Data Filtering - OpenDiabetes/OpenDiabetesVault-engine GitHub Wiki

For data filtering we provide the Filter interface. Filters can be registered at processing services like the DataSlicer and have to implement the Filter interface. Multiple filters are combined with the logical "and" paradigm. As shown in the figure below, filters take a data array together with options. The filter will delete not matching data from the array and returns it together with a list of time spans where continuous data is present (this could be used to find the data in the complete array again). Filter block diagram

We identified some different types of filters. Nevertheless, this has no influence on the interface:

  • Time-Related Filter: This kind of filters are used to find a time point or a time span where some kind of event happend or some kind of data is available or not available. For instance, if you want to check the basal insulin rate you search for time spans where no bolus is given. Filters of the time point type should provide a margin option enable data observation around the found point. Filter Scribble

  • Threshold-Related Filter: This kind of filter are used to find data over or under a certain threshold. Filter Scribble