nsoblath
commented
7 years ago - [x] Fix setting of digitizer parameters (in particular DAQ gain) in streaming_writer and egg_writer
- [x] Tell Katydid that it's dealing with FFTW time series
- [x] Figure out how to make Katydid recognize FFTW time series by itself
- [x] Fix how FFTW time series are handled at the RawTimeSeries level -- this is the underlying cause of weird memory errors that inconsistently appear
katydidoutput.txt
When I created fake data and wanted to process it with katydid I had some data type issues too. But the error then clearly stated that katydid wanted complex 32 float. So I guess this is now at an earlier level.
Katydid starts of very promising: 22:41:09 [DEBUG] rch3/M3Stream.cc(173): Number of acquisitions found: 102; Number of records found: 1867
It also gets the header and the only warnings appearing are: 22:41:09 [ WARN] gleChannelDAC.cc(229): level <129> = 0
and later: 22:41:09 [ WARN] KTForwardFFTW.cc(158): Unable to read FFTW wisdom from file
Katydid also gets the slice headers and has all the information about the slice but then: 22:41:09 [ INFO] KTForwardFFTW.cc(367): binwidth 1e-08 22:41:09 [ INFO] KTForwardFFTW.cc(377): iComponent0 22:41:09 [ERROR] KTForwardFFTW.cc(378): Incorrect time series type: time series did not cast to KTTimeSeriesFFTW. 22:41:09 [ERROR] cessor/KTSlot.hh(322): Something went wrong while analyzing data with type
The lines in ForwardFFTW producing this are: { const KTTimeSeriesFFTW nextInput = dynamic_cast< const KTTimeSeriesFFTW >(tsData.GetTimeSeries(iComponent)); if (nextInput == NULL) { KTERROR(fftwlog, "Incorrect time series type: time series did not cast to KTTimeSeriesFFTW."); return false; }