I've seen that

cace-gui

reads a .txt and writes another .txt with a lot of extra content. Should we use that file the next time we execute

cace-gui

? In other words, should that file overwrites the previous one?

@aquiles viza: Yes, you should be able to read back any of the output files, and all the results should be there. Please keep a backup of the original around---this tool isn't so well developed that you should trust it that far. Most of the records in the file end up in a python dictionary of key:value pairs. If it doesn't recognize something, then it will just ignore it, and it will get written out again in any output file. Where it says "note" isn't mandatory, though: When it writes HTML output, text in a "note" will appear in the datasheet under the "notes" column, so it's meaningful even if it's not mandatory.

To get a parsed .txt file in python I should use "common.cace_read.cace_read()" right? Maybe I could write a verification test that compares two CACE descriptions and indicates differences, including all fields except "results". I think the best option is to have the original .txt and store an upgradable .txt version with the latest simulation results and the html output. It's interesting to see that

cace_read.py

has a _`__main__`_ entry that evaluates syntax. I'm going to integrate that as a validation flag.

Sorry, I am just getting used to the application environment and am no longer really sure how everything works from the command line. I used to have a

cace.py

which drove the main part of the application (generate netlists and run simulations), which is now

cace_cli.py

. Many of the original files were set up to be run directly and provide some kind of behavior on the command line, as you found for

cace_read.py

. (Attn: @Leo Moser)

Thanks @Leo Moser, maybe I went to an extreme when asking this questions. I will set up

cace

as it is now and then, if something happens, I will keep both informed.

Hi @aquiles viza, currently it is expected to run

cace

and

cace-gui

from the command line. If you need any features that are only available through calling a file directly, I think we should make this feature available from the CLI? Feel free to open any issue at CACE's repository, so we don't forget about them 👍

Eventually, if I found something that should be on the cli, I think the best is to inform with an issue or a pull request, Thanks @Leo Moser

message has been deleted

@Leo Moser: Is the matplotlib-3.8.3 requirement really necessary? I'm running matplotlib-3.6.3 under python-3.9 on my laptop and everything works just fine.

Not it's not! I just had to set some required version, so I used that one. I'll update the package shortly!

I think python3.8.10

Thanks, great to hear!