- Requires Mongodb. You will need to start a mongo server from the localhost:
$mongod --dbpath =XXXX etc... - Once the server launched, use main.js to get the info back from the requests and populate the database.
$node main.js - Use compare.js to start comparing the latest info from the various exchanges.
$node compare.js - volCrypto.js is a work in progress and probably doesn't work anymore. The objective being to create some kind of a pricing model in the future.
Each Exchange has its own collection and related model. We tried to include as much information as possible even if not every exchange provides the same amount of info.
-
Market Requests:
- The error logs are not working properly ==> NOT WORKING FOR SOME OBSCURE REASON. NOT VITAL TO THE PROGRAM BUT IT WOULD BE GREAT IF WE COULD INCLUDE SOME KIND OF AUTOMATIC MAIL SENT FROM THE SERVER IN CASE OF A PROBLEM.
- Not able to include Exmo market yet. Problem with the returned data: Gives an array that I would have to save each items separately. The solution might reside in giving a Map function to save each ticker separately... Still have to work on it.
- The requests are not optimized as is since I'm only doing requests on a single market. There's no need for an array anymore but at this point, I don't feel like redoing the code to delete the list option (array). Plus, it might be usefull in the future... The only thing is that for the order book I didn't adapt. The "list" is "hardcoded"
-
main.js:
- Improve the timeStamps since Mongoose supports it directly within the models: https://stackoverflow.com/questions/10006218/which-schematype-in-mongoose-is-best-for-timestamp.
-
compare.js:
- Working on taking the positions. Orders were included in a dedicated collection: next step is trying to compare volumes and select the adequate markets. WORK IN PROGRESS
This is currently running on a Raspberry Pi 3 Model B with pi64 OS (for the mongodb support): https://github.com/bamarni/pi64. I seem to be having troubles with the Raspberry Pi and Mongodb. Everything works correctly up to some point. I believe that the raspberry memory is too low and that Mongodb crashes when it fills the RAM... -- UPDATE: Definitely fills up the RAM. The more information the raspberry is working with the faster the mongo instance crashes... --