Corrector-Hackathon

This is a project of the UAB Multimedia Systems Hackathon. The project is a website hosted in the Google Cloud that must be able to detect misakes, suggest some possible correct words and correct them.

Full-Stack web app to correct grammar mistakes. Deployed on Google Cloud, CI/CD support with Cloud Build Triggers in sync with this repo and external API calls made via Cloud Functions. Use of Python Web framework Flask and its templating engine Jinja.

Local Installation

git clone https://github.com/ggcr/Corrector-Hackathon.git ~/.
cd ~/Corrector-Hackathon
pip3 install -r requirements.txt
python3 main.py

If you have successfully completed the steps at http://localhost:8080 you will be able to see the web UI.

To be able to run the website locally and make it functional, the project credentials in .json are required These credentials are not in this repo.

If you do not have credentials, this message will be displayed on the console:

Google cloud .json credentials are missing! Unable to call Typewise API.

If you have credentials, you must uncomment line 73 from main.py:

os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = './corrector-sm-9ef8799680bd.json'

Google Cloud

With the Google Cloud Build App, we are able to link this repository to the cloud instance. This will allow CI / CD to be made since every time a push is made to the main branch a Build Trigger will be activated in the cloud which will launch a new instance of our backend.

Cloud Builder: cloudbuild.yaml file

This trigger will deploy the app every time it detects changes to the main branch.

• To do this, you must give "App Engine Deployer" permissions to this cloud builder user.

steps:
- name: "gcr.io/cloud-builders/gcloud"
  args: ["app", "deploy"]
timeout: "1600s"

App Engine: app.yaml file

This file will be used to deploy the app. It will be called by the cloud builder.

runtime: python39

handlers:
- url: /static
  static_dir: static
- url: /.*
  script: auto

There are two methods for correcting text:

Local Python Lib: pyspellchecker

It is a python package that allows you to make local text corrections. It runs on the server and only on the server.

External API: Typewise API

Using a cloud function we can call this external API and return the results.

In the 'correct' Cloud Function we have the following main.py code:

def correct_sentence(request):
    import requests
    request_json = request.get_json()
    msg = ""
    if request.args and 'message' in request.args:
        msg = request.args.get('message')
    elif request_json and 'message' in request_json:
        msg = request_json['message']
    
    url = "https://typewise-ai.p.rapidapi.com/correction/whole_sentence"

    payload = {
        "text": msg,
        "keyboard": "QWERTY",
        "languages": [
            "en",
            "de",
            "es"
        ],
    }
    headers = {
        "content-type": "application/json",
        "X-RapidAPI-Host": "typewise-ai.p.rapidapi.com",
        "X-RapidAPI-Key": "***************"
    }

    response = requests.request("POST", url, json=payload, headers=headers)
    return response.text