The Draco Database contains the following collections and relationships
- Farm
- Trubine
- Farm
- Signal
- Sensor
- Turbine
- Signal
- Reading
- Sensor
- PipelineTemplate
- Pipeline
- PipelineTemplate
- MLTask
- Turbine - multiple
- Target
- MLTask
- Experiment
- MLTask
- PipelineTemplate
- Signal - multiple
- ExperimenRun
- Experiment
- PipelineRun
- Pipeline
- ExperimentRun
A Farm represents a physical Wind Turbines Farm. This collection groups together multiple Turbines with shared properties, such as location.
_id (ObjectID)
: Unique Identifier of this Objectname (String)
: Name or code given to this Objectinsert_time (DateTime)
: Time when this Object was insertedcreated_by (String)
: Identifier of the user that created this Object
A Turbine represents a physical Turbine. A Turbine is part of a Farm, and has some particular properties, such as the Turbine manufacturer.
_id (ObjectID)
: Unique Identifier of this Objectfarm_id (ObjectID)
: Unique Identifier of the Farm to which this Turbine belongsname (String)
: Name or code given to this Objectmanufacturer (String)
: Name or code of the manufacturer - EXAMPLEmodel (String)
: Name or code of the model - EXAMPLEinsert_time (DateTime)
: Time when this Object was insertedcreated_by (String)
: Identifier of the user that created this Object
The Signal collection contains the details about each Signal type. This includes shared properties of the signal, such as the sensor type or the measurement units.
_id (ObjectID)
: Unique Identifier of this Objectname (String)
: Name or code given to this Objecttype (String)
: Type of sensor - EXAMPLEcreated_by (String)
: Identifier of the user that created this Objectinsert_time (DateTime)
: Time when this Object was inserted
A Sensor represents a physical sensor that is installed in a Turbine. The Sensor collection specifies the turbine and the signal type, as well as properties about the Sensor such as the Sensor manufacturer and model and its age.
_id (ObjectID)
: Unique Identifier of this Objectturbine_id (ObjectID)
: Unique Identifier of the Turbine where this Sensor is installedsignal_id (ObjectID)
: Unique Identifier of the Signal type of this Sensorname (String)
: Name or code given to this Objectmanufacturer (String)
: Name or code of the manufacturer - EXAMPLEmodel (String)
: Name or code of the model - EXAMPLEinstallation_date (DateTime)
: Time when this Sensor was installed - EXAMPLEinsert_time (DateTime)
: Time when this Object was insertedcreated_by (String)
: Identifier of the user that created this Object
The Readings collection contains all the data readings from a Sensor.
_id (ObjectID)
: Unique Identifier of this Objectsensor_id (ObjectID)
: Unique Identifier of the Sensor to which this Reading belongs.timestamp (DateTime)
: Time where this reading took placevalue (float)
: Value of the reading
The PipelineTemplate collection contains all the pipeline templates from which the pipelines that later on will be used to run an experiments are generated. The template includes all the default hyperparameter values, as well as the tunable hyperparameter ranges.
_id (ObjectID)
: Unique Identifier of this PipelineTemplate objectname (String)
: Name or code given to this Objecttemplate (SubDocument)
: JSON representation of this pipeline templateinsert_time (DateTime)
: Time when this Object was insertedcreated_by (String)
: Identifier of the user that created this Object
The Pipeline collection contains all the pipelines registered in the system, including their details, such as the list of primitives and all the configured hyperparameter values.
_id (ObjectID)
: Unique Identifier of this objectname (String)
: Name or code given to this Objectpipeline_template_id (ObjectID)
: Unique Identifier of the PipelineTemplate used to generate this pipelinepipeline (SubDocument)
: JSON representation of this pipeline objectinsert_time (DateTime)
: Time when this Object was insertedcreated_by (String)
: Identifier of the user that created this Object
An MLTask is a specific Machine Learning Problem consisting on a prediction that is to be made using a Pipeline.
_id (ObjectID)
: Unique Identifier of this objectname (String)
: Name or code given to this Objectdescription (String)
: Short text description of this tasktype (String)
: Type of Machine Learning Taskturbine_set (List[ObjectID])
: List of IDs of the Turbines to which this MLTask is appliedinsert_time (DateTime)
: Time when this Object was insertedcreated_by (String)
: Identifier of the user that created this Object
The Target collection contains the MLTask targets with their cutoff times.
_id (ObjectID)
: Unique Identifier of this Objectmltask_id (ObjectID)
: Unique Identifier of the MLTask to which this target belongsturbine_id (ObjectID)
: Unique Identifier of the Turbine associated with this targetcutoff_time (DateTime)
: Time associated with this Target
An Experiment represents the process of trying and tuning a PipelineTemplate in order to solve a MLTask.
_id (ObjectID)
: Unique Identifier of this Objectname (String)
: Name or code given to this Objectmltask_id (ObjectID)
: Unique Identifier of the MLTask to which this Experiment belongspipeline_template_id (ObjectID)
: Unique Identifier of the PipelineTemplate used in this Experimentsensor_set (List[ObjectID])
: List of IDs of the Sensors used for this Experimentcv_folds (integer)
: Number of folds used for Cross Validationstratified (bool)
: Whether the Cross Validation was stratified or notrandom_state (integer)
: Random State used for the Cross Validation shufflingmetric (string)
: Name of the metric usedinsert_time (DateTime)
: Time when this Object was insertedcreated_by (String)
: Identifier of the user that created this Object
An ExperimentRun represents a single execution of an Experiment.
_id (ObjectID)
: Unique Identifier of this Objectexperiment_id (ObjectID - Foreign Key)
: Unique Identifier of the Experimentstart_time (DateTime)
: When the execution startedend_time (DateTime)
: When the execution endedsoftware_versions (List of Strings)
: version of each python dependency installed in the virtualenv when the execution took placebudget_type (String)
: Type of budget used (time or number of iterations)budget_amount (Integer)
: Budget amountstatus (String)
: Whether the ExperimentRun is still running, finished successfully or failedinsert_time (DateTime)
: Time when this Object was insertedcreated_by (String)
: Identifier of the user that created this Object
A PipelineRun represents a single execution of a Pipeline instance over a MLTask.
It contains information about whether the execution was successful or not, when it started and ended and the cross validation score obtained.
_id (ObjectID)
: Unique Identifier of this Objectexperimentrun_id (ObjectID)
: Unique Identifier of the ExperimentRun to which this PipelineRun belongspipeline_id (ObjectID)
: Unique Identifier of the Pipelinestart_time (DateTime)
: When the execution startedend_time (DateTime)
: When the execution endedscore (float)
: Cross Validation score obtainedstatus (String)
: Whether the Signalrun is still running, finished successfully or failedinsert_time (DateTime)
: Time when this Object was inserted