Use NLP to enhance data quality of ingredients and prepare filter methods

	What?	Techniques	Packages
1	match different forms of an ingredients (e.g. egg --> eggs, organic egg)	Lemmatization, compute similarity by word embeddings, compute string distance (e.g. Levensthein), order alphabetically, coreference resolution	spaCy, flair, GermaNet Uni Tübingen, allennlp
2	extract topics from recipes as basis for filter	Topic Modelling/Extraction	gensim
3	extract keywords from recipes as basis for filter	keyword extraction (based on Rake algorithm), NER, stopwords removal	spaCy, gensim.summarization, multi-rake
4	Develop dictionary of neutral ingredients, dairy and meat/fish as basis for vegetarian/vegan filter	rule based matching	???