ESRA-Backend

This repository is for the backend of ESRA using Django connect with Elasticsearch.

Repository Struture

[backend]
    ├── settings.py     :   setting port, host, database, and etc from django
    └── ...
[esra_backend]
    ├── documents.py    :   elasticsearch analyser and registry document
    ├── helps.py        :   elasticsearch query services
    ├── models.py       :   data model
    ├── serializer.py   :   data serializers
    ├── urls.py         :   api url lists 
    ├── views.py        :   api logics
    └── ...
[synonyms]
    ├── synonyms.py     :   create synonym.txt using embedding vector list of all entities
    └── synonyms.txt    :   file for put in elasticsearch for synonyms feature
.env                    :   enviroments same as in .env.example
new_seed.py             :   dumping items to database
requirements.txt        :   list of all libraries
DockerFile              :   for running with gunicorn
start.sh                :   for running with gunicorn
esra.sqlite3            :   our database file (download: https://bit.ly/35eEFLt)
...

Elasticsearch

1. Create directory mkdir elasticsearch-library
2. Download elasticsearch based-on your system to that directory [ https://www.elastic.co/downloads/elasticsearch ]
3. Extract file tar -xzf elasticsearch-{version}.tar.gz
4. Get in to directory cd elasticsearch-library
5. To start elasticsearch server run this command: ./elasticsearch-{version}/bin/elasticsearch

To use the synonym, please follow the instruction below in Run the project with elascticsearch section.

Enviroment

In order to run the project, you need to create .env file (you can look the example at .env.example).

Administrator

In order to create super user of Django, run this command:

python manage.py createsuperuser

Then put your username, email, and password respectively. You can access the admin at localhost:8000/admin after run the project.

Run the project with elascticsearch

After you clone the project, please follow the two steps before run the application:

1. Copy ./synonyms/synonyms.txt (project directory) into ./elasticsearch-{version}/config/analysis (elasticsearch directory).
2. Download a database file (https://bit.ly/35eEFLt) and add it to the project directory as shown in the repository struture.

Run the following commands:

python manage.py makemigrations
python manage.py migrate
python manage.py search_index --rebuild
python manage.py runserver

Using shell

python manage.py shell

Serve Django application

We use gunicorn for serving our backend application. First install gunicorn(this is not fixed, pick anything that suit your liking) then run the following command:

gunicorn --bind 0.0.0.0:8000 --workers 8 backend.wsgi

This command will create 8 workers and binded them at port 8000

Docker image

The docker image of this project is publicly available at backgroundboy/esra-backend dockerhub repository. The image requires an environment variable ELASTICSEARCH_HOST for the name of external elasticsearch host in order to connenect and run.