Auto-complete

This repository contains the code to create a frequency based autocomplete system. The structure of the project is as follows:

|--app
|    |--db
|    |--routes
|--data   
|    |--convo.text
|--notebooks
|--src
|    |--ngrams.py
|--main.db
|--readme.md
|--runner.py
|--setup.py

app directory contains flask based autocomplete server. notebooks directory contains the runner code to create n-gram frequency files from raw text files present in data directory. src contains the code to generate n-grams, this uses NLTK routines to accomplish creation of frequency files.

If you want to run this project locally, then make sure you carry out the following sequence of steps:

1. Run the notebook in the notebooks folder. This will create three files in data folder namely:

   • unigrams.csv
   • bigrams.csv
   • trigrams.csv

2. Now create a new venv using requirements.txt

3. After this run setup.py, this will create a sqlite3 database named main.db

4. Now setup the following environment variables to run the flask server (Tested on Windows only):

   • $env:FLASK_APP=runner
   • $env:FLASK_ENV = "devlopement"
   • $env:DEBUG = 1

5. Type type flask-run

6. Go to the url http://127.0.0.1:5000/api/status to check if the server is running, you will see a response of {"healthy":true}

7. To get suggestions use query parameters as shown below:

   http://127.0.0.1:5000/api/query?string=word1,word2,word3