GitLab

Import seitosh

@wittkamp @lgass @r 
[libstfinv](https://git.scc.kit.edu/Seitosh/Seitosh/tree/master/src/libs/libstfinv) in [Seitosh](https://git.scc.kit.edu/Seitosh/Seitosh) provides a new feature:

A new taper option for procedures operating in the Fourier domain is implemenetd in libstfinv. This can be used to remove unwanted acausal or coda components of the impulse response of the source time function correction filter. See usage information provided by libstfinv:

    A time domain taper can be applied to the impulse response of the correction
    filter by using option irtap. Four time values are given in units of seconds:
    t1, t2, t3, and t4. They must be in increasing order and (t4-t1) must be
    smaller than the total duration of the time series used to represent signals
    internally. Times value are allowed to be negative. Time series are understood
    to be periodic (due to discrete Fourier transformation). Prior to application
    of the correction filter to the time series passed to the algorithm, the
    correction filter is transformed to the time domain, tapered, and then
    transformed to the Fourier domain again. The values of the taper are:
      0                              if       t <  t1
      0.5-0.5*cos(pi*(t-t1)/(t2-t1)) if t1 <= t <= t2
      1                              if t2 <  t <  t3
      0.5+0.5*cos(pi*(t-t3)/(t3-t4)) if t3 <= t <= t4
      0                              if       t >  t4

Test-cases are set up for [soutifu](https://git.scc.kit.edu/Seitosh/Seitosh/blob/master/src/ts/wf/soutifu.cc) and are described in the accompanying [README.soutifu](https://git.scc.kit.edu/Seitosh/Seitosh/blob/master/src/ts/wf/testcases/README.soutifu).

A vendor update has taken place in [thof/IFOS2D](https://git.scc.kit.edu/thof/IFOS2D) see [issue 5](https://git.scc.kit.edu/thof/IFOS2D/issues/5) there. The code in contrib compiles and installs well. The integration into IFOS2D still is to be tested.

See merge request !3

Merge branch 'vendor_Seitosh' into import_Seitosh

Conflicts:
	contrib/fourier/Ffcommand.f
	contrib/fourier/Ffilters.f
	contrib/fourier/Fpolesnzeroes.f

Commands executed are:

    git merge -X theirs vendor_Seitosh
    git rm contrib/fourier/*.f

Fortran files are removed intentionally, since they are not used in IFOS2D.
This way IFOS2D does not need a Fortran compiler to be installed.

Drop current version of Seitosh code. The contents are taken vom branch master
of [Seitosh](https://git.scc.kit.edu/Seitosh/Seitosh). The current master
commit in Seitosh is 890d2fe565.

With this vendor drop the new taper feature as implemented in libstfinv
becomes available. Source time function correction filters produced in the
Fourier domain can be tapered in the time domain prior to application.

Commands executed in the root directory of IFOS2D:

    rm -rfv contrib
    mkdir -pv contrib
    (cd ~/git/Seitosh/src/libs/libaff; tar cvf - $(git ls-files)) \
      | ( cd ~/git/IFOS2D/contrib; mkdir aff; cd aff; tar xvfps -)
    (cd ~/git/Seitosh/src/libs/libfourier; tar cvf - $(git ls-files)) \
      | ( cd ~/git/IFOS2D/contrib; mkdir fourier; cd fourier; tar xvfps -)
    (cd ~/git/Seitosh/src/libs/libseife; tar cvf - $(git ls-files)) \
      | ( cd ~/git/IFOS2D/contrib; mkdir libcseife; cd libcseife; tar xvfps -)
    (cd ~/git/Seitosh/src/libs/libstfinv; tar cvf - $(git ls-files)) \
      | ( cd ~/git/IFOS2D/contrib; mkdir stfinv; cd stfinv; tar xvfps -)
    git add --force contrib

Since Seitosh on purpose does not provide export branches where source code
exports would be assembled in a hand-picked fashion, whole subdirectories from
Seitosh are simply copied. As a conseqeunce, the directory libcseife now also
contains Fortran code not really required by IFOS2D.

I removed the zero phase option from the code, due to complications depending on the source signal and the time window. However, from a technically point of view it was implemented right.

Position 7 in workflow is now F_HIGH_PASS.
All other parameters have been shifted.

The model, which is used in the step length estimation is outputted if VERBOSE is set to 1.

The sinus3 source wavelet was shifted in the wrong direction.

Added references to D. Koehn's dissertation. Large parts of the theory are extracted from his dissertation.

Position 7 in workflow is now F_HIGH_PASS.
All other parameters have been shifted.

The vector FL was allocated before the value L was exchanged.

This fix will reset the conjugate gradient history each time a frequency or workflow stage is switched.
If such a stage is switched we will have a "new" objective function, thus, it makes no sense to use the old gradients for speeding up the convergence.
This bug was the reason why the objective function could not be reduced for 2-3 iterations after a new frequency stage was appplied.

I switched the naming of the variables of the frequency filtering to a more clear and easy one.
The change affects the JSON input file and the code itself.
The following renaming is used:
FC_START -> F_LOW_PASS_START
FC_END   -> F_LOW_PASS_END
FC_INCR  -> F_LOW_PASS_INCR
FC_EXT   -> F_LOW_PASS_EXT
FC       -> F_LOW_PASS
F_HP     -> F_HIGH_PASS

However, old JSON files will also work for now. If one of the old variables is not found, IFOS will search for the corresponding old one. This option will be removed in a future release.

This option is not supported anymore. Therefore I remove it from the source code.

With the option TRKILL_STF_OFFSET_INVERT it is possible to invert the sense of the STF Offset window.
So TRKILL_STF_OFFSET_INVERT==1 means, do NOT kill the specified offset, instead kill all other offsets.

# Conflicts:
#	src/IFOS2D.c
#	src/wavelet.c