Home       Trees       Indices       Help   
tipy
tipy :: prdct :: PredictorActivator :: Class PredictorActivator
[hide private]
[frames] | no frames]

Class PredictorActivator

source code

object --+
         |
        PredictorActivator

Query the predictors listed in the registry to get their suggestions.

This class has access to a PredictorRegistry and asks the predictors listed in this PredictorRegistry to call their predict() method, store the resulting Prediction instances, merge them into a single Prediction instance and return it.

Class Hierarchy for PredictorActivator
Class Hierarchy for PredictorActivator

Instance Methods [hide private]
 
__init__(self, config, predictorRegistry)
PredictorActivator creator.		 source code
 
pred_worker(self, predictor, queue, factor)
Worker function for the predictor predict() methods.		 source code
Prediction
predict(self, factor=1)
Build a list of every predicted words.		 source code

Inherited from object: __delattr__, __format__, __getattribute__, __hash__, __new__, __reduce__, __reduce_ex__, __repr__, __setattr__, __sizeof__, __str__, __subclasshook__

Properties [hide private]

Inherited from object: __class__

Method Details [hide private]

__init__(self, config, predictorRegistry)
(Constructor)

source code 

PredictorActivator creator.

Parameters:
  • config (drvr.Configuration) - The configuration dictionary is used in order to retrieve the PredictorActivator settings from the config file.
  • predictorRegistry (PredictorRegistry) - The class needs to access the PredictorRegistry to call their predict() method.
Overrides: object.__init__

pred_worker(self, predictor, queue, factor)

source code 

Worker function for the predictor predict() methods.

This method is used as the predictors workers target. It push the predictor's prdct.Predictor.predict method result (a Prediction instance) in a queue (which is used because it is thread-safe).

Parameters:
  • predictor (Predictor based class.) - The Predictor based class instance.
  • queue (multiprocessing.Queue) - A queue in which the result will be pushed.
  • factor (int) - A factor used to increase the number of suggestions.

predict(self, factor=1)

source code 

Build a list of every predicted words.

Call the predict() method of every predictors in the registry then merge their Prediction into a single Prediction instance.

Parameters:
  • factor (int) - A factor used to increase the number of suggestions.
Returns: Prediction
The merged Prediction instance containing every suggestions of every Prediction instances sorted in descending order according to their probabilities.

Change Log:

  • 16/06/15: The method now uses multi-processing. It concurrently runs every predictors's predict() method which allow a significant speed augmentation. The queue is used because it is thread safe. The point is that when the threads args are passed to the PredictorActivator.pred_worker(), they are packed up with pickle, shipped to the other process, where they are unpacked used. A list wouldn't be passed but would be cloned.

Note: Using multi-processing allow significant speed boost. The next benchmark have been maid runing 100 * 10 different contexts predictions:

Total time without multi-processing: 86.785 s Total time wit multi-processing: 76.513 s

To Do (0.0.9): Demonize the processes, set a timeout value. When the time runs out the unfinished workers return their results as is. This can alter the prediction qality but totaly avoid any possible "slow predictions".


   Home       Trees       Indices       Help   
tipy
Generated by Epydoc 3.0.1 on Tue Jul 14 21:07:50 2015 http://epydoc.sourceforge.net