10.3. Algorithms and Widgets#

itom algorithm plugins can contain two different types of code:

Most algorithm plugins consist of one or multiple algorithms (also denoted as filters) that can both be called from Python scripts as well as from any other itom plugin.

To call an algorithm with the exemplary name myAlgo from Python, use one of the two following possibilities:
- call itom.filter("myAlgo", *args, **kwds) for a generic call of any algorithm.
- use the wrapper method myAlgo from the submodule algorithms of itom:
```
import itom
# args and kwds are the mandatory (or optional) parameters of the
# specific algorithm 'myAlgo'. The returned value is either None,
# a single value or a tuple of values.
result = itom.algorithms.myAlgo(*args, **kwds)
```
  This 2nd possibility is available from itom 5.0.0 on and allows using auto completion and calltips with a context sensitive help for the specific algorithm. The set of available methods in the itom.algorithms submodule depends on the loaded algorithm plugins of your itom application.
Algorithm plugins can also contain custom user interfaces. The create and show such an user interface, call the method createNewPluginWidget() or createNewPluginWidget2(). To get an online help, use widgetHelp().

For more information, see also Getting started with algorithm plugins (denoted as filters) or in the demo file demoGaussianSpotCentroidDetection.py.

itom.filter(name, *args, _observer=None, **kwds) → Any#

Invokes a filter (or algorithm) function from an algorithm-plugin.

This function is used to invoke itom filter-functions or algorithms, declared within itom-algorithm plugins. The parameters (arguments) depends on the specific filter function. Call filterHelp() to get a list of available filter functions.

Pass all mandatory or optional arguments of the filter as positional or keyword-based parameters. Some filters, that implement the additional observer interface, can accept another progressObserver object, that allows monitoring the progress of the filter and / or interrupting the execution. If such an observer is given, you have to pass it as keyword-based argument _observer!.

During the execution of the filter, the python GIL (general interpreter lock) is released (e.g. for further asynchronous processes).

Instead of this generic function to call any itom algorithm from an algorithm plugin, these algorithms are also available via the direct algorithm wrapper methods in the submodule itom.algorithms (see Algorithms and Widgets).

Parameters:

namestr: The name of the filter
*argsAny: positional arguments for the specific filter-method
_observerprogressObserver, optional: if the called filter implements the extended interface with progress and status information, an optional progressObserver object can be given (only as keyword-based parameter) which is then used as observer for the current progress of the filter execution. It is then also possible to interrupt the execution earlier (depending on the implementation of the filter). The observer object is reset before passed to the called filter function (using the slot reset()).
**kwdsAny: keyword-based arguments for the specific filter-method. The argument name _observer is reserved for special use.

Returns:

outobj: The returned values depend on the definition of each filter. In general it is a tuple of all output parameters that are defined by the filter function.