Possible approach for arm64 Mac users with GCAM Ver.7 + Conda

[MarcoLeeHT]

It is fascinating that GCAM7 provides pre-built release for Mac OS with arm64 system.
However, I faced several problems while connecting GCAM7 and latest GCAM-Wrapper.
(such as : lack of boost.python library, etc.)

I could solve this through following steps, and I would like to share with others who are having same problems
Even though followings are a bit rough way to solve the issue,
I hope this can help with people who wants to use GCAM / GCAM-wrapper and Conda

<<< Build GCAM, RMK : don't use brew for java and boost>>>

1. Download GCAM7 source code from github (not pre-built version) and unzip
2. install tbb : "brew install tbb"
3. install java : a. download zip file from openjdk
   b. unzip
   c. move unzipped folder under "/Library/Java/JavaVirtualMachines/"
4. install boost : a. download boost source code zip file from the boost website
   b. unzip the boost source code file under "boost-lib" (anywhere you would like)
5. Build boost : "cd boost-lib"
   "cd unzipped_boost_folder"
   "./bootstrap.sh"
   "./b2"
6. put eigen, jars, hector as instructed in GCAM Documentations "How to Setup and Build"
7. "vim ~/.zshrc"
   export BOOST_INCLUDE=your_boost_directory
   export JARS_LIB=your_gcam_directory/libs/jars/*
   export JAVA_INCLUDE=/Library/Java/JavaVirtualMachines/jdk-...jdk/Contents/Home/include
   export JAVA_LIB=/Library/Java/JavaVirtualMachines/jdk-...jdk/Contents/Home/lib/server
   export EIGEN_INCLUDE=your_gcam_directory/libs/eigen
   export TBB_INCLUDE=/opt/homebrew/Cellar/tbb/202*../include
   export TBB_LIB=/opt/homebrew/Cellar/tbb/202*../lib
   "source ~/.zshrc"
8. "cd your_gcam_directory/cvs/objects/build/linux"
   "vim ./config.system"
   export CXX=g++ -fPIC >>>. you can put it in the first line
   "make gcam -j 8"

<<< Build GCAM-Wrapper on Conda environment>>>

1. "conda create -n test_39 python==3.9.0"
   "conda activate test_39"
   "conda install numpy pandas pyyaml" >>>. These three packages should be installed with conda at the first place
2. unzip the boost source code file under "boost-lib" (the name of unzipped folder for this step should be different from before)
   "conda activate test_39" >>>. Just to make sure this boost is build with the conda environment
   "cd boost-lib"
   "cd unzipped_boost_folder"
   "./bootstrap.sh"
   "./b2" >>>. I am not sure about the reason, but this "./b2" is needed to avoid error
   "./b2 --with-python address-model=64 stage "
3. "vim ~/.zshrc"
   • Mask all the exports from previous step (step 7.)
     export GCAM_INCLUDE=your_gcam_directory/cvs/objects
     export GCAM_LIB=your_gcam_directory/cvs/objects/build/linux
     export BOOST_INCLUDE=your_NEW!!!_boost_directory
     export BOOST_LIB=your_NEW!!!_boost_directory/stage/lib
     export EIGEN_INCLUDE=your_gcam_directory/libs/eigen
     export TBB_INCLUDE=/opt/homebrew/Cellar/tbb/202*../include
     export TBB_LIB=/opt/homebrew/Cellar/tbb/202*../lib
     export CXX='c++ -std=c++14'
     "source ~/.zshrc"
4. "conda activate test_39" >>>. Just to make sure this boost is build with the conda environment
5. a. download gcam-wrapper latest source code (for ver.6) from the github
   b. unzip
   c. "cd dir_to_gcam_wrapper"
   d. "vim ./setup.py"
   * Mask the line with "python_requires='>=3.6.*, <4' " >>>. it is a source of error, and it seems
   you just need to use python version between 3.6 and 4
   e. "pip3 install."