Skip to content

Latest commit

 

History

History
340 lines (291 loc) · 16.3 KB

README.md

File metadata and controls

340 lines (291 loc) · 16.3 KB
Raw

PyFerret

The PyFerret program and Python module from NOAA/PMEL.
See https://ferret.pmel.noaa.gov/Ferret/ for more information about Ferret and PyFerret.

This repository is regularly synchronized with PyFerret repository at PMEL (the pyferret branch of the ferret project in the subversion repository at PMEL) using git-svn.

Legal Disclaimer

This repository is a software product and is not official communication of the National Oceanic and Atmospheric Administration (NOAA), or the United States Department of Commerce (DOC). All NOAA GitHub project code is provided on an 'as is' basis and the user assumes responsibility for its use. Any claims against the DOC or DOC bureaus stemming from the use of this GitHub project will be governed by all applicable Federal law. Any reference to specific commercial products, processes, or services by service mark, trademark, manufacturer, or otherwise, does not constitute or imply their endorsement, recommendation, or favoring by the DOC. The DOC seal and logo, or the seal and logo of a DOC bureau, shall not be used in any manner to imply endorsement of any commercial product or activity by the DOC or the United States Government.

Ferret/PyFerret Documentation

For more information on using PyFerret, see the Ferret and PyFerret documentation under https://ferret.pmel.noaa.gov/Ferret/

Information about the Ferret email users group, and archives of past discussions from the group (which should be searched prior to sending a question to the email users group) can be found at https://ferret.pmel.noaa.gov/Ferret/email-users-group/

Jupyter / iPython notebook

The latest ferretmagic module from Patrick Brockmann for using PyFerret with the iPython notebook can be obtained using pip install ferretmagic, or see https://pypi.org/project/ferretmagic/. Note that this only installs the ferretmagic module for PyFerret; it does not install PyFerret.

Anaconda Installation - Linux, OS X, and Windows 10/bash

Note: If you are updating an existing conda installation of PyFerret, you should only need to update the existing installation using conda update -n FERRET pyferret (assuming FERRET is the name of the environment used). Or you can install under a new environment name if you wish to keep multiple versions of PyFerret.
You may first need to update conda itself conda update conda; conda update --all before updating PyFerret. If you are using a Python 2.x version of miniconda, you will need to install the Python 3.x version of miniconda and then install PyFerret in that version of miniconda in order to get the latest version of PyFerret.

Download and install minicoda for your system from https://docs.conda.io/en/latest/miniconda.html Note that Windows 10 bash must use the Linux version! Install the Python 3.x version of minicoda to get the latest versions of PyFerret. (Although PyFerret works with ether Python 2.x or Python 3.x, Python 2.x itself will no longer be supported after 2019.) Allow miniconda to add its initialization code to your start-up scripts (e.g., $HOME/.bashrc) and open a new login window when the installation is complete. One thing with will do is add the conda bin directory to your path of directories to search for executables.

Execute the following command on the terminal to install pyferret as well as ferret_datasets (the default Ferret/PyFerret datasets) into conda:

conda create -n FERRET -c conda-forge pyferret ferret_datasets --yes

(FERRET is the environment name where pyferret is installed. You can change that to any name you like, such as PyFerret.)

A development version of pyferret is also available to install through Anaconda. This version is meant to allow users to test new features and bug fixes ahead of the next stable release. Execute the following command in place of previous one to install the development version alongside the default datasets into conda:

 conda create -n FERRET_DEV install -c conda-forge/label/pyferret_dev pyferret ferret_datasets --yes

(It is recommended to create two enironments, one for the test version of pyferret and one for the stable version. Hence, FERRET and FERRET_DEV)

Apple Silicon Computers (M1, M2, etc.)

To install PyFerret on an Apple Silicon computer, you may either run the above command from a x86_64 version of Anaconda/miniconda installed via a Rosetta terminal or install the usual arm64 version of Anaconda/miniconda and run the following:

CONDA_SUBDIR=osx-64 conda create -n FERRET -c conda-forge pyferret ferret_datasets --yes

Running PyFerret installed via conda

To start using PyFerret, execute the following command:

conda activate FERRET

(replacing FERRET with whatever environment name you used.)

Once you are done working with PyFerret you can leave this environment, if you wish, with the command:

conda deactivate

As mention above FERRET is the environment name where pyferret is installed, and can be any name you like (such as PyFerret). We do not recommend installing pyferret in the root environment of miniconda. The main reason is to take advantage of the activate and deactivate scripts that will set up all the variables that PyFerret needs. (You can test whether the PyFerret environment is activated by issuing the command echo $FER_DATA and see if it returns a directory name.)

If you usually work under a C-shell (such as tcsh) or outside the conda environment, you may want to create a Bourne shell script that activates and runs PyFerret within the Bourne shell that is created for running the script. Such a script would look something like the following (replacing FERRET with whatever environment name you used):

#! /bin/sh -l
## initialize the environment for running PyFerret
conda activate FERRET
## You may wish to add directories to FER_DATA and FER_GO
# export FER_DATA="$FER_DATA /my/path/to/big/data/dir"
# export FER_GO="$FER_GO /my/path/to/custom/ferret/scripts"
## Now execute PyFerret with all the command-line arguments given to this script
pyferret $*

If this script was called pyferret.sh, made executable (chmod 755 pyferret.sh), and placed in a directory listed in your $PATH enviroment variable, you can just type pyferret.sh (followed by any desired arguments) from the command line to run PyFerret.

Installation from prebuilt tar.gz file

You will need to have the following packages installed using your software manager application, or using a command-line package installation program such as yum or apt-get (which needs to be run as the root user or using the sudo privilege escalation program.)

Required packages that may not already be installed:

  • numpy or python-numpy (NumPy)
  • libgfortran (Fortran library; if you install SciPy, it will be installed)
  • pyside2, pyside, pyqt, python-qt5, python-qt4, PyQt5, or PyQt4 (Python bindings for Qt; may already be installed)

Highly recommended but optional packages:

  • scipy or python-scipy (SciPy)
  • pyshp or python-pyshp (PyShp for shapefile functions)

You may also wish to install the netcdf and nco packages to provide some useful programs for working with NetCDF files (such as ncdump and ncattted which are used in the benchmark tests).

If you do not have the Ferret/PyFerret standard datasets, they can be obtained from the https://github.com/NOAA-PMEL/FerretDatasets/ GitHub repo. The contents can be put extracted/cloned to whatever location desired.

Extract the PyFerret tar.gz file in the desired location. Starting with PyFerret v7, there is only one tar.gz file which extracts all its contents to a subdirectory that it creates (as apposed to Ferret which has separate fer_environment and fer_executables tar.gz files that extract into the current directory). If desired, at this time you can change the name of this subdirectory that was created.

Move into this PyFerret installation directory and run the bin/Finstall script to create the ferret_paths.sh, ferret_paths.csh, and pyferret scripts. The value of FER_DIR, the Ferret/PyFerret installation directory, should be this installation directory, which can be specified as . (a period) which means the current directory. (If FER_DIR is already defined for another Ferret/PyFerret installation, you will need to tell the script to use a new value.) For FER_DSETS, the Ferret/PyFerret standard datasets, specify the directory containing these datasets (which you may have created from the FerretDatasets github site mentioned above).

To run PyFerret, you first need to set the PyFerret environment variables. This can be done by executing either . ferret_paths.sh (for Bourne-type shells; e.g., bash) or source ferret_paths.csh (for C-type shells; e.g. tcsh).

  • Note: the pyferret script has recently been updated to automatically set the Ferret environment variables, if not already defined, using the appropriate ferret_paths script.

Building PyFerret from source

Please note that these are general instructions that are not fully verified; names of installation packages may vary slightly for you particular operating system. In particular, some systems have special development (-dev) packages that provide the include files, and shared-object libraries without a numeric extension, that are needed for compiling and linking the PyFerret code. On other systems, these include and library files are part of the standard package. These instructions assume your package manager provides recent versions of HDF5 and NetCDF.

If you have previously built PyFerret from source

The definitions of CC, FC, and LD (the last only for building external functions written in Fortran) were moved to the site_specific.mk.in files. (These were previously defined in the platform_specific.mk.* files.) If you already have customized site_specific.mk files, please appropriately updated your customized files with these additional definitions.

Packages from the package manager

If not already installed on your system, install the following packages using the package manager for your operating system, or a command-line package installation program such as yum or apt-get (which needs to be run as a system administator - as "root" - or using the sudo privilege escalation program):

  • gfortran, or gcc (Gnu Compiler Collection) on some systems - for the gfortran compiler and library
  • libcairo, libcairo-dev, cairo, or cairo-dev - for the cairographics library and include files
  • libpango, libpango-dev, pango, or pango-dev - for the pango and pango-cairo library and include files
  • numpy, or python-numpy - the NumPy python package as well as include and library files
  • pyqt, python-qt5, python-qt4, PyQt5, or PyQt4 - for either PyQt5 or PyQt4
  • netcdf - for NetCDF 4.x include and library files

The NetCDF package should add the HDF5 packages as a dependency. Some package manager programs (such as Homebrew) have their own version of Python separate from the operating system; if so, the NumPy and PyQt packages should add the python package(s) as dependencies.

You may also want, if not already installed:

  • git - to use "git" commands to download the source code; highly recommended
  • scipy, or python-scipy - for statisticaly functions in PyFerret; highly recommended
  • pyshp, or python-pyshp - for shapefile functions in PyFerret

Note that pyshp is pure-python code and can also be installed using pip2 (part of python2.x) or pip3 (part of python3.x).

PyFerret source code

The green Clone or download button at the top of the PyFerret GitHub Code page/tab gives you options for obtaining the latest PyFerret source code. You can get a copy of the latest source as a zip file, but a better option, if you can, is to use "git" commands to clone the repository (the source code, plus history and version control of the source code) to your local system.

The git comands to clone the PyFerret repository look something like the following (the local copy of the repository will be put into $HOME/git/PyFerret):

git clone https://github.com/NOAA-PMEL/PyFerret.git $HOME/git/PyFerret
cd $HOME/git/PyFerret

Working with a cloned repository allows you to quickly and easily update and rebuild PyFerret when updates appear that you wish to use. Executing the command:

git pull

when in the PyFerret cloned repository will download any changes to your local copy.

Configure and build

In the PyFerret source directory, copy the site_specific.mk.in configuration template file to site_specific.mk and edit this site_specific.mk configuration file appropriately for your system; for example:

DIR_PREFIX = $(HOME)/git/PyFerret
INSTALL_FER_DIR = /usr/local/PyFerret
BUILDTYPE = x86_64-linux
CC = $(shell which gcc)
FC = $(shell which gfortran)
PYTHON_EXE = python2.7
GFORTRAN_LIB = $(shell $(FC) --print-file-name=libgfortran.a)
CAIRO_LIBDIR =
PIXMAN_LIBDIR =
PANGO_DIR =
GLIB2_LIBDIR =
HDF5_LIBDIR =
SZ_LIBDIR =
NETCDF_LIBDIR = /usr/lib64

Information about each of these values, as well as suggested values to assign, are included as comments (lines starting with a #) in the site_specific.mk file.

Similarly, copy external_functions/ef_utility/site_specific.mk.in to external_functions/ef_utility/site_specific.mk and edit external_functions/ef_utility/site_specific.mk:

BUILDTYPE = x86_64-linux
CC = $(shell which gcc)
FC = $(shell which gfortran)
LD = $(shell which gfortran)
PYTHON_EXE = python2.7

If you have previously built PyFerret (successfully or not) from this source directory or repository, run the command:

make clean

to make sure you have removed all previously generated files. Then run the command:

make

to build PyFerret. This build will take a bit of time (minutes) and will generate a lot of output, so you may wish to redirect output and run this command in the background.

When the build has successfully completed, install PyFerret in the location given by the value of INSTALL_FER_DIR in the site_specific.mk file by running the following command. (If the installation directory exists and is not empty, you should move or remove any contents of that directory to ensure a clean installation.)

make install

You may need to be logged in as a system administrator (as "root") or use the "sudo" privilege escalation command (thus, sudo make install), to install PyFerret system-wide (such as installing in /usr/local/PyFerret as in the example site_specific.mk file given above.

Standard Ferret/PyFerret datasets

If you do not have the standard Ferret/PyFerret datasets, they can be downloaded from https://github.com/NOAA-PMEL/FerretDatasets/ either as a zip file download or as a git cloned repository (similar to obtaining the PyFerret source). If you already have a copy of these datasets on your system, these datasets can be shared between Ferret and PyFerret, including different versions of these programs. You can also add any of your own datasets that might be frequently used. These datasets will be needed as part of the following PyFerret configuration.

(Py)Ferret configuration

Change to the PyFerret installation directory (the value of INSTALL_FER_DIR) created above and run the script:

bin/Finstall

to create the ferret_paths.sh, ferret_paths.csh, and pyferret scripts. The value of FER_DIR, the Ferret/PyFerret installation directory, should be this installation directory, which can be specified as . (a period) which means the current directory. The value of FER_DSETS should be the directory containing the standard Ferret/PyFerret dataset mentioned above.

Before running PyFerret for the first time in a new terminal window (shell), you will need to set the Ferret/PyFerret environment variables using the appropriate ferret_paths script:

. /my/path/to/ferret_paths.sh

(a period, a space, and the path to the ferret_paths.sh script) for Bourne-type shells (such as bash), or

source /my/path/to/ferret_paths.csh

for C-type shells (such as tcsh). (The source command is also supported by the bash shell.)