How can BITS be used in CI?

[prjemian]

It is foreseeable to use BITS in CI, such as GitHub Actions workflows. We can provide advice for this.

[ravescovi]

I think we should define first, what are the set of "actions/commands" that we do when we are testing bits and what we expect out of it. If this is true, then we can create a real set of tests.

• Clone bits
• Install bits
• run python
• run demo_instrument startup
• run demo_plan on demo_instrument
• Create new package
• install bits
• run new_instrument startup
• run demo_plan on new_instrument
• Create new_device on demo_instrument
• Create new_plan for new_device
• Run new_plan with new_device on new_instrument

If this set of actions works, I think bits is doing its job?

[ravescovi]

Different deployments will be a different stages on this list and we could advice on how to test "from this point forward"

[prjemian]

This Issue is not about creating CI for testing BITS, it's about using BITS to set up a bluesky instrument for testing other packages. That will be very soon on my TODO list.

[ravescovi]

Like APS-tools? This is what I mean by "create new_instrument" and "create new_device". We could expand to almost re-create the whole set of GP devices and test against the GP-container?

…

On Mon, Mar 17, 2025 at 3:57 PM Pete R Jemian ***@***.***> wrote: This Issue is not about creating CI for testing BITS, it's about using BITS to set up a bluesky instrument for testing other packages. That will be very soon on my TODO list. — Reply to this email directly, view it on GitHub <#64 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEFMPQ6Z4M7VMA2CKWKGIR32U4ZNPAVCNFSM6AAAAABZCRKTW6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOMZQHEYDSOJQGQ> . You are receiving this because you commented.Message ID: ***@***.***> [image: prjemian]*prjemian* left a comment (BCDA-APS/BITS#64) <#64 (comment)> This Issue is not about creating CI for testing BITS, it's about using BITS to set up a bluesky instrument for testing other packages. That will be very soon on my TODO list. — Reply to this email directly, view it on GitHub <#64 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEFMPQ6Z4M7VMA2CKWKGIR32U4ZNPAVCNFSM6AAAAABZCRKTW6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOMZQHEYDSOJQGQ> . You are receiving this because you commented.Message ID: ***@***.***>

[prjemian]

Right. The scheme I imagine is to install BITS from the template, then git patch the modifications from a file created with some form of git diff. Probably good to go in a few steps.

[ravescovi]

I am confused about the git patch approach but I think it will be clear once I see it. I also keep thinking that once BITS is on pypi, this will be much more streamlined.
As far as testing environments go, have you seen this caprotro mini beamline ??

[prjemian]

I've seen that. It is trivial in comparison to what the synApps-based IOC provides. As a test case for developing caproto, it is not realistic for simulating any APS beamline. Fundamentally, the caproto IOC provides EPICS ChannelAccess PVs but lacks the internal database necessary for processing of records by the IOC's internal database management. It cannot simulate the sscan record. Since the sscan record is used copiously at APS for so-called fly scans, the caproto mini beamline is not interesting to test software for use at APS.

[prjemian]

the git patch approach

takes BITS as-delivered-from-template and applies a configuration (patch file) from a working instrument in a single step rather than a sequence of steps. The patch file is updated as desired from working instrument configuration (avoids the need to edit the "sequence of steps").

[ravescovi]

It seems this will be something like

• clone DEMO-bits
• pip install apsbits
• create-bits newbits
• Add/edit devices
• Add/edit plans
• check against container.

Feels pretty clean to me now.

[prjemian]

Might be easier still. Make a repo for the BITS instrument to use, then pick it up in the workflow (from tarball of GitHub repo or clone with minimal depth).