terencelim321 / chihuahua2

test

Food Recommendation Flask API

Flask API Endpoints

GET /api/v0.1/recommend

Returns list of recommended restaurants based on provided list of restaurant IDs.

Parameters

• ids: Compulsory. List of IDs (comma separated) to be used to generate recommendations.
• top: Default = 5. Returns this number of recommendations based on descending score.
• prices: Optional. Only recommends restaurants with these prices (comma separated, valid values: 1=$, 2=$$, 3=$$$, 4=$$$).
• zip_codes: Optional. Only recommends restaurants with these zip_codes (comma separated).

Output

• total: Total recommendations returned
• recommendations: List of recommendations with score ranging from 0 to 1 (higher the score, more relevant the recommendation).
• input_ids: Details associated with IDs provided
• rejected: Rejected IDs

Example

• http://:/api/v0.1/recommend?ids=1,4&top=2&price=2,3&zip_codes=012345,712345

{ 
	"total": 2, 
  	"recommendations": [
  		{ "id":"2", "name":"Chicken Rice", "price":"$$", "zip_code":"012345", "tags": ["restaurant"], "score":0.6435643 },
  		{ "id":"9", "name":"Bak Kut Teh", "price":"$$", "zip_code":"712345", "tags": ["restaurant"], "score":0.543154 }	
  	],
  	"input_ids": [
  		{ "id":"1", "name":"Tze Char", "price":"$$", "zip_code":"543210", "tags": ["restaurant"] }
  	]
  	"rejected_ids": [
  		"4"
  	]
}

POST /api/v0.1/recommend

Same as GET method, but also assumes provided restaurant IDs are user's preference & stores them in database.

• Additional "user_id" parameter required.
• All other parameters & output same as GET.

Running Flask API Service

1. Create Heroku App & Heroku Postgres

Prerequisite

• Heroku account
• Install postgreSQL (to use "psql" in command line)

Steps

1. Login to Heroku via web browser.
2. Create new app from Heroku dashboard.
3. Install Heroku Postgres add-on.
4. On Heroku Postgres add-on page, click on "Settings" > "View Credentials"
5. Declare local environment variable based on these credentials.

export DATABASE="<to be filled in>"
export DATABASE_HOST="<to be filled in>"
export DATABASE_PORT="<to be filled in>"
export DATABASE_USER="<to be filled in>"
export DATABASE_PASS="<to be filled in>"
export DATABASE_URL="postgresql://${DATABASE_USER}:${DATABASE_PASS}@${DATABASE_HOST}:${DATABASE_PORT}/${DATABASE}"

6. Connect to Heroku Postgres with following psql command

cd <project folder>
PGPASSWORD=$DATABASE_PASS psql -h $DATABASE_HOST -U $DATABASE_USER $DATABASE

7. Execute following commands to create the required tables

\i postgres/create_restn.sql
\i postgres/create_restn_tags.sql
\i postgres/create_restn_user_rating.sql

8. Confirm that tables have been created

\dt

2a. Deploy App to Heroku

Prerequisite

• Git
• Heroku CLI
• Created Heroku App & Heroku Postgres (instructions above)

Steps

1. Login to Heroku via web browser.
2. Click on created app on Heroku dashboard.
3. Click on "Settings" > "Reveal Config Vars".
4. Add the following configuration variable (there should be a DATABASE_URL variable already present).

KEY = APP_SETTINGS, VALUE = config.ProductionConfig

5. Click on "Deploy" > "Deploy using Heroku Git" section.
6. Following instructions to push codes to Heroku App.

2b. Running Locally

Setting up python virtual environment

Prerequisite

• Python >=3.6 ("python --version" to check)

Steps

1. Change to project folder.

cd <project folder>

2. Create python virtual environment.

python -m venv venv

3. Activate virtual environment. "(venv)" will appear at command prompt if activated.

venv\Scripts\activate (for windows) OR venv/bin/activate (for mac)

4. Install required python packages.

pip install -r requirements.txt

Running python script

Prerequisite

• Setup of python virtual environment (instructions above)
• Created Heroku App & Heroku Postgres (instructions above)

Steps

1. If not activated yet, activate virtual environment ("(venv)" will appear at command prompt if activated).

cd <project folder>
venv\Scripts\activate (for windows) OR venv/bin/activate (for mac)

2. Add local environment variable.

export APP_SETTINGS="config.DevelopmentConfig"

3. Run python script.

cd <project folder>
python run.py