AksaCarma BackEnd API

This is a BackEnd API for AksaCarma mobile apps. Any push to this repository will be automatically deployed in GCP.

This API use python v3.8 for the support of tensorflow docker image and FastAPI.

Run This API

There are two ways to run this API locally. First, run directly in host, second, run in docker.

Run in Host

• Clone this repo.

  git clone https://github.com/AksaCerma/aksacarma-cc.git
  

• Download machine learning model.

  cd saved_model/ml-aksacarma/variables
  wget https://storage.googleapis.com/variables-data-model/variables.data-00000-of-00001 -O variables.data-00000-of-00001
  

• Install the requirements.

  pip install -r requirements_local.txt
  

• Run the API.

  uvicorn main:app --port <your_port>
  

Run in Docker

In this repo, there are two Dockerfile. You can download the Dockerfile.git only or clone the repo and use the Dockerfile.

• Build the image.

  docker build -t <image>:<tag> --file <your_dockerfile> $(for i in `cat .env`; do out+="--build-arg $i " ; done; echo $out;out="") .
  

• Run the image.

  docker run -it -d -p <host_port>:8005 --name <container_name> <image>:<tag>
  

Swagger UI

To access Swagger docs, you can use localhost:<your_port>. The docs will look like this. It contains endpoints information of this API.

Environment Variables

The API needs to have environment variables to be used. The details of these environment variables are in the tables below. Please save this variables to .env file.

API Keys

Variable	Detail
SECRET_KEY	This API key
GOOGLE_KEY_i	API key for Google Search

This API needs five Google Search API key from 1-5 GOOGLE_KEY_1 GOOGLE_KEY_2 and so on.

Database

Variable	Detail
DB_HOST	Database host (IP or hostname)
DB_PORT	Database port
DB_USER	Database username
DB_PASS	Database password
DB_NAME	Database name

Google Bucket

Variable	Detail
BUCKET_NAME	Google Bucket name

Database Tables

This API requires three tables to work on.

Name	Detail
user	User data
history	Detect skin history
search_result	Search result for user history

The queries to create this tables are in create_tables.sql file. Please run the queries sequentialy.

Service Account

This API need a service account to access Google Bucket. There are two ways to manage this.

• Use SA_KEY environment variable. Convert json content from .json file to oneline. Save to .env file.
• Use service account .json file. Rename this file to aksacarma-2e5f57e8e8c5.json. (Recommended)