class Execute(object): def __init__(self): self.depnn = DepNN() self.depnn.extendLookUp('../data/depcheck2') self.depnn.load_wordvector('../data/word-vec.bin') self.deptree = DependencyTree() def build_model(self): self.depnn.build_AutoEncoder(200, '../data/compilation.bin') def load_model(self): self.depnn.load_AutoEncoder('../data/compilation.bin') def train_sentence(self, lines): w = self.deptree.read_sent(lines[0]) y = self.deptree.read_dependency(lines[1]) #self.depnn.saveRAE('../data/weights.bin') #self.depnn.loadRAE('../data/weights.bin') print ' '.join(x[0] for x in w) self.depnn.build_DepRNN_Tree(y) def dump_weights(self, db_file): self.depnn.saveRAE(db_file) def load_weights(self, db_file): self.depnn.loadRAE(db_file)
def main(): print 'ok.' lines = ('[(S (NP (DT Those) (NN space)) (VP (VBZ walks) (SBAR (S (VP (VBP are) (S (VP (TO to) (VP (VB be) (VP (VBN used) (PP (IN for) (S (VP (VBG preparing) (NP (NP (DT the) (NNP ISS)) (PP (IN for) (NP (NP (DT the) (VBN planned) (NN docking)) (NP (JJ next) (NN year)) (PP (IN of) (NP (DT the) (JJ new) (NNP European) (NNP ATV) (NN space) (NN cargo) (NN vessel))))))))))))))))) (. .))]','[det(space-2, Those-1), nsubj(walks-3, space-2), root(ROOT-0, walks-3), ccomp(walks-3, are-4), aux(used-7, to-5), auxpass(used-7, be-6), xcomp(are-4, used-7), prepc_for(used-7,preparing-9), det(ISS-11, the-10), dobj(preparing-9, ISS-11), det(docking-15, the-13), amod(docking-15, planned-14), prep_for(ISS-11, docking-15), amod(year-17, next-16), dep(docking-15, year-17), det(vessel-25, the-19), amod(vessel-25, new-20), nn(vessel-25, European-21), nn(vessel-25, ATV-22), nn(vessel-25, space-23), nn(vessel-25, cargo-24), prep_of(docking-15, vessel-25)]') D = DependencyTree() w = D.read_sent(lines[0]) print ' '.join([x[0]+'/'+x[1] for x in w]) y = D.read_dependency(lines[1]) for d in y: print d pass
def depCheck(): path = '/home/gujt/work/Stanford_Sparser/work/Aquant2-ctree/0/' D = DependencyTree() max = 0 for dir in os.walk(path): for file in dir[2]: print 'read.',file f = open(path+file) line = f.readline() while line: line = f.readline() try: y = D.read_dependency(line) except: print line line = f.readline() continue p = [h[0] for h in y] t = {} for pi in p: if pi not in t: t[pi] = 1 else: t[pi] +=1 t = sorted(t.iteritems(), key =lambda b:b[1], reverse=True) #print t try: if t[0][1] > max: max = t[0][1] print max except: print t #for yi in y: # if yi[4] not in depDict: # depDict[yi[4]] = 1 # else: # depDict[yi[4]] += 1 line = f.readline() f.close() break print max
def depCheck(): D = DependencyTree() max = 0 f = open('test.txt') for line in f.readlines(): try: y = D.read_dependency(line) except: print line line = f.readline() continue p = [h[0] for h in y]#p:parent t = {} for pi in p: if pi not in t: t[pi] = 1 else: t[pi] +=1 t = sorted(t.iteritems(), key =lambda b:b[1], reverse=True) #print t try: if t[0][1] > max: max = t[0][1] except: print t #for yi in y: # if yi[4] not in depDict: # depDict[yi[4]] = 1 # else: # depDict[yi[4]] += 1 line = f.readline() f.close() print 'max:'+str(max)
def __init__(self): self.depnn = DepNN() self.depnn.extendLookUp('../data/depcheck2') self.depnn.load_wordvector('../data/word-vec.bin') self.deptree = DependencyTree()