Is there a way to view the number of cells in an .h5m file?

[jbae11]

Is there a convenient way to 'inspect' an .h5m file? For example printing out how many cells it has?

Thank you.

[makeclean]

I recall someone writing an h5m file inspector @ Get Outlook for Android<https://aka.ms/AAb9ysg>

…

________________________________ From: Jin Whan Bae ***@***.***> Sent: Monday, July 4, 2022 7:49:37 PM To: svalinn/DAGMC ***@***.***> Cc: Subscribed ***@***.***> Subject: [svalinn/DAGMC] Is there a way to view the number of cells in an `.h5m` file? (Discussion #811) Is there a convenient way to 'inspect' an .h5m file? For example printing out how many cells it has? Thank you. — Reply to this email directly, view it on GitHub<#811>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AASTUSTPLZL23YIYOFGNGE3VSMW4DANCNFSM52T6XVOQ>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[shimwell]

Here is a link to the dagmc_h5m_file_inspector https://github.com/fusion-energy/dagmc_h5m_file_inspector it was a community effort

[gonuke]

We are also developing a DAGMC stats package to generate summaries of many characteristics of a DAGMC file: https://github.com/svalinn/dagmc_stats

[jbae11]

thanks! That sounds great.

[shimwell]

@jbae11 give this a go for counting the number of volumes

pip install h5py

Then from within python

import h5py
dagmc_file_contents = h5py.File('dagmc.h5m')
groups=dagmc_file_contents['/tstt/sets/children']
print(len(groups))

[pshriwise]

I think we need to be careful here, children in the .h5m file can refer to surfaces, curves, and geometric vertices also -- not just volumes (a.k.a cell as far as the Monte Carlo code is concerned). The above command might be misleading.

Looking at the structure of the .h5m file a little more I think something like this would do the job:

import h5py
dagmc_file_contents = h5py.File('dagmc.h5m')
category_data = dagmc_file_contents['tstt/tags/CATEGORY/values']
categories_as_str = [c.tobytes().decode().replace('\x00', '') for c in category_data]
n_vols = sum([v == 'Volume' for v in categories_as_str])
print(n_vols)

FWIW @shimwell, I'm liking the use of h5py to get this information from the DAGMC file more and more. The risk is of course that the MOAB file format changes at some point, but I don't see that happening anytime soon and we're well connected enough with the developers that I think we'll see it coming.

[shimwell]

Just wondering if this volume code would be useful in the openmc.DAGMCUniverse class 👀
Are the volumes in a DAGMC file always sequential and start at 1? For example volume ids would be 1,2,3,4

[pshriwise]

We could in theory look up the relevant ID information from tag data as well… here’s where accessing that information using the MOAB interface (via PyMOAB) becomes useful. Doing this via the MOAB library ensures we’re correlating data entries in the tag info correctly for a given entity. It would take some manual book keeping to do this with h5py I think. Looking forward to discussing this further in our bi-weekly meeting.

[makeclean]

Volumes id's are arbitrary and gappy

…

________________________________ From: Jonathan Shimwell ***@***.***> Sent: 27 September 2022 14:49 To: svalinn/DAGMC ***@***.***> Cc: Davis, Andrew ***@***.***>; Comment ***@***.***> Subject: Re: [svalinn/DAGMC] Is there a way to view the number of cells in an `.h5m` file? (Discussion #811) Just wondering if this volume code would be useful in the openmc.DAGMCUniverse class 👀 Are the volumes in a DAGMC file always sequential and start at 1? For example volume ids would be 1,2,3,4 — Reply to this email directly, view it on GitHub<#811 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AASTUSTQXIXZK3KFEOIJTNLWAL3PDANCNFSM52T6XVOQ>. You are receiving this because you commented.Message ID: ***@***.***>

[shimwell]

thanks