| |
- analyze(situation)
- Analyzes a situation. This includes finding LAs that fit the situation
and its utterance and then either reinforcing found LAs or
hypothesizing a new LA.
Return: None.
- chattest()
- Input words and look for a matching linguistic abstraction, then
try to find a new utterance as a reply, different from the input.
- findLAList(scoresWordsMeanings)
- Find a list of LAs that fit the meanings in the argument.
Return: list of LA hashes.
- fits(meaning, LA)
- Return True if supplied meaning fits in the
linguistic abstraction.
Converts both arguments to a flat string representation,
and iterates over the lines of the LA. When a matching line
with the meaning argument is found, a counter is increased.
If the counter reaches the number of lines in the meaning,
the meaning matched the LA. If a line is found which does
not match the LA, the counter is resetted.
Arguments:
meaning -- a meaning frame, without VAR elements
LA -- a linguistic abstraction, with VAR elements in properties or id
- frame2words(frame)
- Searches the linguistic corpus for relevant LAs given a meaning frame. Then
produces a list of 2-word utterances using these LAs.
Return: list of strings.
- frames2wordstest()
- For each situation in the corpus, show the twoword utterances that can
be generated from it with its linguistic abstraction.
- getWordorder(u, word1, word2)
- hypothesize(utterance, scoresWordsMeanings)
- Make a new LA based on the meanings in the second argument. First search for
the highest scoring meaning that has a subframe also in the list of
meanings. Then get the wordorder of the two words corresponding with these
meanings and make a new LA using these frames and the wordorder.
Return: None.
- invertassociations(associations)
- Make a dictionary with framehashes as keys and their most likely
words/scores as values. The data is taken from the associations
dictionary from the oneword stage.
- la2words(la, meaning)
- return utterances matching a LA and a meaning
- onewordtest()
- Ask for a word and show which meanings oneword associates with it.
- reinforce(LAList)
- Reinforce LAs corresponding to the list of LAhashes in the first
argument.
Return: None.
- twMain()
- Create one-word data. Pass situations in the corpus to analyze using this
data.
- uniqueutterancestest()
- similar to frames2words, but count how many utterances are
generated which are unique, as in, not found in the corpus.
- words2frame(w1, w2)
- Searches the linguistic corpus for relevant LAs given two words.
Return: list of frames.
- words2framestest()
- Input words and show a matching linguistic abstraction and meaning for
each pair of words.
|