A softwarer for structuring, visualizing and monitoring of hierarchical data systems.
Many systems in real-life can be viewed as multiple sets of key-value pairs, with hierarchical relations between those sets. An example would be:
Smart Home
├── Living Room [temperature]
│ ├── Smart TV [power, channel, volume]
├── Kitchen [temperature]
│ ├── Smart Oven [temperature, timer]
└── Bathroom [temperature, moisture]
├── Smart Mirror [light on/off]
└── Smart Light [brightness, hue, effect]
As shown, nodes on every level can have their own attributes, as well as children. Another example:
Person
├── Languages
│ ├── Python [lines of code written]
│ ├── C/C++ [segfaults caused, rockets crashed]
├── Projects
│ ├── BLEnky
│ ├── Structure thing
└── Presentations
├── Web&Wine
│ ├── Hacking a closed source smartwatch
│ ├── Node-red
└── Hackerkiste
└── Hacking a closed source smartwatch
In structure-thing, these nodes look like this:
As evident, a node can have a set of attributes, while also containing further children.
To get real-time data into the software, its API can be used. Using the API, the following actions can be performed:
- Setting/Getting single values
- Setting/Getting data for a whole set
- Setting/Getting data for multiple sets based on filters
- Getting data in real-time through websockets
Furthermore, structure-thing provodes the following functions:
- Managing external access to nodes/subtrees with granular permissions
- Filtering incoming data in order to reject values based on JSONata expressions, for instance invalid json documents
- Transforming incoming data, for examples to convert units or to extract values from a JSON document.
- Setting up alert rules for data, for example when a temperature drops below 10° (
$ < 10) - Setting up frontend transformation, for example to add a unit sign:
$ & '°'
At this point it should be mentioned that the software if context-agnostic. As such, neither does it understand a temperature or any other information, nor is it able to retrieve that information by itself. To get any useful live information into the system, usually a middleware is required in order to transfer said information. Here is an example:
Structure-thing <-> connector <-> Stock prices API
If some foreign software sends out JSON webhooks, they can be send directly to structure-thing though. Via transformation the needed data can be extracted and forwarded to the system.
Here is an example:
The Things Network --> Structure-thing
Telegram Bot Api --> Structure-thing
Transformation: data.value
If you want to test the Telegram bot, here you go.
I use this software in order to control my smart home. A close-to-real demo can be viewed here.
Most of my devices are connected using BLE through BLEnky. As such, the infrastructure needs MQTT to BLE bridges, in order to connecte
the raspberry pi to BLE devices through WiFi. Each bridge needs a configuration containing the BLE devices to connecte to.
Since the topology of rooms/devices is known through the hierarchical structure in structure-thing, bridge configuration get derived automatically on change and pushed to the bridges.
If I move a device from one room to another in the software, the devices MAC address gets deleted on one bridge and appended to another.
My projects and presentations can be viewed here.
If you are interested in what electronical and mechanical parts I have in stock, you can check out this.
All services are provided as docker-compose. As such, they can be installed like this:
git clone git@gitlab.com:structure-thing/docker.git structure-thing
cd structure-thing
docker compose up -d
docker inspect structure-nginx-1
# Then visit the IPAddress of the nginx container in your browser.
# This should give you one shot to set up a root account