Class MainWindow

MainWindow creator.

It allow the user to:

• Type text and get predictive suggestions.
• Directly complete input words or insert predicted words in the input.
• See the miners informations.
• Execute mining operations using the defined miners.
• Delete miners database.
• Modify almost every options of the configuration file using the settings window.

Parameters:

• driver (drvr.Driver) - The Driver instance wich contains everything needed for word prediction.

Create a list of every predictors name which use a given database.

Returns: list

The list of the predictors using the database file 'self.dbPath'.

Callback called when the user select a miner in the combo box.

When a miner is selected in the miner selection combo box some settings of the miner have to be retrieved from the configuration in order to display them to the user using the set_miner_widgets_values() method. As every miners store their results (n-grams) in a database the database path and the n-gram size are retrieved from the configuration and stored into the "dbPath" and "nGramSize" instance variables.

Parameters:

• miner (int) - The index of the miner in the combo box.

Update the widgets (labels and list) displaying infos on the miner.

When a miner is selected in the miner selection combo box, its config settings are retrieve from the configuration and the instance variables "dbPath", "nGramSize" and "predsUsingDb" are set. This method modify the value of the widgets displaying the informations about the miner with the above-mentioned variables.

Update the label showing the mining operation and the progress bar.

This method is a callback which is called from the miners. It update the text of the label displaying the minig operation and the value of the progress bar displaying the operation progression.

Parameters:

• perc (float or int) - Computed by the miners calling the callback, it indicate the operation progression and is usually a float but it must be converted to int as the progress bar only show integers.
• text (str) - The text to display in the mining operation label.

Callback called when the user press the "Delete database" button.

The method first identify the current miner selected in the combo box using its index then create an instance of this miner and called its rm_db() method to effectively carry out the database suppression operation. Some miners modify the configuration so the config file needs to be rewrite after the operation. Here is a short description of the miners database suppression operation, please refer to their rm_db() method docstring for more informations:

• CorpusMiner: Remove the database file.
• FbMiner: Remove the database file and set the "last_update" config option to the minimum value so the config have to be writen afterward.

Callback called when the user press the "MINE!" button.

The method first identify the current miner selected in the combo box using its index then create an instance of this miner and called its mine() method to effectively carry out the mining operation. Some miners modify the configuration so the config file needs to be rewrite after the operation. Here is a short description of the miners mining operation, please refer to their mine() method docstring for more informations:

• CorpusMiner: Mine a text corpus (i.e. a set of text files) by extracting n-grams from the files and inserting them into a database.
• FbMiner: Mine a facebook user wall (only text messages from posts) by extracting n-grams from the posts and inserting them into a database. This miner modifies the "last_update" option of its config section so the config have to be writen afterward.
• TwiterMiner: Not implemented yet but it will be very similar to FbMiner.

Update the input buffers and compute the suggestion.

This method is called whenever one or more character(s) is/are added to or removed from the input text. It update the input left and right buffers according to the text change and then generate the suggestions for the new input context. There is three king of input text change:

• Characters have been appened: -> Characters are added to the left input buffer (via callback.update()).
• Part of the characters have been removed: -> Characters are removed to the left input buffer (via callback.update()). It simulates a backspace input.
• Every characters have been removed (while new ones have been added): -> Every characters of the left input buffer are removed (via callback.update()). New characters, if any, are added to the left input buffer (via callback.update()).

Compute and show suggestions.

Request the PredictorActivator to compute the suggestions and show them in the suggestion list.

Callback called when the input text cursor position change.

There's three possible moves types:

• The cursor moved because one or more character(s) have been added to or removed from the input text: -> The cursor position and the input text length are saved and the suggestions are updated according to the current input.
• The cursor moved because the user pressed the left arrow key or clicked somewhere in the text, on the left of the previous cursor position: -> In this case the input text has not change we call the Driver callback to modify the input left and right buffer so that the left buffer now contains what's on the left of the cursor and the right buffer now contains what's on the right of the cursor. The suggestions are then updated according to the current cursor position.
• The cursor moved because the user pressed the right arrow key or clicked somewhere in the text, on the right of the previous cursor position: -> In this case the input text has not change we call the Driver callback to modify the input left and right buffer so that the left buffer now contains what's on the left of the cursor and the right buffer now contains what's on the right of the cursor. The suggestions are then updated according to the current cursor position.

Complete the input with the selected suggestion.

When the user double click on a word of the list the word is used to complete the input text. There is two kind of completion depending on the input cursor position:

• The cursor is at the end of the text: -> The completion for the last token is computed and append to the input text. A space is then added allowing the user to save a key and the program to immediatly suggest the next word prediction.
• The cursor is not at the end of the text: -> The completion for the token in which the cursor is is computed and the end of the token is replaced by the completion.

During this method the input text widget signals are disconnected which means that no suggestions will be compute during the completion process. This is important because this method will modify the input text but, of course, there is no need to generate any suggestions as the user isn't typing anything during the process.

Parameters:

• item (QtWidgets.QListWidgetItem) - The selected item (word) in the suggestion list.