ingredient-based food recommendation system using deep learning

I give a recipe image the model should classify the recipe into which food class it belongs, get the ingredients of the recipe in the image, and get similar recipe images displayed. this is one task, one more is when you provide ingredients you need to get the recipe images that can be prepared with the help of these ingredients. and need a web application to execute the above model.
dataset I have a dataset of images related to 29 different food classes in the below format
dataset/
├── pancakes/
│ ├── 0.jpg
│ ├── 1.jpg
│ └── ...
├── pizza/
│ ├── 0.jpg
│ ├── 1.jpg
│ └── ...
we also have the ingredients.txt file in which there are ingredients related to these images
image_path,ingredients
pancakes/0.jpg,flour, eggs, milk, ...
pancakes/1.jpg,flour, eggs, milk, ...
pizza/0.jpg,flour, tomatoes, cheese, ...
pizza/1.jpg,flour, tomatoes, cheese, ...