-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
External repository for header specification #11
Comments
@trautmane is currently handling dat files. @virginiascarlett is thinking about overarching data flow issues particularly in the realm of open data and recently started to tackle these issues. I'm not sure if there is another way other than LabView license or a PDF / SVG print out of the "code". That said I think we could compare it against more data that we have on hand. We could also compare to other non-LabView code. Should we organize a coordination meeting with @trautmane and @virginiascarlett? |
I'm not too savvy with the technical questions, but I would be happy to facilitate a meeting about this. Mark @mkitti has a lot on his plate right now, but maybe Eric @trautmane or I can help you out. |
Hi @clbarnes - I'll defer to @mkitti regarding anything related to LabView since I can't help much there. I am happy to draft/add a v9 TSV spec to your jeiss-specs repo based on what I know/use from @d-v-b 's fibsem tools repo. With a v9 spec, I can then try to integrate your HDF5 converter into my pipelines and replace the fibsem tools based code I currently use. I support the idea of there being one definitive dat parsing and low-level conversion library (and am happy for that to be your library). I don't have time this week to work on this, but will try to take a look next week. I'm also available to meet next week or later if you think that would help in any way. Best, |
@clbarnes Here is the internal documentation from LabView. I looked over the LabView code quickly and I do not see anything missing from this list. Let me know if you have any specific questions. The next thing I would examine is perhaps the MATLAB and Python code.
|
Here is the MATLAB code
|
Note that we are currently on version 9. |
That's very useful, thanks! The names in my TSVs are from the MATLAB implementation although it would be really helpful to have the units as that original doc from the labview code does. My TSVs were initially generated from @d-v-b 's code so, barring any fixes after that date, should be in sync there. I'm available for a meeting next week! I've just updated my converter to optionally ingest the CSVs the microscopes generate, as I think the metadata provided there has overlap with but isn't entirely contained by the header metadata. P.S. Is that a temperature in Fahrenheit to go with the non-ISO dates 😁 Also a floating point file length? |
Apparently. Now I'm curious what scientific sensor actually reports in Fahrenheit now. |
I am relatively free on Wednesday morning. |
I'm glad it's signed, at least, need to account for those files of negative length. Wednesday afternoon works for me! |
@mkitti : that MATLAB code seems to read |
yeah - I ran into that bug last year ( see janelia-cellmap/fibsem-tools#17 ) |
I spoke with @dmilkie about this, and he pointed to LabView community edition: https://www.ni.com/en-us/shop/labview/select-edition/labview-community-edition.html |
@mkitti @trautmane would you still be available to talk tomorrow about finalising the external spec? I'm free until around 1930 UTC. |
I can meet tomorrow anytime between 1300 and 1800 UTC (I have another meeting at 1830). |
Shall we say 1600 UTC? |
sure - I just emailed you a zoom link - let me know if you did not receive it |
I started https://github.com/clbarnes/jeiss-specs/ as a way to share, in a language-agnostic way, information about the header format. This could then be used as a concrete base for reader implementations, either with code generation or to read dynamically. Having a single source of truth in this way seemed valuable to me. I implemented a lossless HDF5 converter (which can provably regenerate the original .dat from the HDF5) based on that: https://github.com/clbarnes/jeiss-convert .
To me, the main missing piece is actually comparing the data in that repo to the labview implementation. @mkitti , is there a way I can do that without a labview license/ digging handfuls of spaghetti out of the SVG representation, or might you be able to take a look? I'm sure most users are onto v8 or higher by now so I'd be happy to call v1-7 support experimental if we could just get v8 right.
The text was updated successfully, but these errors were encountered: