kerenvascs / openaq-api

OpenAQ Platform API

OpenAQ Platform API

Overview

This is the main API for the OpenAQ project.

Starting with index.js, there is a web-accessible API that provides endpoints to query the air quality measurements. Documentation can be found at https://docs.openaq.org/.

openaq-fetch takes care of fetching new data and inserting into the database. Data format is explained in openaq-data-format.

Getting started

Install prerequisites:

• git
• nvm
• Docker

Clone this repository locally (see these instructions) and activate the required Node.js version with:

nvm install

The last step can be skipped if the local Node.js version matches the one defined at .nvmrc.

Install module dependencies:

npm install

Development

Initialize development database:

npm run init-dev-db

This task will start a PostgreSQL container as daemon, run migrations and seed data. Each of these tasks is available to be run independently, please refer to package.json to learn the options.

After initialization is finished, start the development server:

npm run dev

Access http://localhost:3004.

Stop database container after finishing:

npm run stop-dev-db

Testing

Initialize test database:

npm run init-test-db

This task will start a PostgreSQL container as daemon, run migrations and seed data. After initialization is finished, run tests:

npm run test

Stop database container after finishing:

npm run stop-test-db

Deploying to production

The server needs to fetch data about locations and cities in the measurement history using AWS Athena. This service must be configured via the following environment variables:

• ATHENA_ACCESS_KEY_ID: an AWS Access Key that has permissions to create Athena Queries and store them in S3;
• ATHENA_SECRET_ACCESS_KEY: the corresponding secret;
• ATHENA_OUTPUT_BUCKET: S3 location (in the form of s3://bucket/folder) where the results of the Athena queries should be stored before caching them;
• ATHENA_FETCHES_TABLE: the name of the table registered in AWS Athena.

Automatic Athena synchronization is disabled by default. It can be enabled by setting the environment variable ATHENA_SYNC_ENABLED to true. The sync interval can be set using ATHENA_SYNC_INTERVAL variable, in miliseconds. The default interval is set in file config/default.json.

If needed, the synchronization can be fired manually. First, the WEBHOOK_KEY variable must be set to allow access webhooks endpoint. Sending a POST request to /v1/webhooks, including the parameters key=<WEBHOOK_KEY> and action=ATHENA_SYNC, will start a sync run. An example with curl:

curl --data "key=123&action=ATHENA_SYNC" https://localhost:3004/v1/webhooks

Other environment variables available:

Name	Description	Default
API_URL	Base API URL after deployment	http://:3004
NEW_RELIC_LICENSE_KEY	New Relic API key for system monitoring	not set
WEBHOOK_KEY	Secret key to interact with openaq-api	'123'
USE_REDIS	Use Redis for caching?	not set (so not used)
USE_ATHENA	Use AWS Athena for aggregations?	not set (so not used)
REDIS_URL	Redis instance URL	redis://localhost:6379
DO_NOT_UPDATE_CACHE	Ignore updating cache, but still use older cached results.	not set
AGGREGATION_REFRESH_PERIOD	How long to wait before refreshing cached aggregations? (in ms)	45 minutes
REQUEST_LIMIT	Max number of items that can be requested at one time.	10000
UPLOADS_ENCRYPTION_KEY	Key used to encrypt upload token for /upload in database.	'not_secure'
S3_UPLOAD_BUCKET	The bucket to upload external files to for /upload.	not set

AWS Athena for aggregations

The Athena table is fetches_realtime that represents the fetches from openaq-data and has the following schema:

CREATE EXTERNAL TABLE fetches.fetches_realtime (
  date struct<utc:string,local:string>,
  parameter string,
  location string,
  value float,
  unit string,
  city string,
  attribution array<struct<name:string,url:string>>,
  averagingPeriod struct<unit:string,value:float>,
  coordinates struct<latitude:float,longitude:float>,
  country string,
  sourceName string,
  sourceType string,
  mobile string
 )
 ROW FORMAT SERDE 'org.openx.data.jsonserde.JsonSerDe'
 LOCATION 's3://EXAMPLE_BUCKET'

Uploads & Generating S3 presigned URLs

Via an undocumented /upload endpoint, there is the ability to generate presigned S3 PUT URLs so that external clients can authenticate using tokens stored in the database and upload data to be ingested by openaq-fetch. There is a small utility file called encrypt.js that you can use like UPLOADS_ENCRYPTION_KEY=foo node index.js your_token_here to generate encrypted tokens to be manually stored in database.

Dockerfile

There is a Dockerfile included that will turn the project into a Docker container. The container can be found here and is currently mostly used for deployment purposes for AWS ECS. If someone wanted to make it better for local development, that'd be a great PR!

Contributing

There are a lot of ways to contribute to this project, more details can be found in the contributing guide.

Projects using the API

• openaq-browser site | code - A simple browser to provide a graphical interface to the data.
• openaq code - An isomorphic Javascript wrapper for the API
• py-openaq code - A Python wrapper for the API
• ropenaq code - An R package for the API

For more projects that are using OpenAQ API, checkout the OpenAQ.org Community page.