Skip to content

Latest commit

 

History

History
178 lines (153 loc) · 7.37 KB

README.md

File metadata and controls

178 lines (153 loc) · 7.37 KB
Raw

liteserver 3.1.0

Very Lightweight Data Object Server. It hosts Lite Data Objects (LDO, analog of Process Variables in EPICS) and provides info/set/get/read/subscribe remote access to them using UDP protocol. Data encoding is implemented using CBOR specification, which makes it very fast and efficient.

Data logging and retrieving

Data objects can be logged and retrieved using an apstrim package (https://pypi.org/project/apstrim).

Features

  • Simplicity. The network protocol is UDP, error correction of late/missing/mangled data is implemented. Binary serialization protocol CBOR. Programming interface is very similar to JSON.
  • Low latency, connection-less.
  • Supported requests:
    • info(), returns dictionary with information on requested LDOs and parameters
    • get(), returns dictionary of values of requested LDOs and parameters
    • read(), returns dictionary of changed readable (non-constant) parameters of requested LDOs
    • set(), set values or attributes of requested LDO parameters
    • subscribe(callback), subscribe to a set of the objects, if any object of the requested LDOs have been changed, the server will publish the changed objects to client and the callback function on the client will be invoked.
  • Multidimensional data (numpy arrays) are supported.
  • Access control info (username, program name) supplied in every request
  • Name service
    • file-based
    • network-based using a dedicated liteServer (not commissioned yet)
  • Basic spreadsheet-based GUI: pypeto
  • Architectures. All programs are 100% python. Tested on Linux and partially on Windows.
  • Companion applications:

Supported devices

Server implementations for various devices are located in .device sub-package. A device server can be started using following command:
python3 -m liteserv.device.

  • device.litePeakSimulator: Waveform simulator with multiple peaks and a background noise.
  • device.liteScaler: Test implementation of the liteServer, supporting 1000 of up/down counters as well as multi-dimensional arrays.
  • device.senstation: Server for various devices, connected to Raspberry Pi GPIOs: 1-wire temperature sensor, Pulse Counter, Fire alarm and Spark detector, Buzzer, RGB LED indicator, OmegaBus serial sensors. Various I2C devices: ADC: ADS1x15, Magnetometers: MMC5983MA, HMC5883, QMC5983, TLV493D. I2C multiplexing using TCA9548 or PCA9546. NUCLEO-STM33 mixed signal MCU boards, connected to Raspberry Pi over USB.
  • device.liteGQ: Geiger Counter and a gyro sensor GMC-500 from GQ Electronics.
  • device.liteWLM: Wavelength Meter WS6-600 from HighFinesse.
  • device.liteLabjack: LabJack U3 analog and digital IO module.
  • device.liteUSBCam: Server for USB cameras.
  • device.liteUvcCam: Server for USB cameras using UVC library.
  • device.liteVGM: (Obsolete) Server for multiple gaussmeters from AlphaLab Inc.

Installation

Dependencies:

  • Python3 3.6 or higher.
  • CBOR2

Installation:

python3 -m pip install liteserver

Additional libraries may be required for specific devices.

Examples

Most convenient way to test a base class functionality is by using ipython3,

Test server: liteScaler

Start liteScaler on a local host:
python -m liteserver.device.liteScaler -ilo
To monitor, use:
python -m pvplot L:localhost:dev1:counters

Peak simulator

python -m liteserver.device.litePeakSimulator -ilo
To monitor, use:
python -m pvplot -s.01 -a'L:localhost:dev1' 'x,y'

Labjack U3-HV

python -m liteserver.device.liteLabjack -ilo
To monitor, use:
python -m pvplot -a'L:localhost:dev1' 'tempU3 ADC_HV[0] ADC_HV[1] ADC_HV[2] ADC_HV[3] ADC_LV'

Server for all supported peripherals: senstation:

I2C Support

To detect available devices on the multiplexed I2C chain:
python -m utils.i2cmux -M 112
If multiplexer address on your board is not 112, you can find it using:
i2cdetect -y 1

Start the senstation server

Start the senstation server through default network interface:
python -m liteserver.device.senstation

Interfacing to senstation using python

from liteserver import liteAccess as LA 
from pprint import pprint

Host = 'localhost'
LAserver = Host+':server'
LAdev1   = Host+':dev1'
LAdev2   = Host+':dev2'

#``````````````````Programmatic way, using Access`````````````````````````````
LA.Access.info((Host+':*','*'))# map of all devices and parameters the Host
LA.Access.info((LAserver,'*'))
LA.Access.get((LAserver,'*'))
LA.Access.set((LAdev1,'frequency',2.0))
LA.Access.subscribe(LA.testCallback,(LAdev1,'cycle'))
LA.Access.unsubscribe()
#,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,	
#``````````````````Object-oriented way````````````````````````````````````````
# Advantage: The previuosly created PVs are reused.
allServerParameters = LA.PVs((LAserver,'*'))
pprint(allServerParameters.info())
pprint(allServerParameters.get())# get all parameters from device LAserver
# get all readable parameters from device Scaler1:server, which have been 
# modified since the last read:
pprint(allServerParameters.read())

allDev1Parameters = LA.PVs((LAdev1,'*'))
print(allDev1Parameters.info())

server_performance = LA.PVs((LAserver,'perf'))
pprint(server_performance.info())
pprint(server_performance.get())
# simplified get: returns (value,timestamp) of a parameter 'perf' 
pprint(server_performance.value)

server_multiple_parameters = LA.PVs((LAserver,('perf','run')))
pprint(server_multiple_parameters.info())
pprint(server_multiple_parameters.get())

server_multiple_devPars = LA.PVs((LAdev1,('time','frequency')),(LAserver,('statistics','perf')))
pprint(server_multiple_devPars.get())

# setting
dev1_frequency = LA.PVs((LAdev1,'frequency'))
dev1_frequency.set([1.5])
dev1_frequency.value
dev1_multiple_parameters = LA.PVs([LAdev1,('frequency','coordinate')])
dev1_multiple_parameters.set([8.,[3.,4.]])

# subscribing
ldo = LA.PVs([LAdev1,'cycle'])
ldo.subscribe()# it will print image data periodically
ldo.unsubscribe()# cancel the subscruption

# test for timeout, should timeout in 10s:
#,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
#``````````````````Observations```````````````````````````````````````````````
Timing of Access.get using ipython on localhost.
    from liteserver import liteAccess as LA
    Host='localhost'
    LAdev1   = Host+':dev1'
    %timeit image = LA.Access.get((LAdev1,['image']))
    145 µs ± 1.95 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)
    image[(LAdev1,'image')]['value'].shape
    (120, 160, 3)
Retrieving time of 57600 values (120*160*3) is 145 µs,
which corresponds to 400 mValues/s

Retrieving time of 57600 values (120*160*3) 220 µs,
which corresponds to 260 MValues/s (on entry-level workstation). 
It was 400 MValues/s on top-level workstation.
Note: Msgpack was 4% faster.
#``````````````````Tips```````````````````````````````````````````````````````
# To enable debugging: LA.PVs.Dbg = True
# To enable transaction timing: LA.Channel.Perf = True