tech.md

layout	title	description	header-img
page	Tech Notes	思想有多远，我们就可以走多远. <br>(We can go as far as we thought)	img/top.png

1. Softwares List (mainly for Mac) - updated: 2017-09-29

   • Github Desktop: https://desktop.github.com/
   • MarkDown IDE for Mac OS: https://macdown.uranusjr.com/
   • Anaconda: https://www.anaconda.com/download/
   • QGIS: http://www.qgis.org/en/site/forusers/download.html
   • gcc & gfortran: http://hpc.sourceforge.net/
   • NCL/NCAR: http://www.ncl.ucar.edu/Download/
   • CDO: https://code.mpimet.mpg.de/projects/cdo/
   • Panoply (Gridded data viewer): https://www.giss.nasa.gov/tools/panoply/
   • Texmaker (Tex IDE): http://www.xm1math.net/texmaker/
   • MacTex (Latex for Mac OS):http://www.tug.org/mactex/

2. Install packages in Anaconda - updated: 2017-09-29

   1. Create a NEW python environment

      • Pip 9.0.1
      • Python 2.7.13
      • Setuptools 27.2.0
      • Vs2008_runtime 9.00.30729.5054
      • Vs2015_runtime 14.0.25420
      • Wheel 0.29.0

   2. install pkgs using Terminal BY the ORDER of:

      1. conda install -y -c conda-forge basemap
      2. conda install -y pandas scipy simplejson
      3. conda install -y -c conda-forge rasterio
      4. pip install pyshp pycrs shapely netcdf4
      5. conda install -y -c conda-forge gdal=2.0
      6. conda install -c anaconda statsmodels

      OR specify the version

      1. conda install -y -c conda-forge basemap=1.1 pandas=0.20.3 scipy=0.19.1 simplejson=3.11.1 gdal=2.0 rasterio=0.25.0 shapely=1.6.1 netcdf4 pyshp=1.2.12 statsmodels=0.8.0 krb5=1.14.2 pillow pkg-config gsl gcc

      1. pip install pycrs==0.1.3

   3. install my package:

      1. git clone https://[email protected]/wk1984/quick-analysis-and-visualization.git
      2. cd quick-analysis-and-visualization
      3. python setup.py install

   4. test:

      python -c "from QuickPlots import QuickPlots"

      python -c "import QuickPlots;print QuickPlots.version"

   5. Q&A:

      • Library not loaded: @loader_path/./libgssapi_krb5.2.2.dylib

        conda install krb5

3. Install MacPorts - updated: 2017-10-10

   1. Install Xcode: Huge file > 5 GB

      Apple Store

      see details: https://guide.macports.org/#installing.xcode

   2. Install Xcode Command Line Tools:

      xcode-select --install

   3. Agree to Xcode license in Terminal:

      sudo xcodebuild -license

   4. Download and install MacPorts for your version of the Mac operating system:

      • High Sierra 10.13
      • Sierra 10.12
      • El Capitan 10.11

      Just Double-Click pkg file!

4. Install CDO by using MacPorts:

   CDO = Climate Data Operators

   Homepage = https://code.zmaw.de/projects/cdo

   sudo port install cdo

   To test whether it is installed rightly:

   cdo -V

5. Install ILAMB in Anaconda.

   • create a new env for installing ILAMB (recommend)

   • install all requirements:

     conda install -c conda-forge -y numpy matplotlib netcdf4=1.2.4 sympy scipy mpi4py cfunits basemap

   • download ilamb from Bitbucket, then install:

     git clone https://[email protected]/ncollier/ilamb.git

     cd ilamb

     python setup.py install

   • test:

     python -c "import ILAMB; print ILAMB.version"

6. Install NCL:

Notes: It should be consistent generation between gcc and gfortran from http://hpc.sourceforge.net/ and "gcc-x.y.bin.tar.gz" file. e.g, ncl_ncarg-6.4.0-MacOS_10.12_64bit_nodap_gnu530.tar.gz needs gcc-5.1-bin.tar.gz. If you install any other versions of gcc/gfortran, it would notice like:

> dyld: Library not loaded: /usr/local/lib/libgfortran.3.dylib
> Referenced from: /usr/local/ncl-6.4.0/bin/ncl
> Reason: image not found

• Install XQuartz from https://dl.bintray.com/xquartz/downloads/XQuartz-2.7.11.dmg

• For Bash environment: add three lines to ~/.bash, or ~/.bash_profile

  export DISPLAY=:0.0

  export NCARG_ROOT=/usr/local/ncl-6.4.0

  export PATH=$NCARG_ROOT/bin:$PATH

  Save and Quit: press "esc", and input 'wq' following the ":"

  :wq

  After that, make the change active.

  source ~/.bash_profile

• Test whether it works well (before using NCL, you need to launch XQuartz, otherwise, it will not work!):

  ng4ex gsun01n

The example looks like:

-----