All output lines begin with a keyword identifying the contents of the line as informational (INFO), warnings (ERROR) or one of the following data output types: RAW, SAMPLE, FRAME, SUM, SPEC, ANC, CALIB. The SUM, SPEC, ANC, CALIB and INFO and ERROR lines are produced as part of the pipeline processing and are contained in the correspondingly named Level 1 files (except for ERROR which is in the .info file) The ERROR lines are sent to stderr rather than stdout. Which data output lines are written is controlled by the settings in ini.params (Sec. 3.1). The INFO and ERROR lines are always written. The RAW format is used primarily for engineering purposes and outputs the WB channel for all samples, regardless of the fringe tracker lock state. Due to the order in which the data are processed during calibration, the SAMPLE output has all samples in a frame for channel 0, followed by all samples for channel 1, etc., until all channels have been printed. Then the next set of samples is printed.
An easy way to create processed data files for plotting or other tasks is to grep the output of Kvis using the desired keyword. A header line is produced for each of the data output types. The headers begin with a # so that data output can be directed to plotting programs such as gnuplot without removing this line. Examples of each data output type are given below and the variables in each column are described in more detail in the tables in Appendix A.
The calibration flag in the sum and spec files has the following possible values:
0 okwhere CalTime is currently set to 10 minutes.
1 time difference between fringe data and background > CalTime
2 time difference between foreground and ratio > CalTime
only applies when using ratio1 and foreground, not ratio1 and ratio2
3 1 and 2 true
4 no ratio data within CalTime
5 1 and 4 true