Historical sensors for Home Assistant

Feed historical data into Home Assistant database.

HomeAssistant architecture is built around polling (or pushing) data from devices, or data providers, in "real-time". Some data sources (e.g, energy, water or gas providers) can't be polled in real-time or readings are not accurate. However reading historical data, like last month consumption, it is possible and accurate. This module adds support to this.

This module uses the recorder component and custom state creation to store states "from the past".

Current projects using this module:

• ideenergy energy monitor
• AuroraPlusHA

How to implement a historical sensor

💡 Check the delorian test integration in this repository

1. Import home_assistant_historical_sensor and define your sensor. ⚠️ Don't set the SensorEntity.state property. See FAQ below

from homeassistant_historical_sensor import (
    HistoricalSensor, HistoricalState, PollUpdateMixin,
)


class Sensor(PollUpdateMixin, HistoricalSensor, SensorEntity):
    ...

2. Define the SensorEntity.async_update_historical method and save your historical states into the HistoricalSensor._attr_historical_states attribute.

   async def async_update_historical(self):
        self._attr_historical_states = [
            HistoricalState(state=x.state, dt=x.when) for x in api.fetch()
        ]

3. Done Besides other Home Assistant considerations, this is everything you need to implement data importing into Home Assistant

How to add statistics to your Historical Sensor

Generating statistics it is not an easy generalizable job, HistoricalSensor provides some support and helpers but you have to write some code.

1. Define the statistic_id property for your sensor. For simplicity, you can use the entity_id. ⚠️ Don't set the SensorEntity.state_class property. See FAQ below

@property def statistic_id(self) -> str:
    return self.entity_id

2. Define the get_statistic_metadata method for your sensor.

It's recommended to use just "sum" or "mean" statistics, not both, the one which applies to your sensor. Both are shown here just as example.

def get_statistic_metadata(self) -> StatisticMetaData:
    meta = super().get_statistic_metadata() meta["has_sum"] = True
    meta["has_mean"] = True

    return meta

3. Define the async_calculate_statistic_data method for your sensor.

(Check the delorian integration for a full example)

    self, hist_states: list[HistoricalState], *, latest: StatisticsRow |
    None = None,
) -> list[StatisticData]:
   ...

Technical details

Q. How it is accomplished?.

A. It's a relatively easy answer but needs to be broken into some pieces:

1. A new property for sensors: _attr_historical_states. This property holds a list of HistoricalStates which are, basically, a state value and a dt datetime (with tzinfo), so… the data we want.

2. A new hook for sensor: async_update_historical. This method is responsible to update _attr_historical_states property. This is the only function that needs to be implemented.

3. A new method, implemented by HistoricalSensor class: async_write_ha_historical_states. This method handles the details of creating tweaked states in the past and write them into the database using the recorder component of Home Assistant core.

Q. What is PollUpdateMixin and why I need to inherit from it?

A. Home Assistant sensors can use the poll or the push model to update data.

Poll mode does not mix well with historical data, causes blanks and empty points in history and graphs. Because of that, historical sensors use a false push model: historical sensors are never updated by them self.

PollUpdateMixin solves this and automatically provides the poll functionality back without any code. The sensor will be updated at startup and every hour. This interval can be configured via the UPDATE_INTERVAL class attribute.

Q. Why it is recommended to DON'T set the state and the state_class properties?

A. Because it messes up graphs and statistics

• By setting the state attr you are telling Home Assistant that you have a current state, which is not true if you are importing data from the past. You may think that the last state (now - X hours) may be the current state but it is not, in X hours you will import new data and that point may have a different value. Also messes graphs with inconsistent data points. * Setting state_class causes Home Assistant to calculate statistics data which is not what you want since you are working with historical data and not present data. Statistical data will be incorrect if any is calculated. Historical sensors have helper functions to deal with statistics

Q. Why my sensor is not shown in the energy panel configuration?

A. Energy dashboard doesn't use sensors, it uses statistics (it is a bit confusing, yes). Statistics usually have the same name (id) as the source sensor, hence the confusion.

If your sensor doesn't show up in energy panel options it is because it is not generating statistics. For a standard sensor Home Assistant does that job but HistoricalSensor it is not.

HistoricalSensor basically inserts states into the database, using almost raw SQL INSERT stamens, so any internal process of HomeAssistant doesn't apply.

See "How to implement statistics about".

Statistics will be calculated once new historical data comes in, then it will show up in the energy panel.

Q. I can't calculate energy, water o gas costs

A. Actually it can't be done.

Maybe the energy websocket API can be useful

Q. Why is my historical sensor in an "undefined" state?.

A. Historical sensors don't provide the current state. Keep in mind that the last state (from some hours ago) is NOT the current state. The current state is unknown.

Q. I want to provide current state AND historical data

A. You need to implement two sensors: one for the current state and another one for historical data.

If you are following the data coordinator pattern it should be straightforward

External vs. internal statistics

Due to the way the data is inserted a posteriori into the recorder, Home Assistant cannot calculate statistics on historical states automatically. This is particularly problematic for the Energy dashboard, which relies on these statistics.

It is possible to provide those statistics by

1. Providing a get_statistic_metadata method returning a dictionary of supported statistics. Generally, the sum is sufficient for the energy dashboard, so setting has_sum to True is all it takes

2. Providing an async_calculate_statistic_data method to do the calculation.

   It will take a list of HistoricalState, and the latest homeassistant.components.recorder.statistics.StatisticsRow as arguments, and should return an array of homeassistant.components.recorder.models.StatisticData (with start, state and the statistics, e.g., sum).

Importing CSV files

To be implemented: #3

Licenses

• Logo by Danny Allen (Public domain license) https://publicdomainvectors.org/es/vectoriales-gratuitas/Icono-de-configuraci%C3%B3n-del-reloj/88901.html