import sys
import fnextracter
args = sys.argv
path = args[1]
IDs = fnextracter.ExtractID(path)
Calls = fnextracter.ExtractVerb(path)
print(IDs)
elements = []
slash = 0
while slash < 2:
path = path[:-1]
if path[-1] == '/':
slash += 1
for i in IDs:
print(str(i))
a = fnextracter.ExtractElements(path+'lu/lu'+str(i)+'.xml')
for j in a:
exist = False
for k in range(len(elements)):
if elements[k][0] == j[0]:
elements[k].extend(j[2:])
exist = True
if exist == False:
elements.append(j)
print(Calls)
print(elements)
IDs = fnex.ExtractID(fnframe)
if len(IDs) == 0:
print("no example in " + f.name)
data[f.name] = None
continue
#print(IDs)
elems = []
slash = 0
while slash < 2:
fnframe = fnframe[:-1]
if fnframe[-1] == '/':
slash += 1
for i in range(0, len(IDs)):
newElems = fnex.ExtractElements(fnframe + 'lu/lu' + str(IDs[i]) +
'.xml')
#print(newElems)
#print(elems)
if len(elems) == 0 and len(newElems) > 0:
elems.extend(newElems)
continue
exist = False
for j in newElems:
for k in range(len(elems)):
if j[0] == elems[k][0]:
#print(j[2:])
elems[k].extend(j[2:])
exist = True
break
#用例に含まれるかは関係なく全部の意味役割のタグが各luに書いてあるはずだからこれいるかどうかわからん
for f in FNscores:
#print(args)
#print(args[3] + '/' + f[0])
#ID = fnex.searchLu_Verb(args[3] + '/' + f[0], midashi)
#/を削って前の階層に行く
"""
path = args[3]
slash = 0
while slash < 1:
path = path[:-1]
if path[-1] == '/':
slash += 1
"""
#その動詞のFramenetの用例と格解析結果からの格の対応をとる
luFilePath = args[3] + '/' + f[0]
elements[f[0]] = fnex.ExtractElements(luFilePath)
depends[f[0]] = fnex.getRoleReration(midashi, luFilePath,
"/home/oyanagi/rerater/luReibuns")
reibuns[f[0]] = fnex.ExtractReibun(luFilePath)
print(f[0])
print(elements[f[0]])
#print(elements)
#print("ele")
compareList = []
for kfnum in range(min(kfex.getFrameNum(args[1]), 50)):
compareDict = {}
import fnextracter as fnex
import os
import sys
import gensim
args = sys.argv
notInVocab = 0
#embedding自体はいらんけど語彙の確認のためにいる
embedding = gensim.models.KeyedVectors.load_word2vec_format('./model.vec')
for f in os.scandir(path=args[1]):
elems = fnex.ExtractElements(args[1] + "/" + f.name)
# print(vs)
print(f.name, file=sys.stderr) # とても汚いけど標準出力は吸われるのでこうする
exist = False
additional = []
for role in elems:
if role[1] == "Core":
for elem in role[2:]:
if elem not in embedding:
if not exist:
print(f.name)
exist = True
notInVocab += 1
print(elem)
print(FNscores)
#候補のスコア計算して一番いいやつを正解とする
yorei = kfex.ExtractYorei(args[1])
depends = {}
ukemiDepends = {}
elements = {}
reibuns = {}
for f in FNscores:
e = {}
#その動詞のFramenetの用例と格解析結果からの格の対応をとる
luFilePath = args[3] + '/' + f[0]
e["lu"] = fnex.ExtractElements(luFilePath)
framePath = luFilePath
framePath = pm.movePath(framePath, 2)
framePath = framePath + "/frame/" + fnex.ExtractFrame(luFilePath) + ".xml"
subElements = fnex.ExtractSubElements(luFilePath)
frameReibunElements = fnex.ExtractElements_Frame(framePath)
for i in range(len(subElements)):
for j in range(len(frameReibunElements)):
if len(frameReibunElements[j]) > 2:
if subElements[i][0] == frameReibunElements[j][0]:
subElements[i].extend(frameReibunElements[j][2:])
else:
subElements.append(frameReibunElements[j])