T9-Dictionary

"A long, long time ago, before the touchscreen keyboards of modern phones, primitive beings called humans typed words on their flip phones using numeric keypads. "How does this work?" you wonder. "Numbers aren't letters!" Indeed, they are not; however, each numeric key was marked with a collection of letters it could represent."

The Python program simulates T-9 dictionary behavior on old mobiles and helps users to write fast SMS on mobile phones :).

As presented in the picture above to numeric keys letters are assigned:

keymap = {
    '2': 'abc',
    '3': 'def',
    '4': 'ghi',
    '5': 'jkl',
    '6': 'mno',
    '7': 'pqrs',
    '8': 'tuv',
    '9': 'wxyz',
}

Sequence without T9:

when the user edited SMS without T-9 and inserted the sequence: '226' it resulted in 'bm'. Hitting more than one the same key gave the option to go via a sequence of letters assigned to the key. So twice hit key '2' gave "a->b", so letter b was inserted into SMS, one hit '6' gave 'm' so the final word insterted in sms was: 'bm'.

Sequence with T9:

However, when the user edited SMS with T-9 and inserted the sequence '226', the dictionary was looking at all matching words in the background and presented the proposal of the most common words fitted into the key. Double hitting the same key did not anymore offer key sequence but contributed with one more letter in searching. '226' could result in the following suggestions:

• 'can'
• 'ban'
• 'cam'
• ...
• etc. List of words could be long for better users experience words were suggested based on frequencies of usage the words in certain languages. The more often word was used the higher rank on the list.

File Sources

The program uses frequencies of words gathered by google in dataset:NGram project, released under CC BY 3.0.

Algorithm:

• Recursive dictionary data structure is used for fast search
• Recursive search functions are implemented for data search