Requesting BuildElemLeadField module in SCIRun 5. #2391
Comments
|
Thanks for making the issue here. I looked at the code for that module, and I have good news and bad news. The good news is there is no UI, no algorithm class, and simple inputs (matrix and field). Every line of code but one will be trivial to convert. So other than the one line I can convert this module very easily. That one line though, is the bad news, and relates to the documentation:
SCIRun 5's network editor does not allow loops (yet)! However, there might be a solution. Recently we prototyped a set of modules that use the python interpreter to simulate looping with state variables, and that might be useful here. It will be an interesting challenge at least, and I have an example network to work with from the BioPSE package to guide me. So I've slotted this issue in the next milestone (not the current one, which will be released soon). |
|
I see, ok
Thank you for your help.
In the meantime, I've been having some issues regarding the module in 4.7 where it appears to not run but without any error messages or indications why it's frozen.
Would you be able to help me?
Taseen
…________________________________
From: Dan White ***@***.***>
Sent: Tuesday, July 12, 2022 8:30 PM
To: SCIInstitute/SCIRun ***@***.***>
Cc: Taseen Jawad ***@***.***>; Author ***@***.***>
Subject: Re: [SCIInstitute/SCIRun] Requesting BuildElemLeadField module in SCIRun 5. (Issue #2391)
You don't often get email from ***@***.*** Learn why this is important<https://aka.ms/LearnAboutSenderIdentification>
Thanks for making the issue here. I looked at the code for that module, and I have good news and bad news. The good news is there is no UI, no algorithm class, and simple inputs (matrix and field). Every line of code but one will be trivial to convert. So other than the one line I can convert this module very easily. That one line though, is the bad news, and relates to the documentation<https://can01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fscirundocwiki.sci.utah.edu%2FSCIRunDocs%2Findex.php%2FCIBC%3ADocumentation%3ASCIRun%3AReference%3ABioPSE%3ABuildElemLeadField&data=05%7C01%7Ctaseen.jawad%40mail.utoronto.ca%7C1475b10e9f4b4e28a3f908da6466e0bd%7C78aac2262f034b4d9037b46d56c55210%7C0%7C0%7C637932690266873358%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=9ID9KBKiAo6OHaJSn0CL0s6n%2FJTxZvP%2BfX0PGUvmKeY%3D&reserved=0>:
This module is one of the few that allows, and requires, a looped input.
SCIRun 5's network editor does not allow loops (yet)! However, there might be a solution. Recently we prototyped a set of modules that use the python interpreter to simulate looping with state variables, and that might be useful here. It will be an interesting challenge at least, and I have an example network to work with from the BioPSE package to guide me. So I've slotted this issue in the next milestone (not the current one, that is to be released soon).
—
Reply to this email directly, view it on GitHub<https://can01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FSCIInstitute%2FSCIRun%2Fissues%2F2391%23issuecomment-1182631789&data=05%7C01%7Ctaseen.jawad%40mail.utoronto.ca%7C1475b10e9f4b4e28a3f908da6466e0bd%7C78aac2262f034b4d9037b46d56c55210%7C0%7C0%7C637932690266873358%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=rZUyNMItafUXDmzuBDT4BWCpsYz56iF8MxH2nP4JKDw%3D&reserved=0>, or unsubscribe<https://can01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FA2A2BGPGF6MXLTVDVBI4NMLVTYE2BANCNFSM53LHMQBA&data=05%7C01%7Ctaseen.jawad%40mail.utoronto.ca%7C1475b10e9f4b4e28a3f908da6466e0bd%7C78aac2262f034b4d9037b46d56c55210%7C0%7C0%7C637932690266873358%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Fc4XUYHjDFBZceu1JsNNhnTmL6fIBj9oVxE%2Fxu%2BwxAQ%3D&reserved=0>.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
|
4.7 is no longer in development, but if you sent a network and some (small) data, I could test your setup and see how to fix it in version 5 perhaps. The sample network included in 4.7 was incomplete, but it did seem to work with the looping construct (it just didn't output or visualize anything interesting). If you attach a zip file here I can investigate further. To get some more direct help with 4.7, I am mentioning @jessdtate who is the resident v4 expert. |
![https://r1.res.office365.com/owa/prem/images/dc-zip_20.png]<https://utoronto-my.sharepoint.com/personal/taseen_jawad_mail_utoronto_ca/Documents/Attachments/Taseen_net_and_mesh_files(1).zip](https://r1.res.office365.com/owa/prem/images/dc-zip_20.png]<https://utoronto-my.sharepoint.com/personal/taseen_jawad_mail_utoronto_ca/Documents/Attachments/Taseen_net_and_mesh_files(1).zip){kind=link}
|
To view Taseen_net_and_mesh_files(1).zip, sign in<https://utoronto-my.sharepoint.com/personal/taseen_jawad_mail_utoronto_ca/_layouts/15/acceptinvite.aspx?invitation=%7B32509F11%2D254F%2D42FF%2DBE41%2DEB5270FA4864%7D&listId=d71aafeb%2D6f7d%2D4a6f%2Dbb66%2De247fb17035f&itemId=2dbeb14d%2Deaf8%2D46d7%2D8020%2D4488c5ff831e> or create an account.
|
|
There is a similar example in the forward inverse toolkit, the make lead field matrix https://github.com/SCIInstitute/FwdInvToolkit/blob/master/Networks/potential-based-fem/make_lead_field_matrix.srn5. It uses some different modules but it should have the same functionality. |
|
we may need to rethink this module to help it fit better in our current architecture and use cases. Let's brainstorm some ideas next week. |
|
The key to the lead field matrix is that it is a linear relationship between each individual source and each observation point. It essentially consists of a concatenation of the solution to each individual source. The key points to the provided example are the getmatrixslice, which allows you to iterate through all the source points, and the collectmatrix, which concatenates the solutions into the leadfieldmatrix. Between these modules would be the same as what you would have within the loop in the BuildElemLeadField uses. I haven't had a chance to look at your network, but I will to see if I can provide more specific feedback. |
|
I'm having trouble accessing the files through sharepoint. Is there a way to make them more public for a short time? We can move to email if you prefer. |
Description
BuildElemLeadField module appears to exist in 4.7 but not 5. Requesting an equivalent module in SCIRun 5.
Versions
5
The text was updated successfully, but these errors were encountered: