Python data structure for thesaurus

Question

I need to have synonyms that I define for about 100 words of my of my choice. For testing I am adding the entries manually:

t = {}
t.update({'Strong':['Strong', 'Able', 'Active', 'Big',
                    'Energy', 'Firm',
                    'Force', 'Heavy', 'Robust', 'Secure',
                    'Solid', 'Stable', 'Steady',
                    'Tough', 'Vigor', 'Might',
                    'Rugged', 'Sound']})

t.update({'Fast':['Fast', 'Agile', 'Brisk', 'Hot', 'Quick',
              'Rapid', 'Swift', 'Accel', 'Active',
              'Dash', 'Flash', 'Fly', 'Race', 'Snap',
              'Wing', 'Streak', 'Time', 'Chop', 'Jiffy',
              'Split', 'Bat', 'Crazy', 'Double', 'Scream',
              'Sonic', 'Super', 'Ball', 'Speed']})

So I am creating an empty dictionary, and then taking words like "Strong" and "Fast" and mapping it to synonyms (which I need to be able to choose).

Since I need only 100 different word mappings is this a reasonable approach? Or is there a better way to implement this?

I am also looking at using NLTK and the wordnet module. However, this module takes awhile to run and it seems I have no way of adding synonyms like I need.

Answer 1

I could organize your thesaurus in a graph fashion. First of all, you keep all the words in a dictionary word -> key and then you make a linked-list graph, since it will be sparse.

w = {}
w = {'Fast': 0, 'Strong': 1, 'Able': 2, 'Active': 3, 'Big': 4, ...}

t = {0: [1, 2, 3, ...], ...}

It would scale better for large data sets, since ints use less memory than strings.

Answer 2

In an actual thesaurus, individual words may belong to multiple sets of synonyms. For example, fast as in quick might be one list while fast as in secure might be in another.

I would map each word to a list of "sense groups," and then each sense group would map to a list of words.