Home Search

An app to search for properties on Zoopla and filter on the properties total area. The area is scraped off the floorplan image using OCR.

Local Deployment

This project consists of a frontend (web) application written in Flutter, and a backend service written in Python. The backend requires access to a postgres database.

The simplest way to deploy it locally is to use docker:

1. From the project root directory:
2. Create a Google Maps API key if you don't have one.
3. Create a Google Sign-in Client ID if you don't have one.
4. Create a Zoopla API key if you don't have one.
5. Set environment variables for API keys:

   export GOOGLEMAPSAPIKEY=<your google maps API key>
   export ZOOPLAAPIKEY=<your Zoopla API key>
   export POSTGRES_PASSWORD=<make up a password for the DB user 'admin'>
   export SECRET_KEY=`openssl rand -hex 32`
   export CLIENT_ID=<OAuth2.0 client ID, in GCP API & Services -> Credentials>
   export URL_API=localhost  # or http:/127.0.0.1

   • SECRET_KEY is used for auth and should be a random 32 digit hex value. The openssl command above is a simple way to generate this, but you can use whatever key you like.
   • If you want to connect to an existing database, instead of POSTGRES_PASSWORD, set an environment variable called DATABASE_URL with the full connection string. This is the environment variable name used by Heroku.
6. Build and run:

   docker compose up

   note if you have set URL_API to your remote server, then for local testing run:

   URL_API=http://localhost docker compose up

7. Test the backend:

   curl localhost:80/pdf/acf057f7f02b0cf3552e149de5772640bf0bfd2c.pdf

   This can take around 10s but should return {"area":442.6}
8. The front end should be available at:

   localhost:5001
   

9. The OpenAPI docs foe the backend are available at:

   localhost/docs
   

If you need to force a rebuild of the containers, try:

docker-compose build --no-cache

If you need to clear the DB:

docker volume prune

Server Deployment (Heroku)

The following instructions show you how to create a new deployment for this app on Heroku. If you have already done this, you can run the deploy.sh script instead, which only assumes you have set the APP_NAME environment variable.

1. Install the heroku cli and log in
2. Set the name for your heroku app:

   export APP_NAME=<your app name>
   export API_NAME=${APP_NAME}server

3. Create two apps, one for the frontend, and one for the backend:

   heroku create $APP_NAME
   heroku create $API_NAME

4. Set the following config variables in Heroku:
   • GOOGLEMAPSAPIKEY: your Google Maps API key
   • ZOOPLAAPIKEY: your Zoopla API key
   • SECRET_KEY: (Only for backend) A random key generated usingopenssl rand -hex 32 or something similar.
   • CLIENT_ID: Your OAuth2.0 client ID. In GCP API & Services -> Credentials.
5. Set the following environment variables:

   export URL_API=https://${API_NAME}.herokuapp.com

   note URL_API is a build arg for the FrontEndDockerfile, but you don't have to worry about this if you use docker compose.
6. Add the postgres add-on to your backend Heroku app, i.e. the one name $API_NAME which should end in server. Note you should first check that you haven't already provisioned postgres by running heroku addons --app $API_NAME.

   heroku addons:create heroku-postgresql:hobby-dev --app $API_NAME

7. Log in to the Heroku container registry:

   heroku container:login

8. Build the images locally, most easily done via:

   docker compose up -d

   You can stop the containers using docker compose down as you don't need them.
9. Tag the images:

   docker tag home-search_backend registry.heroku.com/$API_NAME/web
   docker tag home-search_frontend registry.heroku.com/$APP_NAME/web

10. Push to Heroku (wait for first to complete):

    docker push registry.heroku.com/$API_NAME/web
    docker push registry.heroku.com/$APP_NAME/web

11. Release the containers

    heroku container:release --app $API_NAME web
    heroku container:release --app $APP_NAME web

12. Scale up your dynos:

    heroku ps:scale --app $API_NAME web=1
    heroku ps:scale --app $APP_NAME web=1

13. open the app

    heroku open --app $APP_NAME