High level server API¶

Server helper classes for writing Tango device servers.

Device
attribute
command
pipe

device_property
class_property
run()
server_run()

This module provides a high level device server API. It implements TEP1. It exposes an easier API for developing a Tango device server.

Here is a simple example on how to write a Clock device server using the high level API:

import time
from tango.server import run
from tango.server import Device
from tango.server import attribute, command


class Clock(Device):

    time = attribute()

    def read_time(self):
        return time.time()

    @command(din_type=str, dout_type=str)
    def strftime(self, format):
        return time.strftime(format)


if __name__ == "__main__":
    run((Clock,))

Here is a more complete example on how to write a PowerSupply device server using the high level API. The example contains:

a read-only double scalar attribute called voltage
a read/write double scalar expert attribute current
a read-only double image attribute called noise
a ramp command
a host device property
a port class property

from time import time
from numpy.random import random_sample

from tango import AttrQuality, AttrWriteType, DispLevel
from tango.server import Device, attribute, command
from tango.server import class_property, device_property

class PowerSupply(Device):

    voltage = attribute()

    current = attribute(label="Current", dtype=float,
                        display_level=DispLevel.EXPERT,
                        access=AttrWriteType.READ_WRITE,
                        unit="A", format="8.4f",
                        min_value=0.0, max_value=8.5,
                        min_alarm=0.1, max_alarm=8.4,
                        min_warning=0.5, max_warning=8.0,
                        fget="get_current", fset="set_current",
                        doc="the power supply current")

    noise = attribute(label="Noise", dtype=((float,),),
                      max_dim_x=1024, max_dim_y=1024,
                      fget="get_noise")

    host = device_property(dtype=str)
    port = class_property(dtype=int, default_value=9788)

    def read_voltage(self):
        self.info_stream("get voltage(%s, %d)" % (self.host, self.port))
        return 10.0

    def get_current(self):
        return 2.3456, time(), AttrQuality.ATTR_WARNING

    def set_current(self, current):
        print("Current set to %f" % current)

    def get_noise(self):
        return random_sample((1024, 1024))

    @command(dtype_in=float)
    def ramp(self, value):
        print("Ramping up...")

if __name__ == "__main__":
    PowerSupply.run_server()

Pretty cool, uh?

Data types

When declaring attributes, properties or commands, one of the most important information is the data type. It is given by the keyword argument dtype. In order to provide a more pythonic interface, this argument is not restricted to the CmdArgType options.

For example, to define a SCALAR DevLong attribute you have several possibilities:

int
‘int’
‘int64’
tango.CmdArgType.DevLong64
‘DevLong64’
numpy.int64

To define a SPECTRUM attribute simply wrap the scalar data type in any python sequence:

using a tuple: (:obj:`int`,) or
using a list: [:obj:`int`] or
any other sequence type

To define an IMAGE attribute simply wrap the scalar data type in any python sequence of sequences:

using a tuple: ((:obj:`int`,),) or
using a list: [[:obj:`int`]] or
any other sequence type

Below is the complete table of equivalences.

dtype argument	converts to tango type
`None`	`DevVoid`
`'None'`	`DevVoid`
`DevVoid`	`DevVoid`
`'DevVoid'`	`DevVoid`
`DevState`	`DevState`
`'DevState'`	`DevState`
`bool`	`DevBoolean`
`'bool'`	`DevBoolean`
`'boolean'`	`DevBoolean`
`DevBoolean`	`DevBoolean`
`'DevBoolean'`	`DevBoolean`
`numpy.bool_`	`DevBoolean`
`'char'`	`DevUChar`
`'chr'`	`DevUChar`
`'byte'`	`DevUChar`
`chr`	`DevUChar`
`DevUChar`	`DevUChar`
`'DevUChar'`	`DevUChar`
`numpy.uint8`	`DevUChar`
`'int16'`	`DevShort`
`DevShort`	`DevShort`
`'DevShort'`	`DevShort`
`numpy.int16`	`DevShort`
`'uint16'`	`DevUShort`
`DevUShort`	`DevUShort`
`'DevUShort'`	`DevUShort`
`numpy.uint16`	`DevUShort`
`'int32'`	`DevLong`
`DevLong`	`DevLong`
`'DevLong'`	`DevLong`
`numpy.int32`	`DevLong`
`'uint32'`	`DevULong`
`DevULong`	`DevULong`
`'DevULong'`	`DevULong`
`numpy.uint32`	`DevULong`
`int`	`DevLong64`
`'int'`	`DevLong64`
`'int64'`	`DevLong64`
`DevLong64`	`DevLong64`
`'DevLong64'`	`DevLong64`
`numpy.int64`	`DevLong64`
`'uint'`	`DevULong64`
`'uint64'`	`DevULong64`
`DevULong64`	`DevULong64`
`'DevULong64'`	`DevULong64`
`numpy.uint64`	`DevULong64`
`DevInt`	`DevInt`
`'DevInt'`	`DevInt`
`'float32'`	`DevFloat`
`DevFloat`	`DevFloat`
`'DevFloat'`	`DevFloat`
`numpy.float32`	`DevFloat`
`float`	`DevDouble`
`'double'`	`DevDouble`
`'float'`	`DevDouble`
`'float64'`	`DevDouble`
`DevDouble`	`DevDouble`
`'DevDouble'`	`DevDouble`
`numpy.float64`	`DevDouble`
`str`	`DevString`
`'str'`	`DevString`
`'string'`	`DevString`
`'text'`	`DevString`
`DevString`	`DevString`
`'DevString'`	`DevString`
`bytearray`	`DevEncoded`
`'bytearray'`	`DevEncoded`
`'bytes'`	`DevEncoded`
`DevEncoded`	`DevEncoded`
`'DevEncoded'`	`DevEncoded`
`DevVarBooleanArray`	`DevVarBooleanArray`
`'DevVarBooleanArray'`	`DevVarBooleanArray`
`DevVarCharArray`	`DevVarCharArray`
`'DevVarCharArray'`	`DevVarCharArray`
`DevVarShortArray`	`DevVarShortArray`
`'DevVarShortArray'`	`DevVarShortArray`
`DevVarLongArray`	`DevVarLongArray`
`'DevVarLongArray'`	`DevVarLongArray`
`DevVarLong64Array`	`DevVarLong64Array`
`'DevVarLong64Array'`	`DevVarLong64Array`
`DevVarULong64Array`	`DevVarULong64Array`
`'DevVarULong64Array'`	`DevVarULong64Array`
`DevVarFloatArray`	`DevVarFloatArray`
`'DevVarFloatArray'`	`DevVarFloatArray`
`DevVarDoubleArray`	`DevVarDoubleArray`
`'DevVarDoubleArray'`	`DevVarDoubleArray`
`DevVarUShortArray`	`DevVarUShortArray`
`'DevVarUShortArray'`	`DevVarUShortArray`
`DevVarULongArray`	`DevVarULongArray`
`'DevVarULongArray'`	`DevVarULongArray`
`DevVarStringArray`	`DevVarStringArray`
`'DevVarStringArray'`	`DevVarStringArray`
`DevVarLongStringArray`	`DevVarLongStringArray`
`'DevVarLongStringArray'`	`DevVarLongStringArray`
`DevVarDoubleStringArray`	`DevVarDoubleStringArray`
`'DevVarDoubleStringArray'`	`DevVarDoubleStringArray`
`DevPipeBlob`	`DevPipeBlob`
`'DevPipeBlob'`	`DevPipeBlob`

class tango.server.Device(cl, name)¶

Bases: tango.server.BaseDevice

Device class for the high-level API.

All device specific classes should inherit from this class.

add_attribute(self, attr, r_meth=None, w_meth=None, is_allo_meth=None) → Attr¶

Add a new attribute to the device attribute list. Please, note that if you add an attribute to a device at device creation time, this attribute will be added to the device class attribute list. Therefore, all devices belonging to the same class created after this attribute addition will also have this attribute.

Parameters

attr: (Attr or AttrData) the new attribute to be added to the list.
r_meth: (callable) the read method to be called on a read request
w_meth: (callable) the write method to be called on a write request (if attr is writable)
is_allo_meth: (callable) the method that is called to check if it is possible to access the attribute or not

Return

(Attr) the newly created attribute.

Throws

DevFailed

add_command(self, cmd, level=TANGO::OPERATOR) → cmd¶

Add a new command to the device command list.

Parameters

cmd: the new command to be added to the list
device_level: Set this flag to true if the command must be added for only this device

Return

Command

Throws

DevFailed

always_executed_hook()¶: Tango always_executed_hook. Default implementation does nothing

append_status(self, status, new_line=False) → None ¶

Appends a string to the device status.

Parameters: status : (str) the string to be appened to the device status new_line : (bool) If true, appends a new line character before the string. Default is False
Return: None

check_command_exists(self) → None ¶

This method check that a command is supported by the device and does not need input value. The method throws an exception if the command is not defined or needs an input value

Parameters

cmd_name: (str) the command name

Return

None

Throws

DevFailed API_IncompatibleCmdArgumentType, API_CommandNotFound

New in PyTango 7.1.2

debug_stream(self, msg, *args) → None ¶

Sends the given message to the tango debug stream.

Since PyTango 7.1.3, the same can be achieved with:
print(msg, file=self.log_debug)

Parameters

msg: (str) the message to be sent to the debug stream

Return

None

delete_device(self) → None ¶

Delete the device.

Parameters: None
Return: None

error_stream(self, msg, *args) → None ¶

Sends the given message to the tango error stream.

Since PyTango 7.1.3, the same can be achieved with:
print(msg, file=self.log_error)

Parameters

msg: (str) the message to be sent to the error stream

Return

None

fatal_stream(self, msg, *args) → None ¶

Sends the given message to the tango fatal stream.

Since PyTango 7.1.3, the same can be achieved with:
print(msg, file=self.log_fatal)

Parameters

msg: (str) the message to be sent to the fatal stream

Return

None

get_attr_min_poll_period(self) → seq<str>¶

Returns the min attribute poll period

Parameters: None
Return: (seq) the min attribute poll period

New in PyTango 7.2.0

get_attr_poll_ring_depth(self, attr_name) → int ¶

Returns the attribute poll ring depth

Parameters

attr_name: (str) the attribute name

Return

(int) the attribute poll ring depth

New in PyTango 7.1.2

get_attribute_poll_period(self, attr_name) → int ¶

Returns the attribute polling period (ms) or 0 if the attribute is not polled.

Parameters

attr_name: (str) attribute name

Return

(int) attribute polling period (ms) or 0 if it is not polled

New in PyTango 8.0.0

get_cmd_min_poll_period(self) → seq<str>¶

Returns the min command poll period

Parameters: None
Return: (seq) the min command poll period

New in PyTango 7.2.0

get_cmd_poll_ring_depth(self, cmd_name) → int ¶

Returns the command poll ring depth

Parameters

cmd_name: (str) the command name

Return

(int) the command poll ring depth

New in PyTango 7.1.2

get_command_poll_period(self, cmd_name) → int ¶

Returns the command polling period (ms) or 0 if the command is not polled.

Parameters

cmd_name: (str) command name

Return

(int) command polling period (ms) or 0 if it is not polled

New in PyTango 8.0.0

get_dev_idl_version(self) → int ¶

Returns the IDL version

Parameters: None
Return: (int) the IDL version

New in PyTango 7.1.2

get_device_attr(self) → MultiAttribute¶

Get device multi attribute object.

Parameters: None
Return: (MultiAttribute) the device’s MultiAttribute object

get_device_properties(self, ds_class=None) → None ¶

Utility method that fetches all the device properties from the database and converts them into members of this DeviceImpl.

Parameters

ds_class: (DeviceClass) the DeviceClass object. Optional. Default value is None meaning that the corresponding DeviceClass object for this DeviceImpl will be used

Return

None

Throws

DevFailed

get_exported_flag(self) → bool ¶

Returns the state of the exported flag

Parameters: None
Return: (bool) the state of the exported flag

New in PyTango 7.1.2

get_logger(self) → Logger¶

Returns the Logger object for this device

Parameters: None
Return: (Logger) the Logger object for this device

get_min_poll_period(self) → int ¶

Returns the min poll period

Parameters: None
Return: (int) the min poll period

New in PyTango 7.2.0

get_name(self)¶

Get a COPY of the device name.

Parameters: None
Return: (str) the device name

get_non_auto_polled_attr(self) → sequence<str>¶

Returns a COPY of the list of non automatic polled attributes

Parameters: None
Return: (sequence<str>) a COPY of the list of non automatic polled attributes

New in PyTango 7.1.2

get_non_auto_polled_cmd(self) → sequence<str>¶

Returns a COPY of the list of non automatic polled commands

Parameters: None
Return: (sequence<str>) a COPY of the list of non automatic polled commands

New in PyTango 7.1.2

get_poll_old_factor(self) → int ¶

Returns the poll old factor

Parameters: None
Return: (int) the poll old factor

New in PyTango 7.1.2

get_poll_ring_depth(self) → int ¶

Returns the poll ring depth

Parameters: None
Return: (int) the poll ring depth

New in PyTango 7.1.2

get_polled_attr(self) → sequence<str>¶

Returns a COPY of the list of polled attributes

Parameters: None
Return: (sequence<str>) a COPY of the list of polled attributes

New in PyTango 7.1.2

get_polled_cmd(self) → sequence<str>¶

Returns a COPY of the list of polled commands

Parameters: None
Return: (sequence<str>) a COPY of the list of polled commands

New in PyTango 7.1.2

get_prev_state(self) → DevState¶

Get a COPY of the device’s previous state.

Parameters: None
Return: (DevState) the device’s previous state

get_state(self) → DevState¶

Get a COPY of the device state.

Parameters: None
Return: (DevState) Current device state

get_status(self) → str ¶

Get a COPY of the device status.

Parameters: None
Return: (str) the device status

info_stream(self, msg, *args) → None ¶

Sends the given message to the tango info stream.

Since PyTango 7.1.3, the same can be achieved with:
print(msg, file=self.log_info)

Parameters

msg: (str) the message to be sent to the info stream

Return

None

init_device()¶: Tango init_device method. Default implementation calls get_device_properties()

initialize_dynamic_attributes()¶: Method executed at initializion phase to create dynamic attributes. Default implementation does nothing. Overwrite when necessary.

is_device_locked(self) → bool ¶

Returns if this device is locked by a client

Parameters: None
Return: (bool) True if it is locked or False otherwise

New in PyTango 7.1.2

is_polled(self) → bool ¶

Returns if it is polled

Parameters: None
Return: (bool) True if it is polled or False otherwise

New in PyTango 7.1.2

is_there_subscriber(self, att_name, event_type) → bool ¶

Check if there is subscriber(s) listening for the event.

This method returns a boolean set to true if there are some subscriber(s) listening on the event specified by the two method arguments. Be aware that there is some delay (up to 600 sec) between this method returning false and the last subscriber unsubscription or crash…

The device interface change event is not supported by this method.

Parameters

att_name: (str) the attribute name
event_type (EventType): the event type

Return

True if there is at least one listener or False otherwise

push_archive_event(self, attr_name) → None ¶

push_archive_event (self, attr_name, except) -> None

push_archive_event (self, attr_name, data, dim_x = 1, dim_y = 0) -> None

push_archive_event (self, attr_name, str_data, data) -> None

push_archive_event (self, attr_name, data, time_stamp, quality, dim_x = 1, dim_y = 0) -> None

push_archive_event (self, attr_name, str_data, data, time_stamp, quality) -> None

Push an archive event for the given attribute name. The event is pushed to the notification daemon.

Parameters

attr_name

(str) attribute name

data

the data to be sent as attribute event data. Data must be compatible with the attribute type and format. for SPECTRUM and IMAGE attributes, data can be any type of sequence of elements compatible with the attribute type

str_data

(str) special variation for DevEncoded data type. In this case ‘data’ must be a str or an object with the buffer interface.

except

(DevFailed) Instead of data, you may want to send an exception.

dim_x

(int) the attribute x length. Default value is 1

dim_y

(int) the attribute y length. Default value is 0

time_stamp

(double) the time stamp

quality

(AttrQuality) the attribute quality factor

Throws

DevFailed If the attribute data type is not coherent.

push_att_conf_event(self, attr) → None ¶

Push an attribute configuration event.

Parameters: (Attribute) the attribute for which the configuration event will be sent.
Return: None

New in PyTango 7.2.1

push_change_event(self, attr_name) → None ¶

push_change_event (self, attr_name, except) -> None

push_change_event (self, attr_name, data, dim_x = 1, dim_y = 0) -> None

push_change_event (self, attr_name, str_data, data) -> None

push_change_event (self, attr_name, data, time_stamp, quality, dim_x = 1, dim_y = 0) -> None

push_change_event (self, attr_name, str_data, data, time_stamp, quality) -> None

Push a change event for the given attribute name. The event is pushed to the notification daemon.

Parameters

attr_name

(str) attribute name

data

the data to be sent as attribute event data. Data must be compatible with the attribute type and format. for SPECTRUM and IMAGE attributes, data can be any type of sequence of elements compatible with the attribute type

str_data

(str) special variation for DevEncoded data type. In this case ‘data’ must be a str or an object with the buffer interface.

except

(DevFailed) Instead of data, you may want to send an exception.

dim_x

(int) the attribute x length. Default value is 1

dim_y

(int) the attribute y length. Default value is 0

time_stamp

(double) the time stamp

quality

(AttrQuality) the attribute quality factor

Throws

DevFailed If the attribute data type is not coherent.

push_data_ready_event(self, attr_name, counter=0) → None ¶

Push a data ready event for the given attribute name. The event is pushed to the notification daemon.

The method needs only the attribue name and an optional “counter” which will be passed unchanged within the event

Parameters

attr_name: (str) attribute name
counter: (int) the user counter

Return

None

Throws

DevFailed If the attribute name is unknown.

push_event(self, attr_name, filt_names, filt_vals) → None ¶

push_event (self, attr_name, filt_names, filt_vals, data, dim_x = 1, dim_y = 0) -> None

push_event (self, attr_name, filt_names, filt_vals, str_data, data) -> None

push_event (self, attr_name, filt_names, filt_vals, data, time_stamp, quality, dim_x = 1, dim_y = 0) -> None

push_event (self, attr_name, filt_names, filt_vals, str_data, data, time_stamp, quality) -> None

Push a user event for the given attribute name. The event is pushed to the notification daemon.

Parameters

attr_name

(str) attribute name

filt_names

(sequence<str>) the filterable fields name

filt_vals

(sequence<double>) the filterable fields value

data

the data to be sent as attribute event data. Data must be compatible with the attribute type and format. for SPECTRUM and IMAGE attributes, data can be any type of sequence of elements compatible with the attribute type

str_data

(str) special variation for DevEncoded data type. In this case ‘data’ must be a str or an object with the buffer interface.

dim_x

(int) the attribute x length. Default value is 1

dim_y

(int) the attribute y length. Default value is 0

time_stamp

(double) the time stamp

quality

(AttrQuality) the attribute quality factor

Throws

DevFailed If the attribute data type is not coherent.

push_pipe_event(self, blob) → None ¶

Push an pipe event.

Parameters: the blob which pipe event will be send.
Return: None

New in PyTango 9.2.2

register_signal(self, signo) → None ¶

Register a signal. Register this device as device to be informed when signal signo is sent to to the device server process

Parameters

signo: (int) signal identifier

Return

None

remove_attribute(self, attr_name) → None ¶

Remove one attribute from the device attribute list.

Parameters

attr_name: (str) attribute name

Return

None

Throws

DevFailed

remove_command(self, attr_name) → None ¶

Remove one command from the device command list.

Parameters

cmd_name: (str) command name to be removed from the list
free_it: Boolean set to true if the command object must be freed.
clean_db: Clean command related information (included polling info if the command is polled) from database.

Return

None

Throws

DevFailed

classmethod run_server(args=None, **kwargs)¶

Run the class as a device server. It is based on the tango.server.run method.

The difference is that the device class and server name are automatically given.

Args:

args (iterable): args as given in the tango.server.run method: without the server name. If None, the sys.argv list is used
kwargs: the other keywords argument are as given: in the tango.server.run method.

set_archive_event(self, attr_name, implemented, detect=True) → None ¶

Set an implemented flag for the attribute to indicate that the server fires archive events manually, without the polling to be started. If the detect parameter is set to true, the criteria specified for the archive event are verified and the event is only pushed if they are fullfilled. If detect is set to false the event is fired without any value checking!

Parameters

attr_name: (str) attribute name
implemented: (bool) True when the server fires change events manually.
detect: (bool) Triggers the verification of the change event properties when set to true. Default value is true.

Return

None

set_change_event(self, attr_name, implemented, detect=True) → None ¶

Set an implemented flag for the attribute to indicate that the server fires change events manually, without the polling to be started. If the detect parameter is set to true, the criteria specified for the change event are verified and the event is only pushed if they are fullfilled. If detect is set to false the event is fired without any value checking!

Parameters

attr_name: (str) attribute name
implemented: (bool) True when the server fires change events manually.
detect: (bool) Triggers the verification of the change event properties when set to true. Default value is true.

Return

None

set_state(self, new_state) → None ¶

Set device state.

Parameters

new_state: (DevState) the new device state

Return

None

set_status(self, new_status) → None ¶

Set device status.

Parameters

new_status: (str) the new device status

Return

None

signal_handler(self, signo) → None ¶

Signal handler. The method executed when the signal arrived in the device server process. This method is defined as virtual and then, can be redefined following device needs.

Parameters

signo: (int) the signal number

Return

None

Throws

DevFailed This method does not throw exception but a redefined method can.

stop_polling(self) → None ¶

stop_polling (self, with_db_upd) -> None

Stop all polling for a device. if the device is polled, call this method before deleting it.

Parameters

with_db_upd: (bool) Is it necessary to update db ?

Return

None

New in PyTango 7.1.2

unregister_signal(self, signo) → None ¶

Unregister a signal. Unregister this device as device to be informed when signal signo is sent to to the device server process

Parameters

signo: (int) signal identifier

Return

None

warn_stream(self, msg, *args) → None ¶

Sends the given message to the tango warn stream.

Since PyTango 7.1.3, the same can be achieved with:
print(msg, file=self.log_warn)

Parameters

msg: (str) the message to be sent to the warn stream

Return

None

write_attr_hardware(self) → None ¶

Write the hardware for attributes. Default method to implement an action necessary on a device to write the hardware involved in a a write attribute. This method must be redefined in sub-classes in order to support writable attribute

Parameters

attr_list(sequence<int>) list of indices in the device object attribute vector: of an attribute to be written.

Return

None

Throws

DevFailed This method does not throw exception but a redefined method can.

class tango.server.attribute(fget=None, **kwargs)¶

Declares a new tango attribute in a Device. To be used like the python native property function. For example, to declare a scalar, tango.DevDouble, read-only attribute called voltage in a PowerSupply Device do:

class PowerSupply(Device):

    voltage = attribute()

    def read_voltage(self):
        return 999.999

The same can be achieved with:

class PowerSupply(Device):

    @attribute
    def voltage(self):
        return 999.999

It receives multiple keyword arguments.

parameter	type	default value	description
name	`str`	class member name	alternative attribute name
dtype	`object`	`DevDouble`	data type (see Data type equivalence)
dformat	`AttrDataFormat`	`SCALAR`	data format
max_dim_x	`int`	1	maximum size for x dimension (ignored for SCALAR format)
max_dim_y	`int`	0	maximum size for y dimension (ignored for SCALAR and SPECTRUM formats)
display_level	`DispLevel`	`OPERATOR`	display level
polling_period	`int`	-1	polling period
memorized	`bool`	False	attribute should or not be memorized
hw_memorized	`bool`	False	write method should be called at startup when restoring memorize value (dangerous!)
access	`AttrWriteType`	`READ`	read only/ read write / write only access
fget (or fread)	`str` or `callable`	‘read_<attr_name>’	read method name or method object
fset (or fwrite)	`str` or `callable`	‘write_<attr_name>’	write method name or method object
fisallowed	`str` or `callable`	‘is_<attr_name>_allowed’	is allowed method name or method object
label	`str`	‘<attr_name>’	attribute label
enum_labels	sequence	None	the list of enumeration labels (enum data type)
doc (or description)	`str`	‘’	attribute description
unit	`str`	‘’	physical units the attribute value is in
standard_unit	`str`	‘’	physical standard unit
display_unit	`str`	‘’	physical display unit (hint for clients)
format	`str`	‘6.2f’	attribute representation format
min_value	`str`	None	minimum allowed value
max_value	`str`	None	maximum allowed value
min_alarm	`str`	None	minimum value to trigger attribute alarm
max_alarm	`str`	None	maximum value to trigger attribute alarm
min_warning	`str`	None	minimum value to trigger attribute warning
max_warning	`str`	None	maximum value to trigger attribute warning
delta_val	`str`	None
delta_t	`str`	None
abs_change	`str`	None	minimum value change between events that causes event filter to send the event
rel_change	`str`	None	minimum relative change between events that causes event filter to send the event (%)
period	`str`	None
archive_abs_change	`str`	None
archive_rel_change	`str`	None
archive_period	`str`	None
green_mode	`GreenMode`	None	green mode for read and write. None means use server green mode.
read_green_mode	`GreenMode`	None	green mode for read. None means use server green mode.
write_green_mode	`GreenMode`	None	green mode for write. None means use server green mode.
forwarded	`bool`	False	the attribute should be forwarded if True

Note

avoid using dformat parameter. If you need a SPECTRUM attribute of say, boolean type, use instead dtype=(bool,).

Example of a integer writable attribute with a customized label, unit and description:

class PowerSupply(Device):

    current = attribute(label="Current", unit="mA", dtype=int,
                        access=AttrWriteType.READ_WRITE,
                        doc="the power supply current")

    def init_device(self):
        Device.init_device(self)
        self._current = -1

    def read_current(self):
        return self._current

    def write_current(self, current):
        self._current = current

The same, but using attribute as a decorator:

class PowerSupply(Device):

    def init_device(self):
        Device.init_device(self)
        self._current = -1

    @attribute(label="Current", unit="mA", dtype=int)
    def current(self):
        """the power supply current"""
        return 999.999

    @current.write
    def current(self, current):
        self._current = current

In this second format, defining the write implicitly sets the attribute access to READ_WRITE.

New in version 8.1.7: added green_mode, read_green_mode and write_green_mode options

tango.server.command(f=None, dtype_in=None, dformat_in=None, doc_in='', dtype_out=None, dformat_out=None, doc_out='', display_level=None, polling_period=None, green_mode=None)¶

Declares a new tango command in a Device. To be used like a decorator in the methods you want to declare as tango commands. The following example declares commands:

void TurnOn(void)

void Ramp(DevDouble current)

DevBool Pressurize(DevDouble pressure)

class PowerSupply(Device):

    @command
    def TurnOn(self):
        self.info_stream('Turning on the power supply')

    @command(dtype_in=float)
    def Ramp(self, current):
        self.info_stream('Ramping on %f...' % current)

    @command(dtype_in=float, doc_in='the pressure to be set',
             dtype_out=bool, doc_out='True if it worked, False otherwise')
    def Pressurize(self, pressure):
        self.info_stream('Pressurizing to %f...' % pressure)
        return True

Note

avoid using dformat parameter. If you need a SPECTRUM attribute of say, boolean type, use instead dtype=(bool,).

Parameters

dtype_in – a data type describing the type of parameter. Default is None meaning no parameter.
dformat_in (AttrDataFormat) – parameter data format. Default is None.
doc_in (str) – parameter documentation
dtype_out – a data type describing the type of return value. Default is None meaning no return value.
dformat_out (AttrDataFormat) – return value data format. Default is None.
doc_out (str) – return value documentation
display_level (DispLevel) – display level for the command (optional)
polling_period (int) – polling period in milliseconds (optional)
green_mode – set green mode on this specific command. Default value is None meaning use the server green mode. Set it to GreenMode.Synchronous to force a non green command in a green server.

New in version 8.1.7: added green_mode option

New in version 9.2.0: added display_level and polling_period optional argument

class tango.server.pipe(fget=None, **kwargs)¶

Declares a new tango pipe in a Device. To be used like the python native property function.

Checkout the pipe data types to see what you should return on a pipe read request and what to expect as argument on a pipe write request.

For example, to declare a read-only pipe called ROI (for Region Of Interest), in a Detector Device do:

class Detector(Device):

    ROI = pipe()

    def read_ROI(self):
        return ('ROI', ({'name': 'x', 'value': 0},
                        {'name': 'y', 'value': 10},
                        {'name': 'width', 'value': 100},
                        {'name': 'height', 'value': 200}))

The same can be achieved with (also showing that a dict can be used to pass blob data):

class Detector(Device):

    @pipe
    def ROI(self):
        return 'ROI', dict(x=0, y=10, width=100, height=200)

It receives multiple keyword arguments.

parameter	type	default value	description
name	`str`	class member name	alternative pipe name
display_level	`DispLevel`	`OPERATOR`	display level
access	`PipeWriteType`	`READ`	read only/ read write access
fget (or fread)	`str` or `callable`	‘read_<pipe_name>’	read method name or method object
fset (or fwrite)	`str` or `callable`	‘write_<pipe_name>’	write method name or method object
fisallowed	`str` or `callable`	‘is_<pipe_name>_allowed’	is allowed method name or method object
label	`str`	‘<pipe_name>’	pipe label
doc (or description)	`str`	‘’	pipe description
green_mode	`GreenMode`	None	green mode for read and write. None means use server green mode.
read_green_mode	`GreenMode`	None	green mode for read. None means use server green mode.
write_green_mode	`GreenMode`	None	green mode for write. None means use server green mode.

The same example with a read-write ROI, a customized label and description:

class Detector(Device):

    ROI = pipe(label='Region Of Interest', doc='The active region of interest',
               access=PipeWriteType.PIPE_READ_WRITE)

    def init_device(self):
        Device.init_device(self)
        self.__roi = 'ROI', dict(x=0, y=10, width=100, height=200)

    def read_ROI(self):
        return self.__roi

    def write_ROI(self, roi):
        self.__roi = roi

The same, but using pipe as a decorator:

class Detector(Device):

    def init_device(self):
        Device.init_device(self)
        self.__roi = 'ROI', dict(x=0, y=10, width=100, height=200)

    @pipe(label="Region Of Interest")
    def ROI(self):
        """The active region of interest"""
        return self.__roi

    @ROI.write
    def ROI(self, roi):
        self.__roi = roi

In this second format, defining the write / setter implicitly sets the pipe access to READ_WRITE.

New in version 9.2.0.

class tango.server.device_property(dtype, doc='', mandatory=False, default_value=None, update_db=False)¶

Declares a new tango device property in a Device. To be used like the python native property function. For example, to declare a scalar, tango.DevString, device property called host in a PowerSupply Device do:

from tango.server import Device, DeviceMeta
from tango.server import device_property

class PowerSupply(Device):

    host = device_property(dtype=str)
    port = device_property(dtype=int, mandatory=True)

Parameters

dtype – Data type (see Data types)
doc – property documentation (optional)
(optional (mandatory) – default is False)
default_value – default value for the property (optional)
update_db (bool) – tells if set value should write the value to database. [default: False]

New in version 8.1.7: added update_db option

class tango.server.class_property(dtype, doc='', default_value=None, update_db=False)¶

Declares a new tango class property in a Device. To be used like the python native property function. For example, to declare a scalar, tango.DevString, class property called port in a PowerSupply Device do:

from tango.server import Device, DeviceMeta
from tango.server import class_property

class PowerSupply(Device):

    port = class_property(dtype=int, default_value=9788)

Parameters

dtype – Data type (see Data types)
doc – property documentation (optional)
default_value – default value for the property (optional)
update_db (bool) – tells if set value should write the value to database. [default: False]

New in version 8.1.7: added update_db option

tango.server.run(classes, args=None, msg_stream=<_io.TextIOWrapper name='<stdout>' mode='w' encoding='utf-8'>, verbose=False, util=None, event_loop=None, post_init_callback=None, green_mode=None, raises=False)¶

Provides a simple way to run a tango server. It handles exceptions by writting a message to the msg_stream.

The classes parameter can be either a sequence of:

: class:~tango.server.Device or
a sequence of two elements DeviceClass, DeviceImpl or
a sequence of three elements DeviceClass, DeviceImpl, tango class name (str)

or a dictionary where:

key is the tango class name
value is either:
- a : class:~tango.server.Device class or
- a sequence of two elements DeviceClass, DeviceImpl or
- a sequence of three elements DeviceClass, DeviceImpl, tango class name (str)

The optional post_init_callback can be a callable (without arguments) or a tuple where the first element is the callable, the second is a list of arguments (optional) and the third is a dictionary of keyword arguments (also optional).

Note

the order of registration of tango classes defines the order tango uses to initialize the corresponding devices. if using a dictionary as argument for classes be aware that the order of registration becomes arbitrary. If you need a predefined order use a sequence or an OrderedDict.

Example 1: registering and running a PowerSupply inheriting from Device:

from tango.server import Device, DeviceMeta, run

class PowerSupply(Device):
    pass

run((PowerSupply,))

Example 2: registering and running a MyServer defined by tango classes MyServerClass and MyServer:

from tango import Device_4Impl, DeviceClass
from tango.server import run

class MyServer(Device_4Impl):
    pass

class MyServerClass(DeviceClass):
    pass

run({'MyServer': (MyServerClass, MyServer)})

Example 3: registering and running a MyServer defined by tango classes MyServerClass and MyServer:

from tango import Device_4Impl, DeviceClass
from tango.server import Device, DeviceMeta, run

class PowerSupply(Device):
    pass

class MyServer(Device_4Impl):
    pass

class MyServerClass(DeviceClass):
    pass

run([PowerSupply, [MyServerClass, MyServer]])
# or: run({'MyServer': (MyServerClass, MyServer)})

Parameters

classes (sequence or dict) – a sequence of Device classes or a dictionary where keyword is the tango class name and value is a sequence of Tango Device Class python class, and Tango Device python class
args (list) – list of command line arguments [default: None, meaning use sys.argv]
msg_stream – stream where to put messages [default: sys.stdout]
util (Util) – PyTango Util object [default: None meaning create a Util instance]
event_loop (callable) – event_loop callable
post_init_callback (callable or tuple (see description above)) – an optional callback that is executed between the calls Util.server_init and Util.server_run
raises (bool) – Disable error handling and propagate exceptions from the server

Returns

The Util singleton object

Return type

Util

New in version 8.1.2.

Changed in version 8.1.4: when classes argument is a sequence, the items can also be a sequence <TangoClass, TangoClassClass>[, tango class name]

Changed in version 9.2.2: raises argument has been added

tango.server.server_run(classes, args=None, msg_stream=<_io.TextIOWrapper name='<stdout>' mode='w' encoding='utf-8'>, verbose=False, util=None, event_loop=None, post_init_callback=None, green_mode=None)¶: Since PyTango 8.1.2 it is just an alias to run(). Use run() instead.

New in version 8.0.0.

Changed in version 8.0.3: Added util keyword parameter. Returns util object

Changed in version 8.1.1: Changed default msg_stream from stderr to stdout Added event_loop keyword parameter. Returns util object

Changed in version 8.1.2: Added post_init_callback keyword parameter

Deprecated since version 8.1.2: Use run() instead.