Fixes issue with unnecessarry warning message

Erin (Josh) Rigler authored 6 years ago and Claycomb, Abram Earl committed 6 years ago

AdjustedAlgorithm.process() essentially adjusts two values:

 - 3D vector field HEZ_raw -> XYZ_adj
 - 1D scalar field F_raw -> F_adj

Neither should depend on the other, but this was implicitly the case
as originally implemented. When F was missing, this prevented XYZ
from being calculated, at least when called in 'update' mode by the
Controller. Likewise, if HEZ were missing, this prevented adjusted F
from being calculated.

These changes were deployed directly to production (mage2) on November 8. This is a belated update to the official metadata.

Erin (Josh) Rigler authored 6 years ago and Claycomb, Abram Earl committed 6 years ago

Fixes issue #239, whereby the next_starttime state variable was
incremented even if no input was passed to SqDistAlgorithm.process_one(),
and therefore no processing was performed by SqDistAlgorithm.additive().
Also, modified SqDistAlgorithm.additive() so that the state variables
returned were consistent with the last *input* sample, and NOT with the
last output sample (they could differ if fc input parameter were ever
used).

This should fix the travis linting error

As per discussion in geomag-algorithms issue #236, swap the "state" and "process"
strings in the AlgorithmException string.

Jeremy M Fee authored 6 years ago and Claycomb, Abram Earl committed 6 years ago

Update controller to use algorithm next_start_time to pre-process data, move where args.realtime is processed for testing

Jeremy M Fee authored 6 years ago and Claycomb, Abram Earl committed 6 years ago

Add "without_gaps" option to get_stream_start_end_times, add "pad_and_trim_trace" method and update pad_timeseries to use it