Why use pythonIOC?

EPICS IOCs are flexible and modular, why do we need to wrap it in Python? This page attempts to answer that question and list a few good use-cases for it.

Calculating PVs from other values

Some use cases require PVs to be calculated from multiple sources. This is possible in EPICS records with calc or aSub records, but pythonIoc allows you to write this as:

import numpy as np
from cothread.catools import caget, camonitor
from softioc import builder, softioc

# The PVs we want to average and their initial values
PVs = [f"DEVICE{i}:CURRENT" for i in range(100)]
values = np.array(caget(PVs))

# The PV we want to serve
avg = builder.aOut("AVERAGE:CURRENT", np.mean(values))

# Start the IOC
builder.LoadDatabase()
softioc.iocInit()

# Make a monitor on the PVs to keep the value up to date
def update_avg(value: float, index: int):
    values[index] = value
    avg.set(np.mean(values))

camonitor(PVs, update_avg)

# Leave the IOC running with an interactive shell.
softioc.interactive_ioc(globals())

Dynamically created PVs

Other use cases will require PVs to be created based on the features of a device. As EPICS database records are statically created at IOC boot, you can generate the database with a script, but pythonIoc allows you to write this as:

from my_device_lib import Device
import cothread
from softioc import builder, softioc

# Connect to the device
device = Device("hostname")
device.connect()

# Make a record for each parameter it has
records = {}
for param in device.get_params():
    records[param] = builder.aIn(param.name, param.value)

# Start the IOC
builder.LoadDatabase()
softioc.iocInit()

# Poll the device for changes and update the records
def update_params():
    while True:
        for param in device.poll_changed_params():
            records[param].set(param.value)

cothread.Spawn(update_params)

# Leave the IOC running with an interactive shell.
softioc.interactive_ioc(globals())

Existing Python Support

It may be that you have specific device support written in Python that you wish to expose as PVs. This could be either in the form of a device support library or using a Python library to calculate PV values as above.