resdata is a package for reading and writing the result files from the reservoir simulators. The file types covered are the restart, init, rft, summary and grid files. Both unified and non-unified and formatted and unformatted files are supported.

resdata is mainly developed on Linux and macOS, in addition there is a portability layer which ensures that most of the functionality is available on Windows. The main functionality is written in C/C++, and should typically be linked in in other compiled programs. resdata was initially developed as part of the Ensemble Reservoir Tool, other applications using resdata are the reservoir simulator flow and Resinsight from the OPM project.

Regardless of how you build resdata, it will depend on the following system-level components.

Note: The Conan package manager is not available for most Linux systems. Conan recommends installing it via pip. If using pipx, simply pipx install conan and it'll be availabe for your user regardless if you're using a virtualenv or not.

For small interactive scripts, such as forward models, the recommended way to use resdata is by installing it from PyPI. This method doesn't require setting PYTHONPATH or LD_LIBRARY_PATH environment variables:

$ pip install resdata

This is for when you need to link directly with the resdata C library, but don't need the Python bindings. resdata requires a conforming C++11 or later compiler such as GNU GCC, the CMake build system and, optionally, zlib.

$ git clone https://github.com/Equinor/resdata
$ mkdir resdata/build
$ cd resdata/build
$ cmake ..
$ make
$ make install

To install resdata in a non-standard location, add -DCMAKE_INSTALL_PREFIX=/path/to/install to the first cmake command. Remember to set LD_LIBRARY_PATH=/path/to/install/lib64:$LD_LIBRARY_PATH if you do use a non-standard location for your program to find resdata.so.

If you intend to develop and change resdata you should build the tests by passing -DBUILD_TESTS=ON and run the tests with ctest.

It is also possible to install both the C library and Python bindings using CMake. Note that this alternative is incompatible with resdata installed from PyPI (Alternative 1). As before, we require a conforming C++11 or later compiler, CMake and, optionally, zlib.

$ git clone https://github.com/Equinor/resdata
$ mkdir resdata/build
$ cd resdata/build
$ pip install -r ../requirements.txt
$ cmake .. -DENABLE_PYTHON=ON
$ make
$ make install

You will most likely want to install resdata into a Python virtual environment. First activate the virtualenv, then add the argument -DCMAKE_INSTALL_PREFIX=$(python -c "import sys; print(sys.prefix)") to the cmake command when building.

Then, you must tell Python where to find the package[1]:

$ export PYTHONPATH=/path/to/install/lib/python2.7/site-packages:$PYTHONPATH
$ export LD_LIBRARY_PATH=/path/to/install/lib64:$LD_LIBRARY_PATH

Then you can fire up your Python interpreter and try it out:

>>> from resdata.summary import Summary
>>> import sys

>>> summary = Summary(sys.argv[1])
>>> fopt = summary.numpy_vector("FOPT")

The installation with Python enabled is described in a YouTube video by Carl Fredrik Berg.

[1]: The exact paths here will depend on your system and Python version. The example given is for a RedHat system with Python version 2.7.