Programming Assignment (Dictionary lookup and autocomplete using Tries)

Trie

A trie is a special rooted tree. that allows us to store a set of words by sharing their common prefixes.

Example # 1

Consider the following list of five words:

The trie data structure for this list above is given below:

A trie node can be of three types:

1. Root of the trie (shown as a diamond)

2. Internal non-terminal node of a trie (white ovals)

3. Word-ending node of a trie (green ovals).

Note that a word-ending node need not be a leaf. Also, such nodes are associated with a number that represents the frequency of the corresponding word.

Each trie edge is labelled with an alphabet.

Reading the words in the dictionary.

Starting from the root (diamond node), as we traverse a path along the trie, the letters along a path form a word. Whenever we encounter a word-ending node, we can stop and read off a valid word in the dictionary.

For instance, we note the correspondence between the list of words:

and the paths in the trie.

Trie Data Structure: Specification

A trie (or a prefix tree) data structure allows us to store a set of strings from a dictionary along with a frequency count for each strings. The following operations should be supported:

Example

We will illustrate the process of constructing a trie using an example.

At the beginning the empty trie is simply a root node:

Suppose we add the strings in the following order:

1: addWord('test’)

2: addWord('testament’)

3: addword('testing’) 4: addWord('ping’)

5: addWord('pin’) 6: addWord('pink’) 7: addWord('pine’) 8: addWord('pint’)

9: addWord('testing’) 10: addWord('pinetree’)

Lookup of a Word

The operation lookupWord allows us to count the frequency of a word in a trie. If a word does not belong to a trie, then the frequency is 0.

For instance, consider the trie below:

lookupWord('test’) = 1 lookupWord('testing’) = 2 lookupWord('tesla’) = 0 lookupWord('tom’) = 0 lookupWord('pinetree’) = 1

Autocomplete

The operation autoComplete provides a list of “completions” of a given string along with their frequencies.

autoComplete('pi’) = [ ('pin’,1), ('pink’,1), ('pine’,1),('pinetree’,1),('pint’,1), ('ping’,1) ] autoComplete('tom’) = [ ] # emtpy list because no word with prefix tom belongs to the trie autoComplete('testi’) = [ ('testing’,2) ]

Trie Node

A trie node structure has the following attributes:

• isRoot A Boolean flag denoting if the node is a root.

• isWordEnd A Boolean flag denoting if the node is the word ending node.

• count A count field for frequency count for a word ending node.

• next: A dictionary that maps from each of the alphabets 'a’ - 'z’ to the child node corresponding to the alphabet.

Assignment Instructions

The downloadable zip file for the assignment has the following contents:

• trieClass.py This is a skeleton class file for data structuring a trie. You are asked to implement the functions addWord, lookupWord and autoComplete.

• trieTester.py This is the autograding code that will be used to grade your assignment.

• test1.spec to test9.spec different test cases that will insert and lookup words in the trie. You have to pass each of these tests to be successful.

Running a test

To run a test say “test1.spec”, call the python script

This will print a bunch of messages including failure messages and tell you whether the test was passed by your code or not.

Upload Instructions

You are asked to upload a zip file that contains your code with the file trieClass.py containing the class for the trie node as provided in your skeleton. Any extra files you have created for implementing the functionalities can be included. Please do not include anything else.

Uploaded filename should be called firstname_lastname.zip. Do not “submit for grading” on moodle unless your assignment is really final. Reopening submissions causes enormous problems on moodle.