A word on the Python API

The Python API is highly similar to the C++ one. The main modules to import are:

• MEDCoupling
• MEDLoader (with mainly the static MEDLoader class within this module)

The following intuitive rules have been used to map the C++ objects to the Python ones:

• std::vector become standard Python lists, see for example Set part of values of DataArrayInt where the function MEDCoupling::DataArray::setInfoOnComponents() is used.
• the return values of C++ functions (usually passed as last arguments in the C++ prototype) are returned directly as a tuple, see for example the Python usage of MEDCoupling::MEDCouplingUMesh::getReverseNodalConnectivity() in Getting the reverse nodal connectivity. The initial prototype where the return values are passed as argument can however still be used, provided you instantiate the returned objects first.
• the indexing mechanism is greatly simplified in Python, and offers similar functionalities to what NumPy provides thanks to the ':' (column) operator. Once can for example refer to part of an array with something like

import MEDCoupling as mc
d = mc.DataArrayInt([(1,2), (3,4), (5,6)], 3, 2)   # an array of 3 tuples with 2 components
print d[:,1]                                       # show only second component of the array: 2,4,6

Finally for the most common objects (DataArray, MEDCouplingUMesh, etc ...) one can direclty access the Doxygen documentation in the interactive Python interpreter using the built-in help function from Python:

import MEDCoupling as mc
help(mc.DataArrayDouble.getNumberOfTuples)