Make VTK optional

demos/benney_luke/benney_luke.py.rst

@@ -88,6 +88,7 @@ Finally, before implementing the problem in Firedrake, we calculate the total en
 The implementation of this problem in Firedrake requires solving two nonlinear variational problems and one linear problem. The Benney-Luke equations are solved in a rectangular domain :math:`\Omega=[0,10]\times[0,1]`, with :math:`\mu=\epsilon=0.01`, time step :math:`dt=0.005` and up to the final time :math:`T=2.0`. Additionally, the domain is split into 50 cells in the x-direction using a quadrilateral mesh. In the y-direction only 1 cell is enough since there are no variations in y::
 
   from firedrake import *
+  from firedrake.output import VTKFile
 
 Now we move on to defining parameters::
 
@@ -190,7 +191,7 @@ For visualisation, we save the computed and exact solutions to
 an output file.  Note that the visualised data will be interpolated
 from piecewise quadratic functions to piecewise linears::
 
-  output = File('output.pvd')
+  output = VTKFile('output.pvd')
   output.write(phi0, eta0, ex_phi, ex_eta, time=t)
 
 We are now ready to enter the main time iteration loop::

demos/burgers/burgers.py.rst

@@ -19,7 +19,7 @@ the viscosity term by parts:
 
 .. math::
 
-   \int_\Omega\frac{\partial u}{\partial t}\cdot v + 
+   \int_\Omega\frac{\partial u}{\partial t}\cdot v +
    ((u\cdot\nabla) u)\cdot v + \nu\nabla u\cdot\nabla v \ \mathrm d x = 0.
 
 The boundary condition has been used to discard the surface
@@ -28,12 +28,13 @@ stability we elect to use a backward Euler discretisation:
 
 .. math::
 
-   \int_\Omega\frac{u^{n+1}-u^n}{dt}\cdot v + 
+   \int_\Omega\frac{u^{n+1}-u^n}{dt}\cdot v +
    ((u^{n+1}\cdot\nabla) u^{n+1})\cdot v + \nu\nabla u^{n+1}\cdot\nabla v \ \mathrm d x = 0.
 
 We can now proceed to set up the problem. We choose a resolution and set up a square mesh::
 
   from firedrake import *
+  from firedrake.output import VTKFile
   n = 30
   mesh = UnitSquareMesh(n, n)
 
@@ -76,7 +77,7 @@ system's evolution::
   timestep = 1.0/n
 
 Here we finally get to define the residual of the equation. In the advection
-term we need to contract the test function :math:`v` with 
+term we need to contract the test function :math:`v` with
 :math:`(u\cdot\nabla)u`, which is the derivative of the velocity in the
 direction :math:`u`. This directional derivative can be written as
 ``dot(u,nabla_grad(u))`` since ``nabla_grad(u)[i,j]``:math:`=\partial_i u_j`.
@@ -89,15 +90,15 @@ is differentiated by the nonlinear solver::
 
 We now create an object for output visualisation::
 
-  outfile = File("burgers.pvd")
+  outfile = VTKFile("burgers.pvd")
 
 Output only supports visualisation of linear fields (either P1, or
 P1DG).  In this example we project to a linear space by hand.  Another
-option is to let the :class:`~.File` object manage the decimation.  It
+option is to let the :class:`~.VTKFile` object manage the decimation.  It
 supports both interpolation to linears (the default) or projection (by
-passing ``project_output=True`` when creating the :class:`~.File`).
-Outputting data is carried out using the :meth:`~.File.write` method
-of :class:`~.File` objects::
+passing ``project_output=True`` when creating the :class:`~.VTKFile`).
+Outputting data is carried out using the :meth:`~.VTKFile.write` method
+of :class:`~.VTKFile` objects::
 
   outfile.write(project(u, V_out, name="Velocity"))
 
@@ -112,5 +113,5 @@ which amount to applying a full LU decomposition as a preconditioner. ::
       u_.assign(u)
       t += timestep
       outfile.write(project(u, V_out, name="Velocity"))
-    
+
 A python script version of this demo can be found :demo:`here <burgers.py>`.

demos/camassa-holm/camassaholm.py.rst

@@ -62,6 +62,7 @@ As usual, to implement this problem, we start by importing the
 Firedrake namespace. ::
 
   from firedrake import *
+  from firedrake.output import VTKFile
 
 To visualise the output, we also need to import matplotlib.pyplot to display
 the visual output ::
@@ -168,11 +169,11 @@ e.g. for output. ::
   m0, u0 = w0.subfunctions
   m1, u1 = w1.subfunctions
 
-We choose a final time, and initialise a :class:`~.File` object for
+We choose a final time, and initialise a :class:`~.VTKFile` object for
 storing ``u``. as well as an array for storing the function to be visualised::
 
   T = 100.0
-  ufile = File('u.pvd')
+  ufile = VTKFile('u.pvd')
   t = 0.0
   ufile.write(u1, time=t)
   all_us = []

demos/extruded_shallow_water/test_extrusion_lsw.py

@@ -3,6 +3,7 @@
 a sin(x)*sin(y) solution that doesn't care about the silly BCs"""
 
 from firedrake import *
+from firedrake.output import VTKFile
 
 
 power = 5
@@ -37,7 +38,7 @@
 t = 0
 dt = 0.0025
 
-file = File("lsw3d.pvd")
+file = VTKFile("lsw3d.pvd")
 p_trial = TrialFunction(Xplot)
 p_test = TestFunction(Xplot)
 solve(p_trial * p_test * dx == p_0 * p_test * dx, p_plot)

demos/helmholtz/helmholtz.py.rst

@@ -49,6 +49,7 @@ method, so lets go ahead and produce a numerical solution.
 First, we always need a mesh. Let's have a :math:`10\times10` element unit square::
 
   from firedrake import *
+  from firedrake.output import VTKFile
   mesh = UnitSquareMesh(10, 10)
 
 We need to decide on the function space in which we'd like to solve the
@@ -93,7 +94,7 @@ of the manual on :doc:`solving PDEs <../solving-interface>`.
 Next, we might want to look at the result, so we output our solution
 to a file::
 
-  File("helmholtz.pvd").write(u)
+  VTKFile("helmholtz.pvd").write(u)
 
 This file can be visualised using `paraview <http://www.paraview.org/>`__.
 

demos/higher_order_mass_lumping/higher_order_mass_lumping.py.rst

@@ -49,6 +49,7 @@ In the work of :cite:`Chin:1999` and later :cite:`Geevers:2018`, several triangu
 In addition to importing firedrake as usual, we will need to construct the correct quadrature rules for the mass-lumping by hand. FInAT is responsible for providing these quadrature rules, so we import it here too.::
 
     from firedrake import *
+    from firedrake.output import VTKFile
     import finat
 
     import math
@@ -73,7 +74,7 @@ We choose a degree 2 `KMV` continuous function space, set it up and then create
 
 We create an output file to hold the simulation results::
 
-    outfile = File("out.pvd")
+    outfile = VTKFile("out.pvd")
 
 Now we set the time-stepping variables performing a simulation for 1 second with a timestep of 0.001 seconds::
 

demos/linear-wave-equation/linear_wave_equation.py.rst

@@ -50,6 +50,7 @@ This time we created the mesh with `Gmsh <http://gmsh.info/>`_:
 We can then start our Python script and load this mesh::
 
   from firedrake import *
+  from firedrake.output import VTKFile
   mesh = Mesh("wave_tank.msh")
 
 We choose a degree 1 continuous function space, and set up the
@@ -66,7 +67,7 @@ output file::
 
 Output the initial conditions::
 
-  outfile = File("out.pvd")
+  outfile = VTKFile("out.pvd")
   outfile.write(phi)
 
 We next establish a boundary condition object. Since we have time-dependent
@@ -128,7 +129,7 @@ Step forward :math:`\phi` by the second half timestep::
       phi -= dt / 2 * p
 
 Advance time and output as appropriate, note how we pass the current
-timestep value into the :meth:`~.File.write` method, so that when
+timestep value into the :meth:`~.VTKFile.write` method, so that when
 visualising the results Paraview will use it::
 
       t += dt

demos/linear_fluid_structure_interaction/linear_fluid_structure_interaction.py.rst

@@ -64,6 +64,7 @@ The underlined terms are the coupling terms. Note that the first equation for :m
 Now we present the code used to solve the system of equations above. We start with appropriate imports::
 
     from firedrake import *
+    from firedrake.output import VTKFile
     import math
     import numpy as np
 
@@ -278,7 +279,7 @@ Let us set the initial condition. We choose no motion at the beginning in both f
 
 A file to store data for visualization::
 
-    outfile_phi = File("results_pvd/phi.pvd")
+    outfile_phi = VTKFile("results_pvd/phi.pvd")
 
 To save data for visualization, we change the position of the nodes in the mesh, so that they represent the computed dynamic position of the free surface and the structure::
 

demos/ma-demo/ma-demo.py.rst

@@ -75,6 +75,7 @@ We now proceed to set up the problem in Firedrake using a square
 mesh of quadrilaterals. ::
 
   from firedrake import *
+  from firedrake.output import VTKFile
   n = 100
   mesh = UnitSquareMesh(n, n, quadrilateral=True)
 
@@ -188,7 +189,7 @@ and output the solution to a file. ::
 
   u, sigma = w.subfunctions
   u_solv.solve()
-  File("u.pvd").write(u)
+  VTKFile("u.pvd").write(u)
 
 An image of the solution is shown below.
 

demos/matrix_free/navier_stokes.py.rst

@@ -5,6 +5,7 @@ We solve the Navier-Stokes equations using Taylor-Hood elements.  The
 example is that of a lid-driven cavity. ::
 
   from firedrake import *
+  from firedrake.output import VTKFile
 
   N = 64
 
@@ -134,7 +135,7 @@ And finally we write the results to a file for visualisation. ::
   u.rename("Velocity")
   p.rename("Pressure")
 
-  File("cavity.pvd").write(u, p)
+  VTKFile("cavity.pvd").write(u, p)
 
 A runnable python script implementing this demo file is available
 :demo:`here <navier_stokes.py>`.

demos/matrix_free/rayleigh-benard.py.rst

@@ -11,6 +11,7 @@ the Navier-Stokes part, and piecewise linear elements for the
 temperature. ::
 
   from firedrake import *
+  from firedrake.output import VTKFile
 
   N = 128
 
@@ -241,7 +242,7 @@ Finally, we'll output the results for visualisation. ::
   p.rename("Pressure")
   T.rename("Temperature")
 
-  File("benard.pvd").write(u, p, T)
+  VTKFile("benard.pvd").write(u, p, T)
 
 A runnable python script implementing this demo file is available
 :demo:`here <rayleigh-benard.py>`.

demos/matrix_free/stokes.py.rst

@@ -9,6 +9,7 @@ cavity.
 As ever, we import firedrake and define a mesh.::
 
   from firedrake import *
+  from firedrake.output import VTKFile
 
   N = 64
 
@@ -47,7 +48,7 @@ wrapped in a ``try/except`` block so that an error is not raised in
 the case that it is not, to do this we must import ``PETSc``::
 
   from firedrake.petsc import PETSc
-  
+
 To factor the matrix from this mixed system, we must specify
 a ``mat_type`` of ``aij`` to the solve call.::
 
@@ -130,7 +131,7 @@ file.::
   u.rename("Velocity")
   p.rename("Pressure")
 
-  File("stokes.pvd").write(u, p)
+  VTKFile("stokes.pvd").write(u, p)
 
 By default, the mass matrix is assembled in the :class:`~.MassInvPC`
 preconditioner, however, this can be controlled using a ``mat_type``

demos/multigrid/geometric_multigrid.py.rst

@@ -20,6 +20,7 @@ hierarchies are constructed using regular bisection refinement, so we
 must create a coarse mesh. ::
 
   from firedrake import *
+  from firedrake.output import VTKFile
 
   mesh = UnitSquareMesh(8, 8)
 
@@ -114,7 +115,7 @@ appropriate settings using solver parameters. ::
 
   u = run_solve(parameters)
   print('MG F-cycle error', error(u))
-     
+
 A saddle-point system: The Stokes equations
 -------------------------------------------
 
@@ -259,7 +260,7 @@ Finally, we'll write the solution for visualisation with Paraview. ::
   u.rename("Velocity")
   p.rename("Pressure")
 
-  File("stokes.pvd").write(u, p)
+  VTKFile("stokes.pvd").write(u, p)
 
 A runnable python version of this demo can be found :demo:`here
 <geometric_multigrid.py>`.