Wavepacket interface (0.1.1b)
==============================
Features:
- little interface for solving any hermitian stationary/non-stationary 1-dim. Schroedinger equations with "Wavepacket (vers. 0.1.1b)":
https://sourceforge.net/projects/cpp.wavepacket.p/
- non-hermitian and multi-dim. systems are supposed to work as well but need more reformulations and testing of working examples
- external parameters of potentials can be controlled from external configuration files (reduced need for compilation when changing a systems configuration)
- compatibility with latest Wavepacket vers. >= 0.1.2 needs yet to be checked
- for installation see below

Wavepacket (vers. 0.1.1b) prerequisites
=======================================
- Installation
--------------
1) install cmake (via wget, ./bootstrap, make, make install)
2) if no root permissions, this might be necessary: add ".cshrc" or ".<other dependent on used shell (check with "echo $SHELL")>" to home folder and write:
2a) setenv LD_LIBRARY_PATH <e.g.: /share/apps/gcc-4.9.1:$LD_LIBRARY_PATH>
setenv LD_LIBRARY_PATH <e.g.: /share/apps/gcc-4.9.1/lib:$LD_LIBRARY_PATH>
setenv LD_LIBRARY_PATH <e.g.: /share/apps/gcc-4.9.1/lib64:$LD_LIBRARY_PATH>
3) download/clone mps-bundle
4) run: "make cblapack" (C version of lapack for f2c library) in directory "/mps-bundle" and then "make build"
5) download/clone googletest into "/mps-bundle/mps/test/googletest" and run "cmake ." in there 
6) download (wget *.tar.gz) and unpack (via "tar -zxvf <file-path>") fftw and change into its root directory
6a) run: ./configure --prefix="<path into mps-bundle root folder>" (prefix needed if no root permissions available)
6b) run: make 
6c) run: make install
7) download boost into include folder "<boost>" and unpack 
8) download wavepacket into include folder (where other libraries are), create folder "<wavepacket...>", unpack and change into root directory
8a) change Makefile.config: ...
8b) g++ version >= 4.7 : use e.g. "CXX=/share/apps/gcc-4.9.1/bin/g++"
8c) add lines (after "GTestFlags ..."): 
"
BOOST_CPPFLAGS := -I$(LIBS_SRCDIR)/boost_1_62_0 
TENSORLIB_DIR := $(LIBS_SRCDIR)/mps-bundle#(!)
TENSORLIB_CPPFLAGS := -I$(TENSORLIB_DIR)/include 
ExtraCppFlags :=  $(BOOST_CPPFLAGS) $(TENSORLIB_CPPFLAGS)
"
8d) change path value of variable "TENSOR_CONFIG" in Makefile.config:
TENSOR_CONFIG := $(TENSORLIB_DIR)/bin/tensor-config 
8e) add "-Wno-c++11-narrowing" in "CXXFLAGS" at end of file
8f) change following lines in Makefile (i.e. add $(ExtraCppFlags)) as follows:
8fA) $(LibObj):%.o:%.cpp 
	$(CXX) -c $(StdCppFlags) $(ExtraCppFlags) -I$(IncludeDir) $(CXXFLAGS) -o $@ $^
8fB) $(TestObj):%.o:%.cpp 
	$(CXX) -c $(TestCppFlags) $(ExtraCppFlags) -I$(IncludeDir)  -I$(TestDir) $(CXXFLAGS) -o $@ $^
8fC) $(TestLibObj):%.o:%.cpp
	$(CXX) -c $(TestCppFlags) $(ExtraCppFlags) -I$(IncludeDir) -I $(TestDir) $(CXXFLAGS) -o $@ $^
8g) run: "make"

- Demo and test
---------------
13) check demo (e.g. Demos/HarmonicOscillator/Gaussian_1D/1/):
13a) change Makefile variables like: 
CXX := /share/apps/gcc-4.9.1/bin/g++
LD := /share/apps/gcc-4.9.1/bin/g++
13b) append $(ExtraCppFlags) to line "CPPFLAGS:= ..." as follows:
CPPFLAGS := -I$(WAVEPACKET_TOPDIR)/$(IncludeDir) $(ODEINT_CPPFLAGS) $(TENSOR_CPPFLAGS) $(ExtraCppFlags)
13c) run: make

Wavepacket-Interface
====================
- Installation
--------------
1) create copy of sources/Makefile.template and name it sources/Makefile (!!Attention: do not add to git index!!)
2) in /sources/Makefile of "wavepacket interface" root change "WAVEPACKET_TOPDIR := <wavepacket root directory>"
3) download and install libconfig
4) in /sources/Makefile (linker/ld flag) of "wavepacket interface" root add "-L/this-path-to-lib/this-libname.so file>", "-Wl,-R<path to <libname>.so file>" and "-lconfig++"
5) download and install lapack:
5a) copy make.inc.example or /INSTALL/mace.inc.gfortran to make.inc
5b) run: make 
6) download and install armadillo:
6a) run: cmake 
6b) run: make install DESTDIR=<chosen library path>
7) in /sources/Makefile (linker/ld flag) of "wavepacket interface" add "-L/this-path-to-lib/<libname>.so file>", "-Wl,-R<path to <libname>.so file>" and "- armadillo -lapack -lrefblas" 
7a) dependent on the lapack/blas installation the flag -lrefblas might have to be changed into "-lblas" or "-lopenblas" (i.e. for "OpenBLAS"; this is said to be most efficient)
7b) TODO: actually "-lrefblas" should be changed at least to "-lblas" as soon as the system optimised Blas installation is available (supposed to be faster)
8) run: make (in "wavepacket interface" root folder) and follow the wizard

Troubleshooting
===============
- mps-bundle
------------
- interesting note appearing during execution of "Installation step 4)":
"
	Libraries have been installed in:
   /home/kschatz/libs/cpp/include/mps-bundle/lib

If you ever happen to want to link against installed libraries
in a given directory, LIBDIR, you must either use libtool, and
specify the full pathname of the library, or use the `-LLIBDIR'
flag during linking and do at least one of the following:
   - add LIBDIR to the `LD_LIBRARY_PATH' environment variable
     during execution
   - add LIBDIR to the `LD_RUN_PATH' environment variable
     during linking
   - use the `-Wl,-rpath -Wl,LIBDIR' linker flag
   - have your system administrator add LIBDIR to `/etc/ld.so.conf'
"

- lapack: 
---------
1) if errors with "make" try:
1a) run: make clean
1b) run: make blaslib
1c) run: make

- armadillo:
------------
1) if problems with cmake/make install (e.g. lapack/blas not found) try:
1a) edit library paths in /cmake_aux/Modules/ARMA_FindLAPACK.cmake, /cmake_aux/Modules/ARMA_FindBLAS.cmake
1b) it might, but must not, be necessary to do the same in /cmake_aux/Modules/ARMA_FindBLAS.cmake too
1c) rebuild (not to forget to better delete CMakeCache.txt before) 
2) if there are problems with reading symbols in liblapack.a whilst compiling:
2a) recompile lapack with compiler flag "-fPIC" (i.e.: edit make.inc in lapack root folder: append -fPIC to variables FORTRAN, OPTS, NOOPT, LOADER)
2b) then append "-gfortran" to linker/ld flag in /sources/Makefile of "wavepacket interface" root

- Git
-----
some comparably easy options after version merge conflicts (after git pull):
a) "git checkout --ours <sources/file>" for keeping local file version,
b) "git checkout --theirs <sources/file>" for keeping remote file version
-> the following should then finish the resolving of the merging conflicts:
c) git add <"*" or "specific file">
d) git commit -m "note"
e) now it can be pushed: git push