Skip to content


Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Home Brain is an event- and rule-based home automation system.

== The Rules

Sensors are the inputs to Home Brain. These can be direct inputs, like a switch being flipped, or just a way to link into the outside world, such as with the "daylight" sensor (which reports the times of sunrise and sunset, as well as a simple "isDark" boolean).

Devices are the outputs: the light bulbs, the motors, the outlets. Each device class defines any number of axes of control. A switch might just have an on/off boolean, while a color-changing lightbulb might have brightness, hue, and saturation. Each control axis has a default state, defined when an instance of a device is created, so, for instance, in the abscense of any other control, the porch light might turn on while the living room light is off.

Third, there are rules to bind everything together. Any number of devices can be assigned to a rule, as well as how to set any (or all) of the axes on that device. A rule also has an evaluation function to determine whether or not it's active. That function can look at sensors, or anything else, to make its decision.

Each rule, within the "state", has a pre-defined priority. The higher the priority number, the more likely the rule is to have an effect. On each update, the rules are evaluated from lowest to highest priority, building up a new picture of what the world should look like. That means a top-level "lights off, I'm on vacation" will override any normal scheduling, even though both sets of rules are running.

A rule only affects the state of the world if its evaluation function is enabled. They're not really "on" or "off", but instead more "active" or "don't care". This makes it trivial to have rules working on the same devices -- "lights on from 4 to 10" and "lights on for 5 minutes when there's motion" will never conflict. You won't see the motion rule turning the light off at 5:30, since it's only ever responsible for turning lights on. Similarly, if there's motion at 9:58, the lights will stay on until 10:03, even though the 4-to-10 rule implies the should turn off at 10. If there aren't any rules giving a device instructions, it will fall back to its default setting (likely, but not necessarily, off).

== The Events

Once a system is configured with all of its rules, they need to run regularly and, ideally, efficiently. There are two ways that Home Brain's global "update" function will be invoked: by time and by interrupt.

By default, Home Brain will run an update every 30 minutes. However, each rule is allowed to ask for an earlier update. For instance, a "lights on at dark" rule can ask for an update in 13 minutes, if that's when the sun will set. The earliest request update time will be selected, then the server will sleep until then. The update times are only valid for the upcoming cycle, so they're recalculated each go-round (since rules are the best judges of when it's a good time to run). In the example above, if another rule set the update time to 2 minutes, "lights on at dark" would change its timer to 11 minutes during the subsequent update.

Sensors are allowed to interrupt the timing flow and force an immediate update when their values change. The Sensor#changed() method bubbles up to the state and runs the global update. The scheduled update is cancelled, and a new one is scheduled according to the normal timing rules.

== Some Examples


No description, website, or topics provided.






No releases published


No packages published