This seems to be a more stable approach as it does not produce
```
/usr/bin/python3.9: can't open file
'/home/mblatt/src/dune/opm-2.9/opm-common//install.py': [Errno 2] No such file or directory
```
for rebuilds after failed attempts.
Convert the Python opm package from a regular package to a namespace
package such that opm-simulators and opm-common can contribute to the
package from different filesystem paths. In this way, the two packages
opm.simulators and opm.io (in opm-common) can have a different parent
filesystem path.
Remove SYSTEM argument for pybind11_add_module() if cmake version is
greater than 3.12. It appears that this argument has been deprecated
and also removed by the new FindPython mode in the cmake tools for
pybind11.
Instead of always using site-packages we query the sitepackages path
from Python. If it matches dist-packages then we use dist-packages if
we install below /usr, otherwise site-packages is used.
This will also make sure that the Python files end up in the same
location as for the ones of opm-common. This should make using them
easier.
Expect non-reference type shared pointers arguments instead of references
to shared pointer. This will make it clear to the caller that the called
function is making a copy of the pointer for its own use and not trying
to modify the original pointer of the caller.
Adds a new constructor to Main.hpp that takes shared pointers to Deck,
EclipseState, Schedule, and SummaryConfig. This makes it possible to
share these variables with Python without worrying about lifetime issues
of the underlying C++ objects. For example, a Python script can first
create an opm.io.schedule.Schedule object which is modified from Python.
Then, assume the same Python script creates an
opm.simulators.BlackOilSimulator which is initialized with the same
schedule object. Since the underlying C++ object is a shared pointer,
the Schedule object in Python may go out of scope (get deleted by Python)
without having the C++ schedule object being deleted. And the Python
BlackOilSimulator may continue to be used after the Python Schedule object
has been deleted since it still has a valid C++ schedule object.
due to bugs in the openmpi on bionic, this test fails to
execute properly in pbuilder environments. instead
of rebuilding openmpi without dynamic loading
(which is the suggested fix) and potentially break users
systems, this is a non-intrusive workaround to be used
for packaging.
also add explicit option for python support to make it
visible in cmake frontends.
Restores the original cwd after each unittest in test_basic.py. Also
simplifies add_test() in python/simulators/CMakeLists.txt such that the
Bash script wrapper run-python-tests.sh is no longer needed to run the
tests.
A resubmission of commit 8e4f748 in PR #2403 and PR #2444 and continues
the work in #2690 implementing Python bindings to the flow simulator.
The Python step() method advances the simulator one report step. Before
calling step() for the first time, step_init() must have been called.
A resubmission of commit 11eaa3d7 in PR #2403 and PR #2443 and continues
the work in #2555 implementing Python bindings to the flow simulator.
The step_init() method initializes the simulation. It is required for the
Python script to run step_init() before calling the step() method (which
will be implemented in a later commit).
