Class ContextMonitor

ContextMonitor creator.

Parameters:

• config (drvr.Configuration) - It is used to retrieve the ContextMonitor settings from the configuration file.
• predictorRegistry (PredictorRegistry) - It is used to access the predictors's learn() methods. Also, the ContextMonitor is used by the predictors to access the input buffers.
• callback (Callback) - As the callback hold the input buffers and the ContextMonitor operate on these buffers, it is used to access the input buffers from inside the ContextMonitor.

Overrides: object.__init__

Check if a context change occure.

Returns: bool

Return True or False weither a context change occure.

Check if context changes occure and learn what need to be learnt.

This method is called by Driver.predict() after the predictions have been computed. It check if a context change occure in the input buffers and if so, it learn the words that need to be learnt if the predictor's learning mode is ON. Finaly, it update the monitored scope.

Learn n-grams from the input buffers.

Trigger the learn() method of each predictor of the registry. This method use the input buffers to create n-grams and add them to the predictors's databases or memory so that the program learn from the user input.

Parameters:

• string (str) - The string to learn.

Return the token just before the cursor.

Returns: str

The token just before the cursor or an empty string if there is none.

Return the token just after the cursor.

Returns: str

The token just after the cursor or the empty string if there is none.

Return the token at a given index in the left input buffer.

Parameters:

• index (int) - The index of the token to retrieve in the left input buffer.

Returns: str

The token at index 'index' in the left input buffer or an empty string if the token dosen't exists.

Return the token at a given index in the right input buffer.

Parameters:

• index (int) - The index of the token to retrieve in the right input buffer.

Returns: str

The token at index 'index' in the right input buffer or an empty string if the token dosen't exists.

Return the token just before the change token (if any).

This method is called in some predictors's learn() method. It retrieve the token that appear just before the change token and has already been learnt before (or should have). The previous token is used to fill the n-grams.

Parameters:

• index (int) - Index of the previous token.
• change (str) - The change token.

Returns: str

The token just before the change token or an empty string if there is none.

Use the callback to get the value of the left buffer.

Returns: str

The left input buffer.

Use the callback to get the value of the right buffer.

Returns: str

The right input buffer.

Compute the completion string given a suggested word.

This method compute and return the completion string using the token just before the cursor (prefix) and the suggested word (suggestion). The suggestion should be the word that the user choose from the suggested words list.

For instance, if the prefix is:

   "wor"

And the suggestion is:

   "world"

Then this method will compute the completion:

   "ld"

If the character before the cursor is a blankspace or a separator then the prefix should be empty:

   ""

Then if the suggestion is:

   "guilty"

This method will compute the completion:

   "guilty"

If the suggestion and the prefix don't match then False is returned. This should never happen as suggestions completing an input word should always match it. Still, I prefer to check it at the cost of some lower() and startswith() calls.

Parameters:

• suggestion (str) - The suggested word from which to compute the completion.