GitLab

- Allowing users to define analyzers with inputs but no outputs
- In this case, a "sink" unit is created, which always refers to the
root of the sensor tree
- This is useful for control analyzers that tune system knobs but do
not actively produce sensor data

- For plugins that use hierarchical units (e.g., job analyzers) a
"flatten" plugin option was added. If enabled, output sensors in the
sub-units of each unit will be exposed as well, and their output values
propagated
- Minor fixes here and there

- on sampling event record current object file + function name
- add numa-node and cpu info

- If getImportances is set to true, the regressor will print a sorted
list of features and their importances, after performing training
- A REST action to retrieve the same list has been added

- When the start and end timestamps of a query are equal, it is enforced
that only one sensor value will be retrieved

- Staleness checks when performing sensor queries in the cache are
now enforced both in relative and absolute modes
- Fixed a race condition causing wrong results when performing queries
in absolute mode
- Changed logging in the TesterAnalyzer plugin

- Switched internal buffers to ureading_t types to allow the SensorBase
class to perform all basic overflow and wrap-around checks
- Managing overflows explicitly in the SAR parser, as deltas need to be
calculated beforehand

- Fixed install of BACNet headers on MAC OS
- Fixed install_conf target in analytics

- Test plugin that can be used to evaluate the performance and functionality
of the query engine and unit system

- use std::atomic_flag based spin lock for thread safety
- minor fixes and improvements

Pusher: Caliper plugin preparations: introduce sensor mutex to allow for runtime addition of sensors

- Also fixed a few compilation warnings

- An OpenCV-based regression plugin using Random Forest models
- Also added a few missing copy constructors

- A rather large library for machine learning, data analysis, filtering
and so on, that will likely cover ALL use cases we will ever have for the
data analytics framework