raptorbox / raptor Goto Github PK

When retrieving tokens (/auth/token) the api complain about the serialization of the owner` property. It may be mapped to a getter and return just the user UUID.

Add Spring based test for DI services in auth service

Interfaces and configuration are tied together and should be instead refactored to allow a more modular approach, where extension can happens just via plugging a new library

To evaluate:

• SPI
• OSGI
• 3rd party libraries, eg decebals/pf4j

do you have cluster deploy document?

Review sec-related API access for consistency

See ObjecManagementTest.java in raptor-client

Dispatcher should dispatch to the parent group/device parents. Example topic

root/leaf1/leaf2/device1

Sync call must return 2xx response, object creation / deletion must fail otherwise

Currently avail at https://github.com/muka/raptorjs

@see http://docs.raptorbox.eu

After modification in device definition, redis cache needs to clean manually. It should be added to application bootup configuration to remove the cache automatically to avoid errors later.

Support JWT for token genration and validation

• Integrate security API with current Raptor API #13
• Support JWT #10
• Add user management API #9
• Disconnect on permission change #7

Documentation

• #12 Generate Swagger documentation for Raptor API
• #11 Generate Swagger documentation for security API
• #20 Update and migrate previous documentation to Jekill / github.io

SDK

• #4 Java client SDK
• #15 Javascript client SDK

Deployment & Testing

• #18 Docker based deployment
• #17 Docker compose support

An admin endopoint to batch rebuild ES index based on stored data

Split data stream api and object definition api reusing same structure but making two different services out of it

Check that build pass:

• docker-compose start properly?
• Tests passes?

Enable support continuous integration

• Drop support for query
• Add support for multiple database concurrently
• remove dependecy from storage as it is in favour of the indexer

Initial setup of user management, roles and token

Currently it is possible to set parent / children only via dedicate API

Make reference to timestamp consistent in the codebase

Auth service should be connected via broker to the data / object API

The java API is completely missing the in code documentation

Access tree to objects should be inherited from parents

With a tree like this A / B / C where user u1 can read data from A, they will be able to read data from B and C too

See https://issues.oasis-open.org/browse/MQTT-259

Allow for dynamic group building

To delete all the data, the current implementation load the whole dataset and deletes it.
Should be implemented in a more performant way

Use bintray for initial delivery

Add disconnect from broker when Token, Permission or User is disabled

Keep track of the user authoring the data based on its apiKey

Add the deployable jar as docker images in the docker hub

When creating a device permission are created asynchrounusly, this create two side effects

1. Fast clients may get 403 on request (solved by retrying request, as done in the SDK)
2. In some conditions there are chances the message will get lost causing the ACL not being created at all and making the new device unusable, unless it is manually registered with /auth/sync

To avoid that we may use QOS 2 if not already the case or force the ACL creation request to succeed before having the device created

Set the TTL option in settings to allow fine grained data duration configuration

Current search API is pretty cumbersome and limited.

• Review the pull / pull lastupdate feature
• Review and optimize how search and multi-parameter search are managed

HTTP API to manage the groups

Use case:
I want to show online data streaming, I need to expose an api key for read only on a device stream

Integrate Java based client to the repository, leveraging on raptor-model sub project

Needs to fix Swagger documentation

Add debian package support

Current data structure is verbose and complex to handle at client side

[{
  channels: {
    channelName: {
       current-value: <channel value>
    }
  },
  lastUpdate: 1234567890
}]

The new structure will have a data field containing a map of key/value and a timestamp value in place of lastUpdate

[{
   data: {
    channelName: <channel value>
  },
  timestamp: 1234567890
}]

Group

• Management HTTP API #5
• Handle groups in dispatcher #6
• Match hierarchy between objects/ACL #22

Raptor goes home!

There should be a way to store application related metadata for an user account

Support kubernetes deployment configuration and scaling mechanins