type(recipe_ids) == NoneType:
new_pairs = new_pairs.where(RecipeIngredient.recipe_id.in_(recipe_ids))
new_pairs = new_pairs.group_by(groupable)\
.cte()
return new_pairs
def pair_count_query(pairs, countable, recipe_ids = None):
new_counts = select(pairs, func.count(func.distinct(countable)))
if not type(recipe_ids) == NoneType:
new_counts = new_counts.where(or_(pairs[0].in_(recipe_ids),
pairs[1].in_(recipe_ids)))
new_counts = new_counts.group_by(pairs)
return new_counts
def update_graph_connectivity(session = None):
# this is pure SQLAlchemy so it is more portable
# This would have been simpler if I utilized Postgres specific feature
if not session:
session = Session(get_engine())
with session.begin():
ids = select(Recipe.id)\
.join(RecipeGraphed, isouter = True)\
.where(RecipeGraphed.status.is_not(True))
num_recipes = session.execute(select(func.count('*')).select_from(ids.cte())).fetchone()[0]
if num_recipes <= 0:
logging.info("no new recipies")
return
logging.info(f"adding {num_recipes} recipes to the graphs")
new_pairs = pair_query(RecipeIngredientParts.ingredient,
RecipeIngredient.recipe_id,
recipe_ids = ids)
new_counts = pair_count_query(new_pairs.c.pair,
new_pairs.c.recipe_id)
logging.info("addeing new ingredient connections")
for pair, count in session.execute(new_counts):
connection = session.query(IngredientConnection)\
.where(and_(IngredientConnection.ingredient_a == pair[0],
IngredientConnection.ingredient_b == pair[1]))\
.first()
if connection:
connection.recipe_count += count
session.merge(connection)
else:
session.add(IngredientConnection(ingredient_a = pair[0],
ingredient_b = pair[1],
recipe_count = count))
# update RecipeConnection
logging.info("adding new recipe connections")
all_pairs = pair_query(RecipeIngredient.recipe_id,
RecipeIngredientParts.ingredient,
pair_type=Integer)
new_counts = pair_count_query(all_pairs.c.pair,
all_pairs.c.ingredient,
recipe_ids=ids)
i = 0
for pair, count in session.execute(new_counts):
session.add(RecipeConnection(recipe_a = pair[0],
recipe_b = pair[1],
ingredient_count = count))
# flush often to reduce memory usage
i += 1
if (i % 100000) == 0:
session.flush()
# update RecipeGraphed.status
logging.info("updating existing RecipeGraphed rows")
for recipeGraphed in session.query(RecipeGraphed)\
.where(RecipeGraphed.recipe_id.in_(ids)):
recipeGraphed.status = True
session.merge(recipeGraphed)
graphed = select(RecipeGraphed.recipe_id)
# add recipies that aren't in the table
logging.info("adding new RecipeGraphed rows")
for recipe in session.query(Recipe)\
.where(and_(Recipe.id.in_(ids),
not_(Recipe.id.in_(graphed)))):
session.add(RecipeGraphed(recipe_id=recipe.id, status=True))
if __name__ == "__main__":
eng = get_engine()
create_tables(eng)

View File

@@ -1,29 +0,0 @@
from sqlalchemy.orm import sessionmaker
import db
import json
import argparse
import logging
parser = argparse.ArgumentParser(description='Import recipes into database')
parser.add_argument('file', type=str,
help='JSON file with recipe site information')
parser.add_argument('-v', '--verbose', action='store_true')
args = parser.parse_args()
if args.verbose:
logging.basicConfig(level=logging.INFO)
logging.getLogger('sqlalchemy.engine').setLevel(logging.INFO)
with open(args.file) as f:
sites = json.load(f)
eng = db.get_engine()
S = sessionmaker(eng)
with S.begin() as session:
for site in sites:
logging.info(f"Adding {site}")
session.add(db.RecipeSite(**site))

View File

@@ -1,190 +0,0 @@
from ast import alias
from dis import Instruction
import db
import re
from sqlalchemy import select, desc, exists, not_, except_
from sqlalchemy.exc import IntegrityError
from sqlalchemy.orm import sessionmaker
import bs4
from urllib.request import urlopen
from urllib.parse import urljoin
import logging
from argparse import ArgumentParser
def parse_ingredient(ingredient_text):
units = ['teaspoon', 'tablespoon', 'gram', 'ounce', 'jar', 'cup', 'pinch',
'container', 'slice', 'package', 'pound', 'can', 'dash', 'spear',
'bunch', 'quart', 'cube', 'envelope', 'square', 'sprig', 'bag',
'box', 'drop', 'fluid ounce', 'gallon', 'head', 'link', 'loaf',
'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', 'cracked',
'crushed', 'ground', 'cooked', 'cubed', 'crumbled', 'cut',
'cold', 'hot', 'warm', 'day', 'old', 'drained', 'canned',
'dried', 'dry', 'fine', 'firm', '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?(?:\(.+\))?)*)'
ingredient_regex = '([a-zA-Z \'\-]+)'
supplement_regex = ',?(.*)'
units_regex = "|".join([f'[{unit[0]}{unit[0].capitalize()}]{unit[1:]}'
for unit in units])
units_regex = f"((?:(?:{units_regex})e?s?)?)"
instructions_regex = "|".join([f'[{inst[0]}{inst[0].capitalize()}]{inst[1:]}'
for inst in instructions])
instructions_regex = f"((?:(?:(?:{instructions_regex})(?:ly)?)| )*)"
regex = re.compile(number_regex +
units_regex +
instructions_regex +
ingredient_regex +
supplement_regex)
m = regex.match(ingredient_text)
logging.info(f"Parsed {ingredient_text}, found: {m}")
if not m:
return None
return [text.strip() if text else None for text in m.groups()]
def reparse_ingredients(session):
cte = (except_(select(db.RecipeIngredient.id),
select(db.RecipeIngredientParts.id))).\
alias('missing')
missing = session.query(db.RecipeIngredient).where(db.RecipeIngredient.id.in_(cte)).all()
for ingredient in missing:
parts = parse_ingredient(ingredient.text)
if not parts:
continue
quantity, unit, instruction, name, supplement = parts
session.add(db.RecipeIngredientParts(id = ingredient.id,
quantity = quantity,
unit = unit,
instruction = instruction,
ingredient = name,
supplement = supplement))
def load_recipe(recipe_url):
try:
logging.info(f'Loading Recipe: {recipe_url}')
with urlopen(recipe_url) as f:
if f.getcode() == 404:
raise Exception(f"Recipe Does not exist: {recipe_url}")
return bs4.BeautifulSoup(f.read().decode(), 'html.parser')
except Exception as e:
logging.warning(f"Could not download or parse recipe: {recipe_url}")
logging.warning(e)
return None
def parse_recipe(session, recipe, site):
recipe_url = urljoin(site.base_url, str(recipe.identifier))
recipe_page = load_recipe(recipe_url)
if not recipe_page:
return None
name_candidates = recipe_page.find_all(class_=site.name_class)
if len(name_candidates) == 0:
raise Exception(f"Could not extract recipe name: {recipe_url}")
name_div = name_candidates[0]
recipe.name = name_div.text
logging.info(f"Adding Recipe {recipe.name} from {recipe_url}")
session.add(recipe)
session.flush()
ingred_candidates = recipe_page.find_all(class_=site.ingredient_class)
for candidate in ingred_candidates:
ingred = db.RecipeIngredient(text=candidate.text,
recipe_id=recipe.id)
session.add(ingred)
session.flush()
parts = parse_ingredient(ingred.text)
if parts:
quantity, unit, instruction,ingredient, supplement = parts
ingred_parts = db.RecipeIngredientParts(id = ingred.id,
quantity = quantity,
unit = unit,
instruction = instruction,
ingredient = ingredient,
supplement = supplement)
session.add(ingred_parts)
logging.info(f"{len(ingred_candidates)} ingredients found. Inserting into DB")
return recipe
parser = ArgumentParser(description="Scrape a recipe site for recipies")
parser.add_argument('site',
help='Name of site')
parser.add_argument('-id', '--identifier', dest='id',
help='url of recipe(reletive to base url of site) or commma seperated list')
parser.add_argument('-a', '--auto', action='store', dest='n',
help='automaticaly generate identifier(must supply number of recipies to scrape)')
parser.add_argument('-v', '--verbose', action='store_true')
args = parser.parse_args()
if args.verbose:
logging.basicConfig(level=logging.INFO)
logging.getLogger('sqlalchemy.engine').setLevel(logging.INFO)
eng = db.get_engine()
S = sessionmaker(eng)
with S.begin() as sess:
site = sess.query(db.RecipeSite).where(db.RecipeSite.name == args.site).one()
site_id = site.id
recipe_ids = []
starting_id = 0
if args.id and not args.n:
recipe_ids.append(args.id)
logging.info(f'Retreiving single recipe: {args.id}')
elif args.n:
if not args.id:
last_recipe = sess.query(db.Recipe).\
where(db.Recipe.recipe_site_id == site.id).\
order_by(desc(db.Recipe.identifier)).\
limit(1).\
scalar()
starting_id = int(last_recipe.identifier) + 1
else:
starting_id = int(args.id)
recipe_ids = range(starting_id, starting_id+int(args.n))
logging.info(f'Retreving {args.n} recipes from {site.base_url} starting at {starting_id}')
for recipe_id in recipe_ids:
try:
savepoint = sess.begin_nested()
recipe = db.Recipe(identifier = recipe_id, recipe_site_id = site.id)
parse_recipe(sess, recipe, site)
savepoint.commit()
except KeyboardInterrupt as e:
savepoint.rollback()
break
except Exception as e:
savepoint.rollback()
logging.error(e)
continue