xu-xianghua / ofpp Goto Github PK

When reading binary data, I found that sometimes the reader exited early for boundary data. I think this is due to the fact that the } character is sometimes in binary OF data. In the function split_boundary_content this causes a wrong content range.

At line 172 of field_parser.py. I added an ugly fix which checks if the following line is long enough to still contain binary data:

        if in_boundary_field:
            #USER FIX
            if (lc.rstrip() == b'}' and len(content[n+1])<30):
                break
            if in_patch_field:
                if (lc.strip() == b'}' and len(content[n+1])<30):

The keyword in parse_data_nonuniform() for symmetric tensors should be "symmTensor" instead of "symmtensor".

Hi,
Parsing internal non uniform field with linux works fine, but with windows we get this error:

could not convert string to float: b'-3.85361e-19)'

It seems that the character ')' is read inside the string.
Regards
Robert

Hi,
First of all, thank for sharing this package, it's really useful to have this kind of tool to quickly load and manipulate OpenFOAM cases. I noticed a error about the total number of cells. The num_cell attribute return X cells, while there are X +1 cells (as it is written in the note of the constant/owner file)
Best regards,
Jérôme.

Great job !
Many thanks for the modules :)

Do you plan to extend this module to post-processing utilities like extraction of data into an arbitrary plane, plot over line, etc ?
@greole https://github.com/greole/owls has also made an excellent package in this direction but it looks like his repository is not maintained anymore.

Anyway, congratulations for this awesome job and many thanks for sharing it.
Long life to ofpp :)

Regards,
François

Hi,

Firstly, thanks for the library. I discovered it yesterday and it's already been very handy to debug my setup.

In return, I just wanted to point out that the subroutine parse_faces_content() fails for faces with more than 10 points as the line data = [[int(s) for s in ln[2:-2].split()] for ln in content[n + 2:n + 2 + num]] reasonably assumes that faces have less than 10 vertices. Ominously it appears I have just produced a mesh containing a face with 10 vertices. As OpenFOAM supports any shape of cell, perhaps consider changing the line to data = [[int(s) for s in re.findall(b"\d+", ln)[1:]] for ln in content[n + 2:n + 2 + num]]. Of course this requires an import of the re library.