def clickThey():
global d_key
global pa
rep_talk()
return
pa += 1
act.configure(text='批刷' + str(pa) + '词')
f = ''
if platform.system() == 'Windows':
f = open("d:/YYY/词林动物1.txt", 'r', encoding='UTF-8')
else:
f = open("/home/ubuntu/DaiMeng/data/xmlData/词林动物1.txt", 'r', encoding='UTF-8')
words = f.readlines()
for word in words:
word = word.strip('\n').replace('\ufeff','')
trans_json(request_(words))
phsen_ana = get_phsen()
Ds_scan()
fast_scan()
ph_und = trans_ltp.get_thisund()
for key in ph_und.Semdic:
t1.insert(INSERT, key + '扫描：：' + str(ph_und.Semdic[key]) + '\n')
def clickIt():
global d_key
global ra
ra += 1
act.configure(text = '随机' + str(ra) +'词')
if e.get() == '':
ra_wo = test_random()
else:
ra_wo = [e.get().split(' ')[0],get_explans(e.get().split(' ')[0])[0]]
t1.delete('1.0', 'end')
t1.mark_set(INSERT , '2.1')
wo_ex = (ra_wo[0] + ':' + ra_wo[1].strip(' ')).replace('.','=').replace('，','！').replace("|",'').replace(',','！')
trans_json (request_(wo_ex))
t1.insert(INSERT, wo_ex + '\n')
phsen_ana = get_phsen()
Ds_scan()
fast_scan()
hownet_words_list = get_hownet_words(r'/home/ubuntu/DaiMeng/data/xmlData/整理.txt')
b = [i for i in hownet_words_list if i.word == ra_wo[0]]
for c in b:
t_A.insert(INSERT,c.word + c.gc + c.DEF + '\n')
print(c.word, c.gc, c.DEF)
sn = 0 #字典返回解释后，0句0词的信息加载为d_key
wr_in =''
for sen in phsen_ana:
print(sen.sen_mean)
if sn == 0:
set_d_key(wo = sen.w_anas[0].wo ,pos = sen.w_anas[0].pos,syn=sen.w_anas[0],cla = sen.w_anas[0].cla)
sen_result = sentence_match_result(sen.w_anas,js_cla='字典')
if sen_result:
t1.insert(INSERT, '内涵知识：' + str(sen_result[-1]) + '\n')
if ra_wo[0].find(':') == -1:
wr_in += ',' + str(sen_result[-1])
sn += 1
if wr_in != '':
formatok = True
cc = 0
while cc < len(ra_wo[0]): #检查节点名，不能有特殊字符
if '._!|/'.find(ra_wo[0][cc]) > -1:
formatok = False
break
cc += 1
if formatok:
d_key = get_d_key()
Gn_audit(d_key, wr_in.strip(','))
ph_und = trans_ltp.get_thisund()
for key in ph_und.Semdic:
t1.insert(INSERT, key + '扫描：：' + str(ph_und.Semdic[key]) + '\n')
def rep_talk(input='我们出门买橡皮呗'):
trans_json (request_(input))
phsen_ana = get_phsen()
for sen in phsen_ana:
sen_result = sentence_match_result(sen.w_anas)
exp_result = experience_mean_result(sen)
dsent_list = []
lsent_list =[]
if (sen_result):
trans_ltp.thinking_str += '这句注意=' + sen.sen_in + '意味着=' + str(sen_result[-1])
dsent_list.append(sen_result.pop(0))
lsent_list.append(sen_result.pop(0))
phsen_ana[i].l_form = lsent_list
phsen_ana[i].d_form = dsent_list
if exp_result:
phsen_ana[i].exp_mean = exp_result
break # 这里干啥忘了，不敢删除那就跑一轮吧
Ds_scan()
fast_scan()
test_dic = {}
ss = 0
ph_und = trans_ltp.get_thisund()
Has_Qu = False # 疑问句指示
print(trans_ltp.thinking_str)
for sen in phsen_ana:
disp_wo = ''
if_Q = False
Q_if = False
disp_wo += sen.sen_in
for w in sen.w_anas:
if len(disp_wo) < 500:
disp_wo += (w.wo + '::' + w.yuf + '::' + w.pos + '::' + w.wgr + ' :: ' + w.rel + '\n')
if w.pos == '疑问词':
if_Q = True
phsen_ana[ss].sen_mean = phsen_ana[ss].sen_mean.replace('=' + w.wo, '=疑问=' + w.wo) # 做上疑问标记,=号方便劈开
elif w.wo == '？':
Q_if = True
test_dic['句子sn=' + str(ss)] = sen.sen_in + '\n' + sen.sen_mean + '\n'
test_dic['词汇sn=' + str(ss)] = disp_wo
if if_Q and Q_if:
phsen_ana[ss].env_mean += ',句式=疑问'
Has_Qu = True
ss += 1
if len(phsen_ana)>1:
phase_ana() # 这里是段落理解的演示，单句无效
ph_und = trans_ltp.get_thisund()
print('tttttttttt'+ thinking_str)
if '回复' in test_dic:
test_dic['回复'] += Thlog_ana(ph_und.rec[1])
else: test_dic['回复'] = Thlog_ana(ph_und.rec[1])
else: # 这里是分类知识库的搜多，仅限单句
Thlog_ana(myexp_ana())
ph_und = trans_ltp.get_thisund()
if '回复' in test_dic:
test_dic['回复'] += Thlog_ana(myexp_ana())
else: test_dic['回复'] = Thlog_ana(myexp_ana())
if '回复' in test_dic:
test_dic['回复'] += Thlog_ana(myexp_ana(no='我的日记'))
else: test_dic['回复'] = Thlog_ana(myexp_ana(no='我的日记'))
test_dic['扫描'] = str(ph_und.Semdic)
if len(test_dic['回复']) < 3 :
ph_und = trans_ltp.get_thisund()
for key in ph_und.Semdic:
if key == 'V' :
for vv in ph_und.Semdic[key]:
if vv.find('sn='):
continue
thinking_str += '重要的'+ vv + '\n'
test_dic['回复'] += Thlog_ana()
print(ph_und.Semdic)
print('***************')
print(test_dic['回复'])
test_dic['回复']=test_dic['回复'].replace('\n\n','\n').replace('\n','\nΨ').strip('Ψ')
test_dic['回复']=test_dic['回复'].replace('\n\n','\n').replace('\n','\nΨ').strip('Ψ')
return test_dic
def clickMe():
global ti
global ssi
ti += 1
inp = e.get()
ssi += len(re.split(r'[，。！？；]|\.\.\.',inp)) -1
trans_json(request_(inp))
t_nlu.member.append(inp) # 做输入/对话的记录
phsen_ana = get_phsen()
act.configure(text = str(ti) + '段:' + str(len(phsen_ana)) + '句')
t0.delete('1.0','end')
t1.delete('1.0','end')
t_A.delete('1.0','end')
t0.mark_set(INSERT , '2.1')
t1.mark_set(INSERT , '2.1')
t_A.mark_set(INSERT , '2.1')
i = 0
Has_Qu = False
sn = 0
for sen in phsen_ana:
if sn == 0:
set_d_key(wo = sen.w_anas[0].wo ,pos = sen.w_anas[0].pos,syn=sen.w_anas[0],cla = sen.w_anas[0].cla)
sen_result = sentence_match_result(sen.w_anas)
exp_result = experience_mean_result(sen)
dsent_list = []
lsent_list =[]
if (sen_result):
trans_ltp.thinking_str += '这句注意=' + sen.sen_in + '意味着=' + str(sen_result[-1])
dsent_list.append(sen_result.pop(0))
lsent_list.append(sen_result.pop(0))
phsen_ana[i].l_form = lsent_list
phsen_ana[i].d_form = dsent_list
if exp_result:
phsen_ana[i].exp_mean = exp_result
t0.insert(INSERT,sen.sen_mean + '\n')
t0.insert(INSERT,str(lsent_list) + '\n')
t0.insert(INSERT,str(dsent_list) + '\n')
t0.insert(INSERT,str(exp_result) + '\n')
sn += 1
Ds_scan()
fast_scan()
for sen in phsen_ana:
disp_wo = ''
if_Q =False
Q_if =False
disp_wo += sen.sen_in
for w in sen.w_anas:
if len(disp_wo) < 1200:
disp_wo +=(w.wo + '::' + w.yuf + '::'+w.pos + '::' + w.wgr + ' :: ' +w.cla + '::' + w.rel + '\n')
if w.pos=='疑问词':
if_Q = True
phsen_ana[i].sen_mean = phsen_ana[i].sen_mean.replace('='+ w.wo,'=疑问='+w.wo) #做上疑问标记,=号方便劈开
elif w.wo == '？':
Q_if = True
t1.insert(INSERT,disp_wo)
if if_Q and Q_if:
phsen_ana[i].env_mean += ',句式=疑问'
t0.insert(INSERT,'句式=疑问')
Has_Qu = True
i += 1
ph_und = trans_ltp.get_thisund()
if Has_Qu:
Q_res = Query_ana(phsen_ana)
t_A.insert(INSERT, str(Q_res) + '\n')
for key in ph_und.Semdic:
t1.insert(INSERT, key + '扫描：：' + str(ph_und.Semdic[key]) +'\n')
phase_ana()
if len(phsen_ana) > 0:
t_A.insert(INSERT, Thlog_ana())
t_A.insert(INSERT, Thlog_ana(myexp_ana()))
t_A.insert(INSERT, Thlog_ana(myexp_ana(no='我的日记')))
if 'UND' not in t_nlu.features:
t_nlu.features['UND'] = {}
t_nlu.features['UND'][inp] = ph_und # 等完成后，一起收了，注意这是个二维队列，与ph_und平行的还能添加更多的后续分析
print('ssi' * 5 + ' ' + str(ssi))
if ssi > 4:
t_A.insert(INSERT, Thlog_ana(rep_ph_ana(t_nlu))) # 对话段落每5句再做一次分析。
ssi = 0
def fast_scan(js_cla='fast',
sxml=None): # 面向phsen_ana,对所有句子对短小属性做快速、可重复的匹配。已支持@as，@C，@Y，双重变量
global this_und
root = sxml # 添加表达2之外其他描述句式的选择，要送出句式集合节点的上一层
if not root:
root = ET.parse(r'/home/ubuntu/DaiMeng/data/xmlData/表达2.xml').getroot()
this_und = trans_ltp.get_thisund()
Yf_ele = ['主语', 'V', '状语']
for ele in Yf_ele:
this_und.Semdic[ele] = ''
for js_no in root.iter(js_cla): # 增加的第一层是分类标签，一般只有0号有效，所以跑一次就可以了
ss = 0
for sen in this_und.phsen_ana:
Yf_str = sen.sen_mean.replace('::', ',').split(',')
for ele in Yf_ele:
sen_ele = ''
for ys in Yf_str:
if ys.find(ele + '=') > -1:
sen_ele += ',' + ys
if sen_ele != '':
this_und.Semdic[ele] += 'sn=' + str(
ss) + sen_ele + '\n' # 每句scan后写一行
matrec_dic = {} #记录每个主题中该句子各词语匹配情况
for node in js_no.iter('句式'):
logic_express = node.find('逻辑句式').text
sem_t = None
if '主题' in node.attrib:
sem_t = node.attrib['主题']
sen_sem = ''
for ch_no in node.iter('描述句式'):
if '逻辑' in ch_no.attrib:
logic_express = ch_no.attrib[
'逻辑'] # 在特殊情况下，使用单句有效的逻辑句式，用于截胡后边
if sem_t != None:
if not sem_t in this_und.Semdic:
this_und.Semdic[sem_t] = ''
des_str = ch_no.text.split(',')
as_pars = [] # @as参数组的列表
as_par = ['wo', ''] # @as匹配参数的元组
if_as = False
if ch_no.text.find('@as') > -1:
dd = 0
for des in des_str: # 记录装载本次des句式中包含的as变量位置和参数
if des[:3] == '@as':
if_as = True
as_pai = des.split('.')
if int(
as_pai[1]
) == -1: # 如果是公用的dicexpl_key则单独装载。(-1表示)
d_key = get_d_key()
if as_pai[2] == 'wo':
des_str[dd] = d_key.wo
elif as_pai[2] == 'pos':
des_str[dd] = '@P.' + d_key.pos
elif as_pai[2] == 'syn':
des_str[dd] = d_key.syn.split('+')[0]
elif as_pai[2] == 'cla':
cat_cla = d_key.cla.split(
'+')[0].split('.')[-1]
cat_cla = re.sub('\d+', '', cat_cla)
cat_cla = re.sub('[=#@]+', '', cat_cla)
des_str[dd] = '@G.in(' + cat_cla[
cat_cla.find('_') +
1:] + ')' # 只取对应目标的第0个分类的最后一截。
# @as变量的1号位置是跟随的des句式元素序号，从0开始；2号位置是跟随的内容（分类，词性或者词本身），dd是本身的位置。
else:
as_par = [dd, int(as_pai[1]), as_pai[2]]
as_pars.append(as_par)
dd += 1
j = 0
w = 0
dista = 0
par_C = ''
mat_dic = {}
B_C = -1 # @C变量开关
last_mat = -1 # 上一个匹配好的w_anas位置
while j < len(des_str): # 以描述句式为基准进行匹配
if w > len(sen.w_anas) - 1:
break
mat_this = False
mat_des = False
s_mat = des_str[j].replace('opt', '')
if des_str[j].find('@C') > -1: # 不管@C函数
B_C = w
if des_str[j].split('.')[1] == 'wo':
par_C = sen.w_anas[w].wo
if j == len(des_str) - 1: # 最后一项 @C全收
ww = w + 1
while ww < len(sen.w_anas):
par_C += '+' + sen.w_anas[ww].wo
ww += 1
B_C = -1
last_mat = len(sen.w_anas) - 1
j += 1
w += 1
continue
elif des_str[j].find('@Y.') > -1: # 依存句法的relate匹配
for ww in range(w, len(
sen.w_anas)): # 在写@Y变量时要考虑到前后秩序
if sen.w_anas[ww].pos == '标点':
break
# 由wgr词组判断是否关联 一级直连的放行，因为前边只能是HED
if_rel = ('_' + sen.w_anas[ww].wgr + '_').find(
'_' + sen.w_anas[last_mat].wo + '_') > -1
if j == 0:
if_rel = True
elif 'HEDCOV'.find(sen.w_anas[ww].yuf) > -1:
if_rel = True
# 这里要求前后的两项有语法依存关系，HED关联的那一层关系无效，不能跨越标点使用
# print(sen.w_anas[ww].yuf + '++++' + cat_str('@Y.', des_str[j], '+'))
if if_rel and sen.w_anas[ww].yuf == cat_str(
'@Y.', des_str[j] + '+',
'+'): # 会漏一个字符
if des_str[j].find('+') == -1:
this_und.phsen_ana[ss].w_anas[
w].ele = '+' + des_str[j]
elif element_match(
sen.w_anas[ww],
des_str[j].split('+')[-1]):
this_und.phsen_ana[ss].w_anas[
w].ele = '+' + des_str[j]
else:
continue
matrec_dic[w] = des_str[j]
mat_dic[des_str[j]] = sen.w_anas[ww].wo
mat_this = True
last_mat = ww
break
if mat_this: # @Y变量从这里输出了
if j == len(des_str) - 1:
t_sem = str(
logic_sentence(mat_dic,
logic_express)).replace(
'\'', '')
this_und.phsen_ana[ss].w_anas[
last_mat].ele = t_sem #以此记录匹配上的词语位置
if sen_sem.find(
t_sem.split('=')[-1].strip(
']')) == -1:
sen_sem += ',' + t_sem
if len(sen.w_anas) - w > len(
des_str): # 后边够长时可以重复匹配
j = 0
w = max(w, last_mat) + 1
continue
else:
break
else:
j += 1
w = max(w, last_mat) + 1
continue
else:
j = 0
w += 1
continue
if element_match(this_und.phsen_ana[ss].w_anas[w],
s_mat):
last_mat = w
this_und.phsen_ana[ss].w_anas[
w].ele = '+' + des_str[j]
matrec_dic[w] = des_str[j]
mat_dic[des_str[j]] = sen.w_anas[w].wo
if if_as:
for a_par in as_pars: # 每一项匹配成功后，查找后边有没有as项追随自己，发现后按各参数执行替换。
if int(a_par[1]) == j:
if a_par[2] == 'v':
des_str[int(
a_par[0])] = des_sent_list[j]
elif a_par[2] == 'wo':
des_str[int(
a_par[0])] = sen.w_anas[w].wo
elif a_par[2] == 'pos':
des_str[int(
a_par[0]
)] = '@P.' + sen.w_anas[w].pos
elif a_par[2] == 'cla':
des_str[int(
a_par[0]
)] = '@G.in' + sen.w_anas[
w].cla # if_as复原.cla
if B_C > -1:
mat_dic[des_str[j - 1]] = par_C
B_C = -1
if j == len(des_str) - 1:
mat_des = True
else:
dista = 1
j += 1
w += 1
continue
elif B_C > -1:
par_C += '+' + sen.w_anas[w].wo
w += 1
continue
elif des_str[j].find('opt') > -1:
j += 1
continue
elif dista > 0 and '助词副词形容词数词'.find(
sen.w_anas[w].pos) > -1:
dista += -1
w += 1
continue
elif len(
sen.w_anas) - w > len(des_str): # 后边够长时可以也重复匹配
if if_as:
des_str = ch_no.text.split(',') # if_as复原
mat_dic[des_str[j]] = ''
j = 0
w += 1
continue
elif not mat_des:
break
if mat_des:
if if_as:
des_str = des_str = ch_no.text.split(
',') # if_as复原
# 每句在字典栏目中加写一行
t_sem = str(logic_sentence(mat_dic,
logic_express)).replace(
'\'', '')
if this_und.phsen_ana[ss].w_anas[
last_mat].ele == None:
this_und.phsen_ana[ss].w_anas[
last_mat].ele = t_sem
else:
this_und.phsen_ana[ss].w_anas[
last_mat].ele += '\n' + t_sem
if sen_sem.find(t_sem.split('=')[-1]) == -1:
sen_sem += ',' + t_sem
if len(sen.w_anas) - w > len(
des_str): # 后边够长时可以重复匹配
j = 0
w += 1
continue
else:
break
if sen_sem != '':
this_und.Semdic[sem_t] += 'sn=' + str(ss) + sen_sem + '\n'
if sen_sem.find('发送信息=') > -1:
if sen.sen_in[-1] != '”':
sss = ss + 1
while sss < len(this_und.phsen_ana):
this_und.Semdic[sem_t] += 'sn=' + str(
sss) + ',信息={' + this_und.phsen_ana[
sss].sen_in.replace(' ', '+') + '\n'
sss += 1
if this_und.phsen_ana[sss].sen_in[-1] == '”':
break
ss += 1
ele_sts(this_und.Semdic)
break
this_und.rec[0] = trans_ltp.input_a
this_und.rec[1] = trans_ltp.thinking_str
trans_ltp.set_thisund(ph=this_und)
return this_und
def rep_talk(input='我们出门买橡皮呗'):
trans_ltp.thinking_str = ''
global ssi
ssi += len(re.split(r'[，。！？；]|\.\.\.',input)) -1
trans_json(request_(input))
t_nlu.member.append(input) # 做输入/对话的记录
if len(input)>4:
if input[:4] == '@古诗#':
print(input[4:])
phase_json.phase_json_ana(input[4:], idiom = True)
print(trans_ltp.thinking_str)
trans_json (request_(input))
phsen_ana = get_phsen()
for sen in phsen_ana:
sen_result = sentence_match_result(sen.w_anas)
exp_result = experience_mean_result(sen)
dsent_list = []
lsent_list =[]
if (sen_result):
trans_ltp.thinking_str += '这句注意=' + sen.sen_in + '意味着=' + str(sen_result[-1])
dsent_list.append(sen_result.pop(0))
lsent_list.append(sen_result.pop(0))
phsen_ana[i].l_form = lsent_list
phsen_ana[i].d_form = dsent_list
if exp_result:
phsen_ana[i].exp_mean = exp_result
break # 这里干啥忘了，不敢删除那就跑一轮吧
Ds_scan()
fast_scan()
test_dic = {}
ss = 0
ph_und = trans_ltp.get_thisund()
Has_Qu = False # 疑问句指示
print(trans_ltp.thinking_str)
for sen in phsen_ana:
disp_wo = ''
if_Q = False
Q_if = False
disp_wo += sen.sen_in
for w in sen.w_anas:
if len(disp_wo) < 500:
disp_wo += (w.wo + '::' + w.yuf + '::' + w.pos + '::' + w.wgr + ' :: ' + w.rel + '\n')
if w.pos == '疑问词':
if_Q = True
phsen_ana[ss].sen_mean = phsen_ana[ss].sen_mean.replace('=' + w.wo, '=疑问=' + w.wo) # 做上疑问标记,=号方便劈开
elif w.wo == '？':
Q_if = True
test_dic['句子sn=' + str(ss)] = sen.sen_in + '\n' + sen.sen_mean + '\n'
test_dic['词汇sn=' + str(ss)] = disp_wo
if if_Q and Q_if:
phsen_ana[ss].env_mean += ',句式=疑问'
Has_Qu = True
ss += 1
if len(phsen_ana)>1:
print(321,phsen_ana)
phase_ana() # 这里是段落理解的演示，单句无效
ph_und = trans_ltp.get_thisund()
# print('tttttttttt'+ trans_ltp.thinking_str)
if '回复' in test_dic:
test_dic['回复'] += Thlog_ana(ph_und.rec[1])
else:
test_dic['回复'] = Thlog_ana(ph_und.rec[1])
#print('***************')
#print(test_dic['回复'])
else: # 这里是分类知识库的搜多，仅限单句
Thlog_ana(myexp_ana())
ph_und = trans_ltp.get_thisund()
if '回复' in test_dic:
test_dic['回复'] += Thlog_ana(myexp_ana())
else: test_dic['回复'] = Thlog_ana(myexp_ana())
#print('********2*******')
#print(test_dic['回复'])
if '回复' in test_dic:
trans_ltp.thinking_str = ''
test_dic['回复'] += Thlog_ana(myexp_ana(no='我的日记'))
else: test_dic['回复'] = Thlog_ana(myexp_ana(no='我的日记'))
print('********3*******')
print(test_dic['回复'])
test_dic['扫描'] = str(ph_und.Semdic)
if len(test_dic['回复']) < 3 :
ph_und = trans_ltp.get_thisund()
for key in ph_und.Semdic:
if key == 'V' :
for vv in ph_und.Semdic[key]:
if vv.find('sn='):
continue
thinking_str += '重要的'+ vv + '\n'
test_dic['回复'] += Thlog_ana()
print(ph_und.Semdic)
if 'UND' not in t_nlu.features:
t_nlu.features['UND'] = {}
t_nlu.features['UND'][input] = ph_und # 等完成后，一起收了，注意这是个二维队列，与ph_und平行的还能添加更多的后续分析
print('ssi' * 5 + ' ' + str(ssi))
if ssi > 4:
test_dic['回复'] += Thlog_ana(rep_ph_ana(t_nlu)) # 对话段落每5句再做一次分析。
ssi = 0
test_dic['回复']=test_dic['回复'].replace('\n\n','\n').replace('\n','\nΨ').strip('Ψ')
print('******4*****')
print(test_dic['回复'])
return test_dic