NLP.WordNet.Prims

WordNet

NLP.WordNet.Prims

Description

NLP.WordNet.Prims provides primitive operations over the word net database. The general scheme of things is to call initializeWordNet to get a WordNetEnv. Once you have this, you can start querying. A query usually looks like (suppose we want Dog as a Noun:

getIndexString on Dog. This will give us back a cannonicalized string, in this case, still dog. We then use indexLookup to get back an index for this string. Then, we call indexToSenseKey to with the index and the sense number (the Index contains the number of senses) to get back a SenseKey. We finally call getSynsetForSense on the sense key to get back a Synset.

We can continue to query like this or we can use the offsets provided in the various fields of the Synset to query directly on an offset. Given an offset and a part of speech, we can use readSynset directly to get a synset (instead of going through all this business with indices and sensekeys.

Synopsis

initializeWordNet :: IO WordNetEnv

initializeWordNetWithOptions :: Maybe FilePath -> Maybe (String -> Exception -> IO ()) -> IO WordNetEnv

closeWordNet :: WordNetEnv -> IO ()

getIndexString :: WordNetEnv -> String -> POS -> IO (Maybe String)

getSynsetForSense :: WordNetEnv -> SenseKey -> IO (Maybe Synset)

readSynset :: WordNetEnv -> POS -> Offset -> String -> IO Synset

indexToSenseKey :: WordNetEnv -> Index -> Int -> IO (Maybe SenseKey)

indexLookup :: WordNetEnv -> String -> POS -> IO (Maybe Index)

Documentation

initializeWordNet :: IO WordNetEnv

initializeWordNet looks for the word net data files in the default directories, starting with environment variables WNSEARCHDIR and WNHOME, and then falling back to defaultPath as defined in NLP.WordNet.Consts.

initializeWordNetWithOptions :: Maybe FilePath -> Maybe (String -> Exception -> IO ()) -> IO WordNetEnv

initializeWordNetWithOptions looks for the word net data files in the specified directory. Use this if wordnet is installed in a non-standard place on your machine and you don't have the appropriate env vars set up.

closeWordNet :: WordNetEnv -> IO ()

closeWordNet is not strictly necessary. However, a WordNetEnv tends to hog a few Handles, so if you run out of Handles and won't be using your WordNetEnv for a while, you can close it and always create a new one later.

getIndexString :: WordNetEnv -> String -> POS -> IO (Maybe String)

getIndexString takes a string and a part of speech and tries to find that string (or something like it) in the database. It is essentially a cannonicalization routine and should be used before querying the database, to ensure that your string is in the right form.

getSynsetForSense :: WordNetEnv -> SenseKey -> IO (Maybe Synset)

getSynsetForSense takes a sensekey and finds the appropriate Synset. SenseKeys can be built using indexToSenseKey.

readSynset :: WordNetEnv -> POS -> Offset -> String -> IO Synset

readSynset takes a part of speech, and an offset (the offset can be found in another Synset) and (perhaps) a word we're looking for (this is optional) and will return its Synset.

indexToSenseKey :: WordNetEnv -> Index -> Int -> IO (Maybe SenseKey)

indexToSenseKey takes an Index (as returned by, ex., indexLookup) and a sense number and returns a SenseKey for that sense.

indexLookup :: WordNetEnv -> String -> POS -> IO (Maybe Index)

Produced by Haddock version 0.5