luke-unicorn / augur-node

Public Augur node for serving market data

augur_node

An Augur Node is a public node that caches/serves market data to help with performance. It also provides a public geth node.

To setup a node:

1. Install Docker
2. Clone this repository
3. run docker-compose up -d

Instructions to do this on Ubuntu can be found here: https://github.com/AugurProject/augur-node/blob/master/ubuntu_install_commands.sh

You can test that your augur_node is working by navigating to http://your_server:8547/getMarketsInfo in your browser. There will be a delay between when you spin up your node and when you actually see market data as geth must first sync.

An augur_node does not manage your networking for you, so please verify that ports 8545 (geth rpc), 8546 (geth websockets), 8547 (augur_node) are open on your machine.

endpoints

Your cache node will have the following endpoints available:

getMarketsInfo

filtered mode

http://your_server:8547/getMarketsInfo?branchId=branchId

If branch isn't specified, the default branch will be used.

You can specify if you want only active or inactive markets:

http://your_server:8547/getMarketsInfo?active=true|false If active flag isn't specified, you will get all markets.

By default, the getMarketsInfo info endpoint will return results sorted by volume. You can specify a different sort order if you'd like. For example, http://your_server:8547/getMarketsInfo?sort=newest_market Supported sort options are:

• newest_market
• oldest_market
• most_volume
• least_volume
• soonest_expiry
• furthest_expiry
• lowest_maker_fee
• lowest_taker_fee
• highest_maker_fee
• highest_taker_fee

You can also easily paginate results, and specify how big each page size is. For example, this will return 20 results per page, and give you the second page of results:

http://your_server:8547/getMarketsInfo?sort=soonest_expiry&limit=20&page=2

It also may useful to return sorted results, limited to a specific. This can be done useing the tag param, which will do an exact match on markets that contain the tag that you specify.

http://your_server:8547/getMarketsInfo?limit=20&tag=sports

All of these options mentioned above can be used together or individually to build whatever view into the market data that you desire.

Query Mode

The getMarketsInfo endpoint also supports searching for markets when the query param is included. This paramter works with all of the others mentioned above with the exception of sort, since in query mode results should be ordered by closest match to the queried term. Results are ordered with a preference for matches to description, then tags, then extraInfo.

Examples:

//Returns top 5 results matching "Gary Johnson"
http://your_server:8547/getMarketsInfo?limit=5&query=Gary%20Johnson

//Searches for "2025", limited to markets tagged "climate change"
http://your_server:8547/getMarketsInfo?tag=climate%20change&query=2025

//Searches active markets only for "president", returns page 2 of results.
http://your_server:8547/getMarketsInfo?query=president&active=true&page=2

getTags

If you are interestd in retrieving the top n most popular tags, you can do: http://your_server:8547/getTags?limit=1000

This will return you the top 1000 tags. The page parameter is not supported here due to a limitation in how aggregrations work in elasticsearch.

You can also specify a branchId to return tags from. Default branch will be used if not specified.

Development

You can run this without Docker during development if you'd like, assuming you have elasticsearch (port 9200) and geth running on your local machine configured to accept RPC and websocket requests (ports 8545 and 8546). Your local augur_node can be run by doing node index.js

Doing it this way is fine for development, but be careful using your own geth instance especially if it is used for other purposes, if it contains anything of value, or if you unlock it on occassion. Production augur_nodes should ideally be run on their own machine/vm using docker-compose mentioned above for isolation, predictable deployment, and seamless handling of container networking, dependencies, restarts, etc.

If for some reason you still decide you want to manually run augur_node in production, you should put it in a service (like upstart) so it at the very handles restarts on errors, reboots, etc, and ensure that you are using safe geth settings.

Some additional commands for installing elasticseach on ubuntu can be found here: https://github.com/AugurProject/augur-node/blob/master/ubuntu_setup_elasticsearch.sh

Also, if you change anything in package.json, please run npm shrinkwrap.