=============================================================
hapPyTango depends on Python (< 3), gevent 1.0 (>= RC2) and Fnorb (>= 1.3).
You can download gevent from here: https://github.com/SiteSupport/gevent/downloads
And you can get Fnorb from the project SourceForge web page: http://sourceforge.net/projects/fnorb/files/latest/download
Once dependencies are satisfied: python setup.py install
hapPyTango started as a fun week-end project, to have an alternative Tango client library for Python that would be lightweight, easy to install (especially on Windows! No compilation, no C++ library needed), easy to use and Pythonic, and that would take advantage of gevent.
Communication is handled by gevent coroutines, thus I/O operations are non-blocking and any host application running the gevent loop can benefit from asynchronous communication with Tango devices. Using a dedicated input hook for IPython, for example, it becomes easy to write complex sequences involving Tango devices taking advantage of cooperative multi-tasking.
Old-style events (with notifd) does not work. Support was planned initially, but due to some weird error in CDR decoding, it has been dropped. New style events (with zmq) will be implemented in a future version when stable ZMQ library v3.2 will provide backward compatibility with ZMQ 3.1.
The code is less than 500 lines. Of course it is far from being a complete Tango client like PyTango, the official Tango client for Python.
Python 2.6.5 (r265:79063, Oct 1 2012, 22:07:21)
[GCC 4.4.3] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> from hapPyTango import client
>>> dev = client.Device("sys/tg_test/1")
>>> dir(dev)
['DevBoolean', 'DevDouble', 'DevFloat', 'DevLong', 'DevShort', 'DevString', 'DevULong', 'DevUShort', 'DevVarCharArray', 'DevVarDoubleArray', 'DevVarDoubleStringArray', 'DevVarFloatArray', 'DevVarLongArray', 'DevVarLongStringArray', 'DevVarShortArray', 'DevVarStringArray', 'DevVarULongArray', 'DevVarUShortArray', 'DevVoid', 'Init', 'State', 'Status', 'SwitchStates', '__doc__', '__init__', '__module__', '_attributes', '_boa', '_commands', '_connection_greenlet', '_db_host', '_device', '_device_impl_number', '_do_connect_event', '_ev_channel', '_keepalive_greenlet', '_nodb_flag', '_orb', '_tango_db', 'ampli', 'boolean_image', 'boolean_image_ro', 'boolean_scalar', 'boolean_spectrum', 'boolean_spectrum_ro', 'cmd_inout', 'connect', 'connect_event', 'double_image', 'double_image_ro', 'double_scalar', 'double_scalar_rww', 'double_scalar_w', 'double_spectrum', 'double_spectrum_ro', 'float_image', 'float_image_ro', 'float_scalar', 'float_spectrum', 'float_spectrum_ro', 'impl', 'init', 'initialized_event', 'long_image', 'long_image_ro', 'long_scalar', 'long_scalar_rww', 'long_scalar_w', 'long_spectrum', 'long_spectrum_ro', 'name', 'no_value', 'ping', 'read_attr', 'short_image', 'short_image_ro', 'short_scalar', 'short_scalar_ro', 'short_scalar_rww', 'short_scalar_w', 'short_spectrum', 'short_spectrum_ro', 'string_image', 'string_image_ro', 'string_scalar', 'string_spectrum', 'string_spectrum_ro', 'throw_exception', 'uchar_image', 'uchar_image_ro', 'uchar_scalar', 'uchar_spectrum', 'uchar_spectrum_ro', 'ushort_image', 'ushort_image_ro', 'ushort_scalar', 'ushort_spectrum', 'ushort_spectrum_ro', 'wait_initialized', 'wave', 'write_attr']
Attributes are read and written like Commands.
When calling a command (or reading/writing an attribute), the "wait" keyword argument allows to specify if the behaviour should be synchronous (wait=True, default) or asynchronous (wait=False). In the asynchronous case, a greenlet object is returned. Doing get() on this object to wait for completion and getting the result (for more options, see gevent documentation).
Other keyword arguments can be passed: "timeout" allows to specify a timeout in seconds (default: 3 seconds), "from_cache" specifies if values should be returned from cache or from device (default: from device). For arguments reading:
- "return_value" indicates if the attribute value should be returned, or the complete attribute_value object (default: True)
- "return_set_value" indicates if the set value has to be returned too (default: False)
It is fairly easy to add an input hook to IPython to run gevent loop, just as it is done for GUI loops for example: see the ipython.py module. When starting IPython, make sure "enable_gevent_hook" is called:
start_ipython.sh
#!/usr/bin/env python
from IPython.frontend.terminal.interactiveshell import TerminalInteractiveShell
from hapPyTango.client import Device
from hapPyTango.ipython import enable_gevent_hook
shell = TerminalInteractiveShell(user_ns={"TangoDevice": Device, "gevent": gevent})
enable_gevent_hook()
shell.mainloop()
Enjoy smooth integration of IPython, Tango and gevent :)