A serverless web app to check spelling and suggest alternate words.

Spell checking is based on two algorithms, n-gram Decomposition and Levenshtein Distance. N-gram decomposition is the process of breaking apart a string into each proper n-letter substring (with two extra ones added for the beginning and ending of the word). Levenshtein distance is the measurement of the distance in edits between two strings.

The database can be seeded using the program seed_database.py. You will need a file where the words are listed one-per-line like in the sample dictionary provided in this repo. I usually use the one at https://github.com/dwyl/english-words.

• /check?word=MISSPELLEDWORD
• /ngrams?word=WORD
• /levenshtein?a=WORD1&b=WORD2

• [app.py] -- Flask app for API
• [find_word.py] -- Use the database to find similar words
• [get_ngrams.py] -- Decompose words into n-grams
• [seed_database.py] -- Seed the database from a word list file
• [tables.py] -- SQLAlchemy table definitions
• [test_get_ngrams.py] -- PyTest test cases for get_ngrams.py
• [tiny_sample_database.txt] -- A really short word list for testing