Potential AntiDoc addon for TestStand - suggest features

[leahmedwards]

It was great to meet so many people at GDevCon#4 Europe!

Olivier mentioned in his presentation that I'd started a proof of concept TestStand addon for AntiDoc, to automatically document .seq files. A few people asked me about this and requested features so I thought I'd open up the discussion 🙂

My rationale for making this addon is that:

1. NI's Tools>Sequence File documentation is not as visually polished as AntiDoc reports
2. I wanted more visualisation in the documentation such as that given in 'View > Sequence File > Hierarchy', e.g. also showing variable usage
3. There are other details that I would like to include in the report. I'd made a tool at work specific to our architecture that achieved some of these things, so I knew it was possible! Cyril Gambini (another collaborator) also advised me on what he would like to see.

You can view an example of what I have made so far, attached at the bottom of the post. It is a pdf and html report for the NI example "C:\Users\Public\Documents\National Instruments\TestStand 2017 (32-bit)\Examples\Demos\Computer Motherboard Test\LabVIEW\Computer Motherboard Test Sequence.seq"

This is only a proof of concept, so there are many improvements needed!

Features/Improvements I would like to add

1. Fix variable parsing recursion bug (perhaps different approach needed)
2. Improve document organisation (subheadings, links to other sections?)
3. In the graphs, group variable types e.g. file globals, parameters etc. into subgraphs - involves updating AntiDoc's graphBuilder.lvlib dependency version
4. Direction of arrows on variables graph should indicate read/write - need to look at how, not sure if checking which side of the equals sign the variable is on covers all cases or if there is a better way
5. Add pie/bar chart of module type of externally called code (LabVIEW, python, .NET etc.). Because it's useful to see at a glance what languages are used.
6. Add pie/bar chart of step types in sequence file by % (maybe unnecessary but interesting)
7. List custom data types info (structure, version)
8. List custom step types used
9. State whether results output is disabled for all steps of subsequence - could be displayed in subsequence table with columns: subsequence name, description/comment, type (Normal/Callback/EntryPoint etc), link to section, symbol if results output is disabled for all steps
10. Show whether parameter is an input, output or both. As a LabVIEW person, I like seeing whether data will be changed
11. For subsequence, add table of external files called by steps: File, type (adapter/from extension?), path, calling step(s), VI icon image if is VI
12. Add AntiDoc menu item to TestStand IDE - currently launched from LabVIEW!

I'm not interested in putting too much of the step details in (in the same way that AntiDoc LabVIEW addons don't include the block diagram) as it can be seen in the IDE and test reports, and is very lengthy.

Improvements/Features Suggested by Community

1. Have optional 'show requirements on steps' setting

I'll add more of these as we go. Please put any suggestions or feedback in this thread!

[Petru_Tarabuta]

Hi Leah,

Thanks for working on a TestStand addon for AntiDoc! It sounds very useful, and I would definitely use it. The Sequence-Documentation.pdf in the attached zip file is already useful.

Is the plan for the addon to be free to use and open-source or paid-for and closed-source? (I'd be happy either way)

If open-source, it might be useful to use the issue tracking functionality of the git repository platform (GitHub, Gitlab, etc) to track the features/improvements. For example, if using GitHub, a GitHub issue could be created for each feature/improvement. This essentially enables the use of a rich-text webpage to describe the feature, including the ability to add screenshots, and for other users to leave comments or ask questions.

Thanks

[leahmedwards]

Hi Petru.

Glad to hear you think it would be useful. 🙂

The addon will be free and open source. Though I'm sure Olivier appreciates any contributions to his Patreon if people find the tool useful.

The code is already hosted on GitLab here and I've made a few (very basic) issues already - hoping to make more once I've figured out what features people want! I just thought a forums post would be a good way of gathering wider feedback first. There's also a Zulip Chat channel if you want to discuss any particular feature in detail.

Cheers,

Leah

[Petru_Tarabuta]

Thanks!

[LFBaute]

Im very interested in this topic, if any information is shared besides gitlab, Id like to be added. Currently looking at the AntiDoc Zulip chat, Thanks!

[Olivier-JOURDAN]

---

@LFBaute  a écrit :

Im very interested in this topic, if any information is shared besides gitlab, Id like to be added. Currently looking at the AntiDoc Zulip chat, Thanks!

---

It's great to see that many people interested in Leah's project 👍

The GitLab project is public. You can access all issues, add issues, comments, etc…

Issues are made to organize the work and keep track of the issue.

I'm actively encouraging more open discussions to take place in the Zulip chat and only create issues when it's necessary

I'm not a TestStand developer. I can't help with that part, but I'm definitely here to help with all the Antidoc/Asciidoc stuff. 

That being said, Leah is the boss of this project, so I let her lead as she likes 😅

[leahmedwards]

Good suggestion Olivier - means we can check for alignment on how a new feature might work/be implemented before spending lots of time writing issues.

[smi38]

Hi,

Really really cool project.

Point 12 is very easy to achieve with a custom menu entry.

I'd like to have the version and hash of code modules in the documentation.

Example : version and hash of a called LV PPL

Maybe also an option to have the full properties list of a step (big report expected...).

Also, as a workspace or project is not mandatory, it should work with client sequence files alone.

Thx

[leahmedwards]

Thank you for the suggestions! PPL versions world be super useful, agreed. 

I'm starting with .seq files only, so yes it will definitely work without workspaces or projects.