Convert a D3plot to HDF5

Hello everyone,

this article shows the newest and coolest feature of our qd python library in order to make the real data in a D3plot accessible. In version 0.6.6, a new class called RawD3plot has been introduced, which gives access to the raw data arrays of a d3plot. This was crucial to us, because now one can not only check the raw data very easily (yes we found bugs in commercial post-processors), but also transform it into new data formats, such as HDF5. For those who don’t know it, a HDF5 file is like a filesystem, in which one can save large data arrays. The library also provides methods for parallel access from multiple threads and thus is very attractive for High Performance Computing (HPC). With the following code, one can easily convert a d3plot into an HDF5 file:

from qd.cae.dyna import RawD3plot

RawD3plot("path/to/d3plot").save_hdf5("path/to/d3plot.h5")

With this solution, we want to give engineers more power, so that they can build apps for their own needs and do not rely on expensive commercial software. In consequence, a reasonable and easily accessible data format was crucial for us. By the way, one can also read data from the HDF5 file in python with the h5py package. Native interface libraries for C++, C, Fortran etc. are also available, but I assure you this is no fun at all.

And the best thing of course is, that we provide advanced solutions like this for free 🙂

Please don’t forget to comment, mail or star us on github, if you like our work. We are also open for new projects and welcome support for existing ones, such as our AI research.

Edit:

For version 0.6.7, which is not released yet, also compression was tested and enabled. The compression ratio of the raw data achieved a factor of 2 to 4 for our test suite. The compressed HDF5 in comparison to FEMZIP was still 2 times bigger for big files and 3 times bigger for small files, which is quite fine for a free solution.


It has been a while ever since I posted and made some videos, but I’m still working on my PhD and enjoy at the same time the slow internet at the countryside of china, so be patient.

Be the first to comment

Leave a Reply

Your email address will not be published.


*