Add PVPositionerSoftDone #531
Conversation
|
I really want unit testing for this support but this project does not have an EPICS IOC available until #235. |
|
One way to apply a functional test (to evaluate this PR independently) is to configure one of the synApps userCalcs ( swait configuration details
In [6]: tsim = PVPositionerSoftDone("gp:userCalc8", name="tsim", readback_pv=".VAL", setpoint_pv=".B", tolerance=1.0)
In [7]: tsim.read()
Out[7]:
OrderedDict([('tsim',
{'value': 29.90983443961242, 'timestamp': 1632159680.426064}),
('tsim_setpoint',
{'value': 29.537445639734493, 'timestamp': 1632159674.42613})])
In [8]: tsim.move(25)
Out[8]: MoveStatus(done=True, pos=tsim, elapsed=2.1, success=True, settle_time=0.0)
In [9]: tsim.move(35)
Out[9]: MoveStatus(done=True, pos=tsim, elapsed=9.8, success=True, settle_time=0.0)
In [10]: tsim.move(25)
Out[10]: MoveStatus(done=True, pos=tsim, elapsed=9.4, success=True, settle_time=0.0) |
|
The functional test seems about right. Device reported comparable times to raise and lower the (simulated) temperature by the same amount to a fixed tolerance of 1 degree. Given the noise on this signal, a lower tolerance could invoke an indefinite wait. |
apstools/devices.py
Outdated
| Motion tolerance. The motion is considered done if | ||
| `abs(readback-setpoint) <= tolerance`. Defaults to `10^(-1*precision)`, | ||
| where `precision = setpoint.precision` | ||
| target_attr : str, optional |
There was a problem hiding this comment.
Can the target_attr kwarg be refactored into an always-present attribute named target which is initially set to None? Means more logic changes below but less likely to be misconfigured (or means more easily understood) by end-user of this device.
|
Also, complexity of |
|
This pull request introduces 2 alerts when merging 486e7bd into 40180d1 - view on LGTM.com new alerts:
|
|
This is a candidate for unit testing with an EPICS IOC (#235). |
|
Needs to resolve merge conflicts with main branch. |
remove target_attr, keep target signal @gfabbris Thanks!
The
PVPositionerSoftDoneis basically the same asophyd.PVPositionerthat uses a soft (instead of EPICS) done signal. A few other (current) differences:where, for instance, the readback PV is "prefixPV:myreadbackPV". The advantage of using this is that the readback/setpoint signals are already marked as hinted/normal and forces auto monitor and PV completion. Alternatively, you can still do something like:
or
If not set, it will use
10^(-1*precision), where the setpoint precision is used.PVPositionerSoftDonewill recognize as a "done" motion before it's truly done. To fix this, you can define a dummytargetsignal:Closes #530.