TML Messaging Suite  1.1.5
Python Extension
 All Data Structures Namespaces Functions Variables Groups Pages
Introduction to libTML-python

Many use cases are covered by the features of libTML. The example in this chapter is a brief overview of the most frequently used functions.

All examples in this introduction contain only code lines to explain a special aspect of the library. Important things may be left out in favor of better readability.

How to use the library

The libTML-python library provides access to the libtml-c low level functions and simplifies the usage with a set of Python objects. Two fundamental modules are available. TML (The Missing Link) provides the messaging and network functions and SIDEX (Simple Data Exchange) provides the API to access message data. After installing the library you can import TML and SIDEX into your project.

1 import tml # TML low level API
2 import tml.constants # constant values
3 import tml.common # helper functions
4 import tml.core # TMLCore()
5 import tml.profile # TMLProfile()
6 import # stream classes
7 import tml.command # TMLCmd()
9 import sidex # SIDEX low level API
10 import sidex.constants # constant values
11 import sidex.variant # SIDEX variant classes and type conversion
12 import sidex.document # SDXDocument()

The following examples are using the name space in front of library calls to show which particular library is used. Using the low level API is similar to the C API in most aspects. The error handling is different in Python. Low level functions do not return an error code but raise an exception instead. This introduction concentrates on the simplified usage of TML and SIDEX objects. The usage of low level functions is discussed more deeply in the C API documentation.

Implementing the Interface

The TML communication is peer to peer (P2P) and the Client/Server pattern is a subset. The latter is used in this example to demonstrate how to implement and access a remote interface with TML.

Create and prepare a TML core and profile

To use TML at least one TMLCore object is required. To accept incoming traffic the listener thread needs an IP address, port and one or more profiles to be published. A TML profile is the interface to refer to by a remote call.

1 tmlcore = tml.core.TMLCore()
2 tmlcore.listener_binding = ""

To identify a profile as a unique interface, a unique profile ID is required. Any string can be used, but usually a URN is selected. In this example urn:your-domain:com:pyexample01 is used. The profile is referred several times and to avoid repeating, is assigned to profileid which is subsequently used.

1 profileid = 'urn:your-domain:com:pyexample01'

The profiles property of the TMLCore allows to access the interfaces assigned to it. It is an instance of TMLProfiles. To publish a profile it is added to the core.

1 # register a command and profile
2 tmlcore.profiles.add_profile(profileid)

Add command handler functions

A command handler function needs two parameters. A TML profile and a TML command. All handlers are implemented as seperate functions in this example but if an interface is implemented as an object, methods can be used as well. A TML profile accepts methods as call handler and ignores the self parameter.

1 # return a number and a unicode string
2 def cmd42handler(tmlprofile, cmd):
3['Result']['Value'] = 42
4['Result']['Test'] = "一切是伟大的"
6 # return a number and a list of values
7 def cmd43handler(tmlprofile, cmd):
8['Result']['Value'] = 43
9['Result']['Test'] = [1, 2, 3, 'test', 3.14]
11 # return a progress while executing the command
12 def cmd44_progress_test(tmlprofile, cmd):
13  time.sleep(1)
14  for p in range(0, 100, 10):
15['Result']['Progress'] = p
16  time.sleep(0.5)
17  cmd.progress = p

To publish a command handler it is added to a profile using the register_cmd() method of the TMLProfile instance returned by the profiles property.

1 # register commands
2 tmlcore.profiles[profileid].register_cmd(42, cmd42handler)
3 tmlcore.profiles[profileid].register_cmd(43, cmd43handler)
4 tmlcore.profiles[profileid].register_cmd(44, cmd44_progress_test)

Start the Listener

Finally the listener is started to publish the interface to the network. TML uses multiple threads for parallel execution of the handler functions. Meanwhile the main thread needs to wait or can do different things. The run() function in this example is waiting for a keyboard input. In GUI programs it can easily be replaced with the application loop.

1 tmlcore.listener_enabled = True
3 run() # wait until CTRL-C is pressed

Calling the Interface

To call a particular interface a TCP/IP address and a profile ID is required.

1 profile = 'urn:your-domain:com:pyexample01'
2 port = '12345'
3 host = ''

The TMLCore is handling inbound and outbound messages. Before sending a command an instance of TMLCore is created.

1 core = tml.core.TMLCore()

Use the same TMLCore instance as long as the application is sending/receiving messages to keep the connections open. Reusing connections is much faster than starting a TCP/IP connection every time a message is sent.

Before sending a command, a TMLCmd object is created with the command ID of the remote interface function. It is sent with the call_sync() method of TMLCore. Using synchronous calls with call_sync() waits until a reply was received or an error occurs before returning. In case of an error an exception is raised.

1 # call command 42
2 cmd = tml.command.TMLCmd(42)
3 core.call_sync(host, port, profile, cmd, 5000)
4 print(['Result']['Value'])
5 print(['Result']['Test'])
7 # call command 43
8 cmd = tml.command.TMLCmd(43)
9 core.call_sync(host, port, profile, cmd, 5000)
10 print(['Result']['Value'])
11 print(['Result']['Test'])

Receiving progress replies

TML allows to send back status or progress replies to the caller. Besides providing information about the status of the command processing at the remote peer, progress replies reset the timeout passed to the call_sync() or call_async() method. In cases of long or unknown reply times, the receiver can keep the sender waiting.

The timeout reset is performed whether the progress reply is handled or not. Handling a progress reply is used to visualize the progress or any additional data.

The progress handler function can be assigned to a profile or, in this example, to the TMLCore. The object calling the progress handler is passed as the sender in the first parameter. The second parameter (cmd) is the command instance currently processed by the remote peer. The progress value is part of the command data, but it is passed for convenience as a third parameter.

1 def progress_handler(sender, cmd, progress):
2  print(progress)

The progress handler has to be assigned to the core before sending the first command. Any subsequent call with progress replies will use the handler function as well.

1 # call progress command
2 core.on_progress = progress_handler
3 cmd = tml.command.TMLCmd(44)
4 core.call_sync(host, port, profile, cmd, 5000)

If the data of the command is modified either at sender (progress handler) or receiver (command handler) side, the modifications are transferred together with the progress change. Applications can use this behavior to implement advanced features.

A script calling the commands of the interface in this example is producing the output below.

[1, 2, 3, 'test', 3.14]
progress 10
progress 20
progress 30
progress 40
progress 50
progress 60
progress 70
progress 80
progress 90

Using the data property and variants

The SIDEX API (Simple Data Exchange) provides powerful functions to organize data. Any TMLCmd instance has it's data stored in the SDXDocument available through the data property.

1 # create a SDXDocument
2 sdoc = sidex.document.SDXDocument('SDXDocTest')

An SDXDocument instance is storing data groups identified by a name (string). Each group can contain one or more values identified with keys (string). The python SDXDocument class mimics a dictionary to simplify data access.

1 sdoc['General']['int'] = 42
2 sdoc['General']['float'] = 3.14
4 sdoc['General']['str'] = u"some text äüö !^s-§$"
5 sdoc['General']['chn'] = u"一切是伟大的"
6 sdoc['General']['bool'] = True
7 sdoc['General']['None'] = None

All the above values are automatically converted into SIDEX variants and added to the document.

The SIDEX variant types can be created and assigned like python values as well. The SDXDateTime example shows how to create a variant and assign it to the document.

Python date and time values can be converted automatically like other python values. The example with SDXDateTime is chosen for demonstration only.

1 sdxval = sidex.variant.SDXDateTime("2014-01-01 12:30:00:000")
2 sdoc['General']['Date'] = sdxval

Reading data from a SDXDocument is just using the dictionary behavior and address the value with group and key names.

1 print(sdoc['General']['chn'])

All values are automatically converted into Python values. If required set the autocast property of the SDXDocument to False to return all values as SDX variants.

Add Lists and Dictionaries

Python container types like lists and dictionaries are a powerful tool to create hierarchical data structure. The SDXDocument supports lists and dictionaries as well and they can simply be added by assignment.

1 sdoc['Container']['List'] = [1, 2, 3, u"一切是伟大的", None, False, sdxval,
2  [1, 2, 3], None,
3  {'test': [1, sdxval, 3]}]

The example shows the assignment of a list with an embedded list, a mixture of values including a SDX variant datetime (sdxval) and a dictionary with a list as value.