Recipe Recommendation System (RecRes)

Demo

Models Weights & Data

Unless otherwise specified, please download all models and place the entire folder under /saved_models

• Ingredients Image Classification [Download]
• Cuisine Classification [Download]
• D2VModel [Download]
• Recipe Database for D2VModel, put it under RecRes folder [Download]
• Models & Data for Attention Encoder Decoder [Download]

Installation

pip install -r requirements.txt

[Mac Users Only]

Pip-installed Tensorflow might not work, instead, install it with conda.

pip uninstall tensorflow
conda install tensorflow

Start Streamlit

streamlit run app.py

---

Related Datasets

Recipe

Dataset	Features	#Data	Remark	Used
Food Recommendation System - schemersays	name, ingredients, cuisines, ratings	400
Food.com - Recipes	name, preparing time, date, tags, nutrition, # cooking steps, description, ingredients, #ingredients	180K	Provides raw and tokenize data. Paper.
Food.com - Review of Recipes	date, rating	700K	Provides raw and tokenize data. Paper.
Indian Food and Its Recipes Dataset (With Images)	name, image_url, description, cuisine, course, diet, prep_time, ingredients, instructions	4226	Scraped from Archana's Kitchen
MealRec	name, #reviews, category, aver_rate, image_url, ingredients, cooking_directions, nutritions, reviews, tags	7280	There are multiple reviews for one recipe
Indian Food 101	name, ingredients, diet, prep_time, cook_time, flavor_profile, course, state, region	255
foodRecSys-V1	recipe_name, image_url, ingredients, cooking_directions, nutritions, rating	45568

Restaurant

Dataset	Features	#Data	Remark	Used
Restaurant Data with Consumer Ratings	payment type, operating hours, operating days, parking_lot, latitude, longitude, the_geom_meter, name, address, city, state, country, fax, zip, alchohol, smoking_area, dress_code, accessibility, price, url, Rambience, franchise, area, other_services, rating, food_rating, service_rating	-	Consist of multiple csv of users and restaurants with different length
Micheline Guide Restaurants	name, address, location, price, cuisine, longitude, latitude, phoneNumber, Url, WebsiteUrl	6653

Related Tutorials and Repositories

Name	Description	Remark
Scraping Google Reviews with Selenium(Python)	Web scraping google reviews via Selenium and BeautifulSoup
recipe-recommendation-system	Data-driven recipe recommendation system using web-scraped recipe data (including but not limited to data like ingredients, health facts, etc.) and user’s historical preference	No access to dataset
recipes-telegram-bot	Telegram bot that can recommend recipes based on the ingredients you already have	Technical article on Medium
Food Recommendation using BERT	Based on cosine similarities computed on embedding from BERT
Restaurant_recommendation_system	Recommend users based on ratings and comments of other visitors taking into consideration of their location and preferences	Slides