• Add units to override time field
• Remove eV spectrogram (for now)
• Add enum display of BYKIK rate
• Remove dialog boxes from override activate, deactivate, and fault ack/reset buttons

Remove pulse energy field from all areas of the diagnostic.

Periodic error spam in the terminal:

js: WebSocket connection to 'ws://ctl-logsrv01:3000/ctl/grafana/api/live/ws' failed: Error during WebSocket handshake: Unexpected response code: 403

I am on tmo-console because I'm doing a cursory check of #51

The column headers and the table contents size and scroll independently, sometimes creating a desync:

Small nitpick:

This might be slightly more intuitive to read if the eV range values are left-justified (or is it top-justified?).

Probably need to figure out why and submit a patch (or change code here)

Both lines use the same diagnostic tool, but an instance might be launched for the L line or the K line. Right now it's not immediately obvious which line (KFE or LFE) the diagnostic is pointing to.

Need something to distinguish between the two lines.

Currently there's a much-repeated pattern in this repository where we create a pydm channel, connect it, and set it to be cleared when the containing widget is destroyed.

This should be consolidated into a re-usable function, either specifically here or in a shared utilities repository.
Conversation here: #53 (comment)

If the arbiter clock drifts again, the user should know by how much at a glance! If this number gets bad, we should get a call sooner rather than later.

Let's add a red +0, +1, etc. seconds diff counter that compares the arbiter PV with your local system clock.

I was lurking on the lfe screen today and I noticed that the majority of PVs we check are unused. We could probably do better by configuring smaller numbers here and cut the startup time.

Either an embedded ms calculator, for translation from Months/Weeks/Days/Hours/Minutes or some other way to enter a time for the overrides time.

Here's the minimum size. Note that when you extend this horizontally, the output indicator continues to be glued to the right side of the screen. The large empty space and no row lines makes it very hard to see what's going on here.

This is the most useful filter, might as well start with it open- consider the consequences though, because it might be unclear if the gui is loading properly when it opens with a blank page

There's (one? several?) accelerator link nodes that show the status of the system fast faults. These should be included in the diagnostic so that if there's some sort of problem with the node it becomes apparent to anyone trying to look for issues in the existing diagnostic pages.

There was an issue this weekend where there was a problem with the link node but it wasn't clear why the photon-side systems were faulting the KFE line.

Top-level "FAULT" or "NO FAULTS" to make it clear that there isn't an issue without having to dig around.
Starts as "LOADING", probably

It'd be nice to include a summary of the veto logic in the diagnostic

Need to add a button to Play/Pause the beam. It enter a requested rate of 0Hz and hit apply to the Line Beam Parameters Control element and be accessible outside the tab views (so always available).

Currently, this is an application that lives outside of a python environment.

If this could be installed inside a python environment via pip or conda, then it would be easier to distribute to places like the accelerator prod machines.

Expected Behavior

There are places where we import modules/functions inside functions, rather than in the module preamble. For example:

Current Behavior

We should structure the package more sensibly

Possible Solution

Copy paste, see if it breaks?

Context

#100

Your Environment

pcds-5.8.1

It's possible for twincat IOCs to be in a state where they are reading from stale memory locations after a PLC update. In these cases the IOC needs to be rebooted.

In the GUI, this may look something like this:

There are probably numerous ways we could look at the incoming data and produce suitable "reset the IOC please" warnings.

The simplest would be to catch negative assertion ids and cohorts.

As of now the code uses the YAML configuration file at the current folder.
We need to add an option to specify the file to be used so we can launch the PMPS UI for LFE and KFE.

Rendering using MobaXterm, Windows 10, 13" laptop, 3200x1800.

PMPS display can only show a few words in fast faults, and some of the filter mechanisms are obscured.

For large cycle counts, we get negative numbers in the PLC IOC status tab because EPICS is signed while the underlying data type in the PLC is unsigned. It should be possible to force these to always be positive.

These may be pointing to the overall system readbacks instead of the line beam parameters readback, which makes the screen confusing to use when other devices are restricting the beam.

Expected Behavior

We should have some visual indication of loading progress

Current Behavior

Screen takes a very long time to load, with little feedback outside of command line print statements

Possible Solution

QSplashScreen was my first look at this, and led me to go about refactoring the package to expose the entrypoints to our manipulation.

This likely has to go in the PMPS class, since the QApplication is created during PyDMApplication instatntiation.

Context

Jira ticket link

Your Environment

pcds-5.8.1

Should also rename "Requested" column to "Limit"

Would be really helpful to be able to sort or filter the preemptive request list like the fast fault list.

As a follow up to #100, set up our standard docs workflow.

Be sure to include pre-release notes

Screenshot:

When a PLC doesn't have enough PMPS fast fault allocations, the PLC output will fault without any other indication in the pmps-ui.

Goal: improve the diagnostic to let us know when this situation is happening.

We currently show the counts of allocated/in use/faulted FFOs on the ioc/plc status page. Since the counts here are summed from FFO1, FFO2, FFO3, ... they can't be considered a useful diagnostic tool for when the PLC has more in-use fast faults than allocated fast faults.

Ideas:

• Separate the allocated/in use counters per FFO
• Create a warning indicator somewhere visible on the main screen when this happens
• Create a fast-fault-looking thing and add it to the fast fault table to represent this event

I guess we have a label that is receiving data it doesn't expect

$ [2023-05-04 10:45:21,834] [INFO    ] - Using PyDM via SSH. Reverting to Software Rendering.
Added 1465 fast faults
Added 410 preemptive requests
Added 29 arbiter outputs
[2023-05-04 10:46:05,214] [ERROR   ] - Could not decode [247 204 177  88   1] using utf_8 at widget named 'device_label'.
[2023-05-04 10:46:05,259] [ERROR   ] - Could not decode [128  63   1] using utf_8 at widget named 'device_label'.
[2023-05-04 10:46:05,908] [ERROR   ] - Could not decode [129  88 189   1] using utf_8 at widget named 'device_label'.
[2023-05-04 10:46:06,758] [ERROR   ] - Could not decode [129  88 189   1] using utf_8 at widget named 'device_label'.

Expected Behavior

CLI should be simple and intuitive (read: follow the other conventions present in our stack)

Current Behavior

The package and accompanying .ui files currently use pyDM's macro substitution to pick the proper config.

Possible Solution

Use normal args to specify the config

Context

#100

Your Environment

pcds-5.8.1

• If a PLC boots and has all FFOs as not in-use, this should be represented somewhere in the diagnostic
• If an IOC is down, this should be represented somewhere in the diagnostic

Awaiting further user data if I can't reproduce it myself.

I am not exactly sure where to direct this question, but the PMPS screens have been freezing. They become non-responsive after being open for ~10 min.
This happened consistently for the 24 hour LX15 shift. The screen would freeze after 10 minutes and need to be Force Quit and relaunched

GUI lags a lot when we try to show 1000 fast faults at once
Might be possible to dynamically assign pages to the faults using a second tab widget, perhaps with numbered tabs at the bottom

Expected Behavior

From Zach:

I think ideally the code looks more like "import the pmps class and create it" 
and less like "here's a string path to a python file to exec", so long as the
pydm application would support it.

Current Behavior

We call PyDMApplication and direct its ui_file arg to the right python file (source)

Possible Solution

Refactor the package a smidge. Doing this might require us to do more PyDM initialization.

A note: I did try creating the display under a vanilla QApplication, but got a lot of EPICS connection errors. For example this fails on teardown:

app = QApplication([])
display = PMPS()
display.show()
app.exec()

Context

Brought up in #100

Your Environment

pcds-5.8.1

Under the line parameters control tab, Photon Energy Ranges needs to be adjustable by selecting bits, in a manner similar to how it's displayed in the Beam Parameter Status panel.