/sbn-name-resolution

Name resolution service for cross identification of astronomical objects

Primary LanguageHTML

SBN Name Resolution Service

Deployed Openapi Interface

Go here

What's This?

This repo houses scripts and code to build REST API services that enable a user to carry out cross-identification tasks. For example, if you want the official designation for "Halley's Comet" in order to perform a query on some astronomical-data service but do not know it ahead of time, then this service will allow you to perform a "fuzzy word search" (i.e. submit arbitrary text) and receive back candidate matches with meta data describing, e.g., in what domain a particular technical designation would be used.

Code Features

  • PostgresDB
  • Flask API layer
  • Connexion used to generate swagger interface
  • Gunicorn/Apache used for production deployment

Development

Common Steps

This repo has code for:

  • Building the indexed postgresdb
  • Running a flask-connexion API
  • Testing

The following steps are needed to set up the code base whatever aspect you want to work on:

  • The codebase is operated using bash scripts that begin with the \_ underscore character
  • Prerequisites for local development:
    • Running postgresql server with credentials to read/write
    • python (& pip) v3.5+
  • Clone the repo locally: 
        git clone https://github.com/Small-Bodies-Node/sbn-name-resolution
    cd sbn-name-resolution
  • Run cp .env-template .env and edit the variables therein
  • Always begin by source _init_setup.sh. This will:
    • Create/activate a python virtual environment
    • Install dependencies to virtual env
    • Make available to your shell the variables .env

Name-Search-Build

  • To build the indexed tables in your postgresql database, you need to to have the pg_trgm extension made available to your system.
    • If you are on a Mac, then the postgresql installed via homebrew will have this already available
    • If youre on linux, then you'll need to have the postgresql-contrib installed
  • Log into your database and run the following query to enable the relevant indexing: CREATE EXTENSION pg_trgm;
  • You can now run a pipeline using the _name_search_build script to download the files from MPC, clean the data, and load into the postgresql db.

APIs

  • If you have nodemon globally installed, then you can develop your api code and have it automatically update on changes by running _develop_apis. Otherwise, just run python src/api/app.py

Deployment

A script is supplied called _gunicorn_manager that takes the arguments start|stop|status|restart to launch the app as a background process with the gunicorn WSGI server for production serving. The number of workers is controlled with the env variable LIVE_GUNICORN_INSTANCES. If you have trouble getting gunicorn to work, you can run the manager with 0 as the 2nd argument to start it off in non-daemon mode.

It is recommended that you make the gunicorn-powered server accesible to the outside world by proxy-passing requests through an https-enabled web server like apache. Contact me if you need assistance setting that up.