<div dir="ltr">Hi!<div><br></div><div>Last week I was working on modifying autoconf files and continue polishing the psurface_convert code. </div><div><br></div><div>I have movded the hdf5.m4 file under commons directory to the <a href="http://configure.ac">configure.ac</a> file to make up a test for linking library hdf5. </div>
<div><br></div><div>On Thurday's chat with Oliver, he give me several suggestions on the output type. </div><div>I have follow his ideas and made some changes. (I will push my changes to git tomorrow, since I have problem linking wireless internet with my computer. I do not have wired internet access tonight.)</div>
<div><br></div><div>1. The hdf5 file generated from amiramesh file are twice as large as the original file. Oliver suggest that moving out the redundant data in hdf5. So I add a base hdf5 writer that store as much data as amiraio does. You can choose which hdf5 output type psurface_convert write using commend line options.</div>
<div> </div><div>2. Write the vtu writer that did not write planar graph into vtu. The output vtu file only display base grid triangle. You can also control it using line options. </div><div><br></div><div>3. I modified the hdf5 writer so that the output h5 file have similar structure with amiramesh. I used to store x coordinate, y coordinate and z coordinate in 3 different arrays. Now I use on array to store them. The topology information, ie, the triangle index and the edge end are store in one hdf5 entry. Now they are separated. I also changed the xdmf file, so the basegrid and planer graph are put into two different grids in xdmf. You can now view the whole graph or view base grid triangle only using buttons in paraview. </div>
<div><br></div><div>Best,</div><div>Xiaoxue</div></div>