FlexMeasures is on the Pypi index! Installing can now be done with a simple
pip install flexmeasures
Before, you'd have to clone the code from Github and run a make command. Just installing via pip should amplify the installation process and also make it easier to use on non-Unix platforms.
This work was done in Pull Request 54.
CLI commands to get started
We are also continually working to make it easier to get started with a new project using FlexMeasures (which we also documented better).
The command-line interface (CLI) is crucial for this. There are now four groups of commands:
$ flexmeasures --help
Usage: flexmeasures [OPTIONS] COMMAND [ARGS]...
Management scripts for the FlexMeasures platform.
add FlexMeasures: Add data.
db-ops FlexMeasures: Reset/Dump/Load/Restore/Save the DB.
delete FlexMeasures: Delete data.
jobs FlexMeasures: Job queueing.
Many of these FlexMeasures commands existed before and now it's easier to find them. There are two new ones:
flexmeasures add asset
flexmeasures add weather-sensor
Now it's possible to add a user, assets and weather sensors right after installing FlexMeasures, so a new project has all necessary data so adding measurements, forecasts and schedules can begin (we also add a dummy market now as part of
flexmeasures add structure).
Finally, note that there are more useful commands available, from underlying libraries, for example for inspecting available API routes or to handle database migrations.
This work was done in Pull Request 74.
Set belief time of data in API: "prior"
We increased clarity for a crucial aspect about time series data (we call them "beliefs"): When was the data recorded / known (the "belief time")? Next to
horizon, the FlexMeasures now understands the new field
prior, both for GETting and for POSTing beliefs through the API. I'll cite from the FlexMeasures API documentation about GETting beliefs using the two methods:
The "prior" field (a timestamp) can be used to select beliefs recorded before some fixed moment in time. It can be used to "time-travel" to see the state of information at some moment in the past.
In contrast, the "horizon" field (a duration) can be used to select beliefs recorded before some moment in time, relative to each event.
Also, we retired the concept of denoting rolling (recurring) horizons, e.g.
R/PT6H. We only allow
PT6H now to denote a 6 hour horizon and all horizons are interpreted as rolling horizons if the context fits.
This work was done in Pull Request 41.