Re: Generation of state diagram documentation from SNL

[Ed Janke via Tech-talk <tech-talk at aps.anl.gov>]

Hi Rees,

 

Thanks for reaching out and the suggestion – I gave both ChatGPT and CoPilot a try, but it looks like many of the sequence files I tried breached the character limits of the tool (sequence contains (48k characters)!  Clearly I need to be making smaller sequences, haha. 

 

But still an interesting idea – I’ll keep hunting around for other LLM/systems that might have larger character limits as this could still prove interesting.

 

 

Cheers,

Ed Janke

 

From: Rees Simon <simon.rees at psi.ch>
Date: Friday, June 27, 2025 at 12:05 AM
To: Ed Janke <ejanke at keck.hawaii.edu>, tech-talk at aps.anl.gov <tech-talk at aps.anl.gov>
Subject: Re: Generation of state diagram documentation from SNL

[You don't often get email from simon.rees at psi.ch. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]

Dear Ed

An alternative, low-effort, approach that might work as a “quick
look” at the states in your SNL file, is to ask ChatGPT to
generate a PlantUML (or perhaps Mermaid) diagram for you.

Just for fun I asked the GPT-4o model to visualize the states
in the homing procedure that initializes one of our
stepper-motor controlled range-shifter devices.

It seems to have made a reasonable job of it (see attached
screenshot), though not as detailed as the diagram I made
years ago, laboriously, using a UML modelling tool.

Just throwing the idea out there, although most of you are
probably already aware of it…

Best regards

Simon Rees


On 27.06.2025 02:55, Ed Janke via Tech-talk wrote:
> Hello!
>
> Does anyone know if there is software out there that can parse the SNL
> in an EPICS Sequence file to generate a state diagram for every state set?
>
> I was considering doing this by hand but figured I would ask the greater
> community in case this already exists.
>
> Thanks,
>
> Ed Janke
>