def storage():
datapath = os.path.dirname(__file__) + '\data'
imgpath = r'D:\\IMG'
gmp = GMP(imgpath)
regisration = Regisration(imgpath)
license = License(imgpath)
certificate = ProductionCertificate(datapath, imgpath)
pga = Improtdrug(imgpath)
for file in os.walk(datapath):
id_code = randomidcode()
for file_name in file[2]:
# if 'GMP证书' in file_name:
gmp.gmp(file[0], id_code)
# elif "营业执照" in file_name:
license.license(file[0], id_code)
# elif "药品再注册批件" in file_name:
regisration.regisration(file[0], id_code)
# elif '药品生产许可证' in file_name:
certificate.recognize(file[0], id_code)
# elif '说明书' in file_name:
introduction.run_introduction(file[0], id_code)
# elif '进口药品注册证' in file_name:
try:
pga.start(file[0], id_code, 'shuai', '')
except Exception as e:
logmgr = LogMgr()
logmgr.error(file[0]+ ":" + str(e))
continue
break
def __init__(self, typeid, app_id = APP_ID, api_key = API_KEY, secret_key = SECRET_KEY):
self.client = AipOcr(app_id, api_key, secret_key)
#self.client = AipOcr(appid[1], apikey[1], secretkey[1])
self.typeid = typeid
self.codepath = os.path.dirname(__file__)
self.datapath = self.codepath + '\data'
os.makedirs(self.datapath, exist_ok=True)
self.log = LogMgr()
def json2word(wordlist,savepath,savename):# savepath = './word' # savename = 'test1'
emb_filename = os.path.join(savepath, savename+'.doc')
if not os.path.isdir(os.path.split(emb_filename)[0]):
os.makedirs(os.path.split(emb_filename)[0])
try:
with open(emb_filename, "w",encoding='utf-8') as f:
for i in wordlist:
f.write(i + "\n")
f.close()
except Exception as e:
print(e)
log_mgr = LogMgr()
log_mgr.error('[mylog]This is error log')
# savepath = './word'
# savename = 'test1'
# wordlist = ["淋日期有合","【有效期】24个月","请仔细阅读说明书井在医师指导下使用"]
# json2word(wordlist,savepath,savename)
class JobTable(object):
'''
工作表
'''
db = cxOracle()
logmgr = LogMgr()
def __init__(self):
self.jobdict = dict()
#self.jobdict['SER_IP'] = '10.67.28.8'
self.dbtable = 'OCRWORKFILE'
self.dbflag = 2
def job_add(self, jobtmp):
self.jobdict = jobtmp
def job_del(self):
if self.jobdict:
self.jobdict.clear()
def update_item(self, find_key, find_value, update_key, update_value):
'''
????????????
@find_key ----????????
@find_value ----?????????
@update_key ----????????
@update_value ----????????
'''
self.db.update('OCRWORKFILE', find_key, find_value, update_key,
update_value)
def job_todb(self):
try:
jobsql, jobparam = self.db.getsavesql(self.dbtable, self.jobdict,
self.dbflag)
self.db.insert(jobsql, jobparam)
except Exception as e:
self.logmgr.error(str(e))
class License(Tools):
"""
识别营业执照
"""
def __init__(self, imgpath):
Tools.__init__(self)
self.imgpath = imgpath
self.logmgr = LogMgr()
def _recognize(self, datas, nums):
"""
程序的主逻辑
"""
keylist = []
datadict = dict()
for (word, i) in zip(datas, range(0, nums)):
'''
循环读识别出的数据，然后根据judge_keywords函数是否提取到了关键信息；
若提取到了，则保存到datadict中。
若未提取到，list_result为空。有两种情况，
1.这段信息不是我们所需要的。
2.这段信息是上个关键字的值。
然后执行else，进行更精确的判别。若是需归到上个字段，则循环递减，根据
keylist[1],也就是list_reault[2]是否出现再上面的某个字段。若有则追加。
'''
list_result = self._judge_keywords(word['words'])
if '名' == word['words'] and datas[i + 1]['words'][0] == '称':
datadict['ENT_NAME'] = datas[i + 1]['words'][1:]
continue
elif '类' == word['words'] and '型' == datas[i + 1]['words'][0]:
datadict['ENT_TYPE'] = datas[i + 1]['words'][1:]
continue
elif '住' == word['words'] and '所' == datas[i + 1]['words'][0]:
datadict['住所'] = datas[i + 1]['words'][1:]
continue
if list_result != None:
if list_result[
0] in datadict and keylist[-1][0] != list_result[0]:
datadict[list_result[0]] += list_result[1]
flag = 1
else:
datadict[list_result[0]] = list_result[1]
flag = 1
#保存关键字段的信息，以及这段信息原本关键字段的信息
keylist.append([list_result[0], list_result[2]])
else:
j = i
while j > 0:
if not keylist:
break
if keylist[-1][0] == '统一社会信用代码':
if re.search(r'[\u4e00-\u9fa5]+', word['words']):
break
if flag:
if keylist[-1][1] in datas[j]['words']:
datadict[keylist[-1][0]] += word['words']
break
j -= 1
return datadict
def _judge_keywords(self, strword):
'''
判断关键字,若识别到关键字，返回一个包含关键字的list。
$resultlist[0] -----要入库的关键字
$resultlist[1] -----提取到内容
$resultlist[2] -----需判断的信息中本来的关键字
如:'证书编号:H12345',resultlist = ['证书编号', 'H12345', '证书编号']
'证书号:H123', resultlist = ['证书编号', 'H123', '证书号']
'''
re_name = re.compile(r"名称")
re_social_code = re.compile(r"统*一社*会信用代码|统一*社会*信用*代码")
re_type = re.compile(r"类型")
re_residence = re.compile(r"住所")
re_legal_representative = re.compile(r"法定*代表*人|法*定代表人*")
re_capital = re.compile(r'注册*资本*|注*册资*本')
re_establish = re.compile(r'成立*日期*|成*立日*期')
re_period = re.compile(r'营业*期限*|营*业期*限')
re_scope = re.compile(r"经营*范围*|经*营范*围")
re_authority = re.compile(r"登记*机关*|登*记机*关")
if len(strword) >= 10:
index = 8
elif len(strword) >= 8:
index = 6
else:
index = len(strword)
if re_social_code.search(strword[:index]):
return [
'统一社会信用代码', strword[re_social_code.search(strword).span()[1]:],
re_social_code.search(strword).group()
]
elif re_legal_representative.search(strword[:index]):
return [
'法定代表人',
strword[re_legal_representative.search(strword).span()[1]:],
re_legal_representative.search(strword).group()
]
elif re_capital.search(strword[:index]):
return [
'注册资本', strword[re_capital.search(strword).span()[1]:],
re_capital.search(strword).group()
]
elif re_establish.search(strword[:index]):
return [
'成立日期', strword[re_establish.search(strword).span()[1]:],
re_establish.search(strword).group()
]
elif re_period.search(strword[:index]):
return [
'营业期限', strword[re_period.search(strword).span()[1]:],
re_period.search(strword).group()
]
elif re_scope.search(strword[:index]):
return [
'经营范围', strword[re_scope.search(strword).span()[1]:],
re_scope.search(strword).group()
]
elif re_authority.search(strword[:index]):
return [
'登记机关', strword[re_authority.search(strword).span()[1]:],
re_authority.search(strword).group()
]
else:
return None
def license_deploy(self, imgs, id_code):
flag = 0
tmp = ''
for file in imgs:
file_name = file['imgpath'].split('/')[-1]
id = file['imgpath'].split('/')[-2]
if re.search(r'[\u4e00-\u9fa5]+', id):
dragname = re.search(r'[\u4e00-\u9fa5]+', id).group()
else:
dragname = re.search(r'[\u4e00-\u9fa5]+', file_name).group()
if dragname.find('(') > 0:
dragname = dragname[:dragname.find('(')]
if 'error_code' in file['imgjson']:
self.logmgr.error(file['imgpath'] + " : Img Size Error!")
continue
datas = file['imgjson']['words_result']
nums = file['imgjson']['words_result_num']
if len(datas) > 0 and nums > 0:
datadict = self._recognize(datas, nums)
######################################增加部分###########################################
datadict['ID_CODE'] = id_code
datadict['REMARK'] = ''
datadict['ADD_USER'] = '******'
datadict['JOB_ID'] = self._generatemd5(file['imgpath'])
######################################增加部分###########################################
if not datadict:
nums = self._cleandata(datadict, datas, nums)
return datadict
if '登记机关' in datadict:
del datadict['登记机关']
return datadict
#try:
# #self._data_to_db('BUSINESSLICENCE', datadict)
# nums = self._cleandata(datadict, datas, nums)
#except Exception as e:
# print('Error: ', e)
# self.logmgr.error(file[0] + '\\' + file_name + "insert error!! : " + str(e))
# self._update_item('OCRWORKFILE','JOB_ID', jobid,'IS_TO_DB','F')
# nums = self._cleandata(datadict, datas, nums)
def license(self, path, id_code):
flag = 0
temp = ''
jobdict = {}
for file in os.walk(path):
page = 1
for file_name in file[2]:
if '营业执照' in file_name:
imgname = file_name.split('.')[0]
curpath = file[0].split('data')[1]
index = imgname.rfind('_')
id = curpath[curpath.rfind('\\') + 1:]
if re.search(r'[\u4e00-\u9fa5]+', id):
dragname = re.search(r'[\u4e00-\u9fa5]+', id).group()
else:
dragname = re.search(r'[\u4e00-\u9fa5]+',
file_name).group()
if dragname.find('(') > 0:
dragname = dragname[:dragname.find('(')]
datajson = self._load_json(file[0] + '\\' + file_name)
original_path = self.imgpath + '\\' + curpath + '\\' + imgname[:index
-
2] + '.' + 'pdf'
#服务器
jobdict['SER_IP'] = '10.67.28.8'
#job id
jobdict['JOB_ID'] = self._generatemd5(file[0] + imgname)
jobid = jobdict['JOB_ID']
jobdict['SRC_FILE_NAME'] = imgname[:index -
2] + '.' + 'pdf'
jobdict['SRC_FILE_PATH'] = original_path
#原文件
jobdict['CUT_FILE_NAME'] = imgname[:index] + '.' + imgname[
index:].split('_')[1]
#原路径
jobdict['CUT_FILE_PATH'] = 'G:\\IMG' + '\\' + curpath
#时间
jobdict['HANDLE_TIME'] = time.strftime(
"%Y-%m-%d %X", time.localtime())
#药品名
jobdict['DRUG_NAME'] = dragname
#影像件类型
jobdict['FILE_TYPE'] = '营业执照'
#同一套影像件识别码
jobdict['ID_CODE'] = id_code
#分公司
jobdict['SRC_CO'] = curpath.split('\\')[1]
#源文件相对路径
jobdict[
'FILE_REL_PATH'] = '\\' + imgname[:
index] + '.' + imgname[
index:].split(
'_')[1]
#文件服务器域名
jobdict['SYS_URL'] = '10.67.28.8'
#页数
jobdict['PAGE_NUM'] = page
#文件ocr解析识别状态 fk sysparams
jobdict['OCR_STATE'] = 'T'
#备注说明
jobdict['REMARK'] = ''
#创建用户
jobdict['ADD_USER'] = '******'
#图片过大或者一些原因，没有识别出来就会有error_code字段
if 'error_code' in datajson:
jobdict['IS_TO_DB'] = 'F'
self.job.job_add(jobdict)
self.job.job_todb()
self.job.job_del()
self.logmgr.error(file[0] + '\\' + file_name +
": img size error!")
continue
datas = datajson['words_result']
nums = datajson['words_result_num']
flag = 1
#中间文件
jobdict['MID_FILE_NAME'] = file_name
#中间文件路径
jobdict['MID_FILE_PATH'] = file[0]
#评分
jobdict['OCR_SCORE'] = int(self._getscore(datas, nums))
#影像件内容是否入库
if len(datas) > 0 and nums > 0:
jobdict['IS_TO_DB'] = 'T'
else:
jobdict['IS_TO_DB'] = 'F'
#文件文本内容
jobdict['FILE_TEXT'] = self._middict(
datas, self.codepath + '\\middata\\' + curpath,
imgname)
###########################
temp = jobdict['FILE_TEXT']
###########################
#jobdict['JOB_ID'] = self._generatemd5(jobdict['FILE_TEXT'])
###############
page += 1
self.job.job_add(jobdict)
self.job.job_todb()
self.job.job_del()
if flag:
if len(datas) > 0 and nums > 0:
datadict = self._recognize(datas, nums)
######################################增加部分###########################################
datadict['ID_CODE'] = id_code
datadict['REMARK'] = ''
datadict['ADD_USER'] = '******'
datadict['JOB_ID'] = self._generatemd5(temp)
######################################增加部分###########################################
print(datadict)
if not datadict:
nums = self._cleandata(datadict, datas, nums)
continue
if '登记机关' in datadict:
del datadict['登记机关']
try:
self._data_to_db('BUSINESSLICENCE', datadict)
nums = self._cleandata(datadict, datas, nums)
except Exception as e:
print('Error: ', e)
self.logmgr.error(file[0] + '\\' + file_name +
"insert error!! : " + str(e))
self._update_item('OCRWORKFILE', 'JOB_ID', jobid,
'IS_TO_DB', 'F')
nums = self._cleandata(datadict, datas, nums)
continue
def __init__(self, imgpath):
Tools.__init__(self)
self.imgpath = imgpath
self.logmgr = LogMgr()
class Regisration(Tools):
"""
识别药品再注册批件
"""
def __init__(self, imgpath):
Tools.__init__(self)
self.imgpath = imgpath
self.logmgr = LogMgr()
def _recognize(self, datas, nums):
"""
程序的主逻辑
"""
keylist = []
datadict = dict()
for (word, i) in zip(datas, range(0, nums)):
'''
循环读识别出的数据，然后根据judge_keywords函数是否提取到了关键信息；
若提取到了，则保存到datadict中。
若未提取到，list_result为空。有两种情况，
1.这段信息不是我们所需要的。
2.这段信息是上个关键字的值。
然后执行else，进行更精确的判别。若是需归到上个字段，则循环递减，根据
keylist[1],也就是list_reault[2]是否出现再上面的某个字段。若有则追加。
'''
list_result = self._judge_keywords(word['words'])
if list_result != None:
if list_result[
0] in datadict and keylist[-1][0] != list_result[0]:
datadict[list_result[0]] += list_result[1]
flag = 1
else:
datadict[list_result[0]] = list_result[1]
flag = 1
#保存关键字段的信息，以及这段信息原本关键字段的信息
keylist.append([list_result[0], list_result[2]])
else:
j = i
while j > 0:
if not keylist:
break
if keylist[-1][0] == '批准文号':
if re.search(r'.?[a-zA-z][0-9]+', word['words']):
break
if keylist[-1][0] == '规格':
if not re.search(r'.*m*g|.*m*l', word['words']):
break
if flag:
if keylist[-1][1] in datas[j]['words']:
datadict[keylist[-1][0]] += word['words']
break
j -= 1
return datadict
def _judge_keywords(self, strword):
'''
判断关键字,若识别到关键字，返回一个包含关键字的list。
$resultlist[0] -----要入库的关键字
$resultlist[1] -----提取到内容
$resultlist[2] -----需判断的信息中本来的关键字
如:'证书编号:H12345',resultlist = ['证书编号', 'H12345', '证书编号']
'证书号:H123', resultlist = ['证书编号', 'H123', '证书号']
'''
re_coname = re.compile(r"名称")
re_num_orig = re.compile(r"原始*编号*|原*始编*号")
re_drug_standord = re.compile(r"药品*标准*|药*品标*准")
re_drug_valid = re.compile(r"药品*有效期*|药*品有*效期")
re_drug_class = re.compile(r"药品*分类*|药*品分*类")
re_common_name = re.compile(r'药品*通?用名称?|药*品通?用名?称')
re_product_name = re.compile(r'商?品名称?|商?品名?称')
re_english = re.compile(r'英文?名称?|英文名?称')
re_pinyin = re.compile(r'汉语?拼音?|汉?语拼?音')
re_coaddr = re.compile(r"生产*地址*|生*产地*址")
re_conclution = re.compile(r"审批*结论*|审*批结*论")
re_drug_approval = re.compile(r"药品*批准文*号|药*品批*准文号")
re_drug_approval_valid = re.compile(r"药*品批准文号有*效期|药品*批准文号*有效*期")
#TODO:有些注册批件的生产厂家
re_annex = re.compile(r"附件")
re_zhusong = re.compile(r"主送")
re_chaobao = re.compile(r"抄报")
re_regisnum = re.compile(r"注册*证号*|注*册证*号")
re_regisnum_valid = re.compile(r"注册*证号有效期*|注*册证号有效*期")
re_specification = re.compile(r'规格')
re_jixing = re.compile(r'剂型')
if len(strword) >= 8:
index = 6
else:
index = len(strword)
if (re.match(r'.+?(?:\:)', strword[:index])):
if re_common_name.search(strword[:8]):
return [
'药品名称', strword[re_common_name.search(strword).span()[1]:],
re_common_name.search(strword).group()
]
elif re_pinyin.search(strword[:index]):
return [
'汉语拼音', strword[re_pinyin.search(strword).span()[1] + 1:],
re_pinyin.search(strword).group()
]
elif re_coname.search(strword[:4]):
return [
'名称', strword[re_coname.search(strword).span()[1] + 1:],
re_coname.search(strword).group()
]
elif re_coaddr.search(strword[:index]):
return [
'生产地址', strword[re_coaddr.search(strword).span()[1] + 1:],
re_coaddr.search(strword).group()
]
else:
return None
else:
if re_common_name.search(strword[:8]):
return [
'药品名称', strword[re_common_name.search(strword).span()[1]:],
re_common_name.search(strword).group()
]
elif re_pinyin.search(strword[:index]):
return [
'汉语拼音', strword[re_pinyin.search(strword).span()[1]:],
re_pinyin.search(strword).group()
]
elif re_coname.search(strword[:4]):
return [
'名称', strword[re_coname.search(strword).span()[1]:],
re_coname.search(strword).group()
]
elif re_coaddr.search(strword[:index]):
return [
'生产地址', strword[re_coaddr.search(strword).span()[1]:],
re_coaddr.search(strword).group()
]
elif re_conclution.search(strword[:index]):
return [
'审批结论', strword[re_conclution.search(strword).span()[1]:],
re_conclution.search(strword).group()
]
elif re_drug_approval.search(strword[:index]):
return [
'再注册证批准文号',
strword[re_drug_approval.search(strword).span()[1]:],
re_drug_approval.search(strword).group()
]
elif re_drug_approval_valid.search(strword[:index]):
return [
'药品批准文号有效期',
strword[re_drug_approval_valid.search(strword).span()[1]:],
re_drug_approval_valid.search(strword).group()
]
elif re_regisnum.search(strword[:index]):
return [
'注册证号', strword[re_regisnum.search(strword).span()[1]:],
re_regisnum.search(strword).group()
]
elif re_regisnum_valid.search(strword[:8]):
return [
'批准文号有效期',
strword[re_regisnum_valid.search(strword).span()[1]:],
re_regisnum_valid.search(strword).group()
]
elif re_zhusong.search(strword[:index]):
return [
'主送', strword[re_zhusong.search(strword).span()[1]:],
re_zhusong.search(strword).group()
]
elif re_specification.search(strword[:self._short_index(strword)]):
return [
'规格', strword[re_specification.search(strword).span()[1]:],
re_specification.search(strword).group()
]
elif re_jixing.search(strword[:self._short_index(strword):]):
return [
'剂型', strword[re_jixing.search(strword).span()[1]:],
re_jixing.search(strword).group()
]
elif re_drug_class.search(strword[:index]):
return [
'药品分类', strword[re_drug_class.search(strword).span()[1]:],
re_drug_class.search(strword).group()
]
else:
return None
def regisration_deploy(self, imgs, id_code):
flag = 0
tmp = ''
for file in imgs:
file_name = file['imgpath'].split('/')[-1]
id = file['imgpath'].split('/')[-2]
if re.search(r'[\u4e00-\u9fa5]+', id):
dragname = re.search(r'[\u4e00-\u9fa5]+', id).group()
else:
dragname = re.search(r'[\u4e00-\u9fa5]+', file_name).group()
if dragname.find('(') > 0:
dragname = dragname[:dragname.find('(')]
if 'error_code' in file['imgjson']:
self.logmgr.error(file['imgpath'] + " : Img Size Error!")
continue
datas = file['imgjson']['words_result']
nums = file['imgjson']['words_result_num']
if len(datas) > 0 and nums > 0:
datadicttmp = self._recognize(datas, nums)
datadict = dict()
if '药品名称' in datadicttmp:
if re.match('[:：]', datadicttmp['药品名称']):
datadict['药品名称'] = datadicttmp['药品名称'][1:]
else:
datadict['药品名称'] = datadicttmp['药品名称']
if '剂型' in datadicttmp:
if re.match('[:：]', datadicttmp['剂型']):
datadict['剂型'] = datadicttmp['剂型'][1:]
else:
datadict['剂型'] = datadicttmp['剂型']
if '规格' in datadicttmp:
if re.match('[:：]', datadicttmp['规格']):
datadict['规格'] = datadicttmp['规格'][1:]
else:
datadict['规格'] = datadicttmp['规格']
if '生产厂家' in datadicttmp:
if re.match('[:：]', datadicttmp['生产厂家']):
datadict['生产厂家'] = datadicttmp['生产厂家'][1:]
else:
datadict['生产厂家'] = datadicttmp['生产厂家']
if '日期' in datadicttmp:
if re.match('[:：]', datadicttmp['日期']):
datadict['日期'] = datadicttmp['日期'][1:]
else:
datadict['日期'] = datadicttmp['日期']
if not datadict:
return 'None'
return datadict
def regisration(self, path, id_code):
flag = 0
temp = ''
for file in os.walk(path):
page = 1
jobdict = {}
for file_name in file[2]:
if '药品再注册批件' in file_name:
imgname = file_name.split('.')[0]
curpath = file[0].split('data')[1]
index = imgname.rfind('_')
id = curpath[curpath.rfind('\\') + 1:]
if re.search(r'[\u4e00-\u9fa5]+', id):
dragname = re.search(r'[\u4e00-\u9fa5]+', id).group()
else:
dragname = re.search(r'[\u4e00-\u9fa5]+',
file_name).group()
if dragname.find('(') > 0:
dragname = dragname[:dragname.find('(')]
datajson = self._load_json(file[0] + '\\' + file_name)
original_path = self.imgpath + '\\' + curpath + '\\' + imgname[:index
-
2] + '.' + 'pdf'
#服务器
jobdict['SER_IP'] = '10.67.28.8'
#job id
jobdict['JOB_ID'] = self._generatemd5(file[0] + imgname)
jobid = jobdict['JOB_ID']
jobdict['SRC_FILE_NAME'] = imgname[:index -
2] + '.' + 'pdf'
jobdict['SRC_FILE_PATH'] = original_path
#原文件
jobdict['CUT_FILE_NAME'] = imgname[:index] + '.' + imgname[
index:].split('_')[1]
#原路径
jobdict['CUT_FILE_PATH'] = 'G:\\IMG' + '\\' + curpath
#时间
jobdict['HANDLE_TIME'] = time.strftime(
"%Y-%m-%d %X", time.localtime())
#药品名
jobdict['DRUG_NAME'] = dragname
#影像件类型
jobdict['FILE_TYPE'] = '药品再注册批件'
#同一套影像件识别码
jobdict['ID_CODE'] = id_code
#分公司
jobdict['SRC_CO'] = curpath.split('\\')[1]
#源文件相对路径
jobdict[
'FILE_REL_PATH'] = '\\' + imgname[:
index] + '.' + imgname[
index:].split(
'_')[1]
#文件服务器域名
jobdict['SYS_URL'] = '10.67.28.8'
#页数
jobdict['PAGE_NUM'] = page
#文件ocr解析识别状态 fk sysparams
jobdict['OCR_STATE'] = 'T'
#备注说明
jobdict['REMARK'] = ''
#创建用户
jobdict['ADD_USER'] = '******'
#图片过大或者一些原因，没有识别出来就会有error_code字段
if 'error_code' in datajson:
jobdict['IS_TO_DB'] = 'F'
self.job.job_add(jobdict)
self.job.job_todb()
self.job.job_del()
self.logmgr.error(file[0] + '\\' + file_name +
": img size error!")
continue
datas = datajson['words_result']
nums = datajson['words_result_num']
flag = 1
#中间文件
jobdict['MID_FILE_NAME'] = file_name
#中间文件路径
jobdict['MID_FILE_PATH'] = file[0]
#评分
jobdict['OCR_SCORE'] = int(self._getscore(datas, nums))
#影像件内容是否入库
if len(datas) > 0 and nums > 0:
jobdict['IS_TO_DB'] = 'T'
else:
jobdict['IS_TO_DB'] = 'F'
#文件文本内容
jobdict['FILE_TEXT'] = self._middict(
datas, self.codepath + '\\middata\\' + curpath,
imgname)
###############
temp = jobdict['FILE_TEXT']
#jobdict['JOB_ID'] = self._generatemd5(jobdict['FILE_TEXT'])
###############
page += 1
self.job.job_add(jobdict)
self.job.job_todb()
self.job.job_del()
if flag:
if len(datas) > 0 and nums > 0:
datadicttmp = self._recognize(datas, nums)
datadict = dict()
if '药品名称' in datadicttmp:
if re.match('[:：]', datadicttmp['药品名称']):
datadict['药品名称'] = datadicttmp['药品名称'][1:]
else:
datadict['药品名称'] = datadicttmp['药品名称']
if '剂型' in datadicttmp:
if re.match('[:：]', datadicttmp['剂型']):
datadict['剂型'] = datadicttmp['剂型'][1:]
else:
datadict['剂型'] = datadicttmp['剂型']
if '规格' in datadicttmp:
if re.match('[:：]', datadicttmp['规格']):
datadict['规格'] = datadicttmp['规格'][1:]
else:
datadict['规格'] = datadicttmp['规格']
if '生产厂家' in datadicttmp:
if re.match('[:：]', datadicttmp['生产厂家']):
datadict['生产厂家'] = datadicttmp['生产厂家'][1:]
else:
datadict['生产厂家'] = datadicttmp['生产厂家']
if '日期' in datadicttmp:
if re.match('[:：]', datadicttmp['日期']):
datadict['日期'] = datadicttmp['日期'][1:]
else:
datadict['日期'] = datadicttmp['日期']
######################################增加部分###########################################
datadict['ID_CODE'] = id_code
datadict['REMARK'] = ''
datadict['ADD_USER'] = '******'
datadict['JOB_ID'] = self._generatemd5(temp)
######################################增加部分###########################################
print(datadict)
###########################
###########################
if not datadict:
nums = self._cleandata(datadict, datas, nums)
continue
try:
self._data_to_db('DRUGREGAPPROVAL', datadict)
nums = self._cleandata(datadict, datas, nums)
except Exception as e:
print('Error: ', e)
self.logmgr.error(file[0] + '\\' + file_name +
"insert error!! : " + str(e))
self._update_item('OCRWORKFILE', 'JOB_ID', jobid,
'IS_TO_DB', 'F')
nums = self._cleandata(datadict, datas, nums)
continue
import json
from DatabaseToolsNew import cxOracle
import re
from FindKeyword import findImportWords
import HowManyColumn4 as hmc
#import openpyxl
import xlwings as xw
import time
import hashlib
from log import LogMgr
from job import JobTable
import random
from json2word import json2word
from tool import Tools
logmgr = LogMgr()
'''
使用openpyxl太慢了,改用xlwings
wb = openpyxl.load_workbook('C:\\Users\\DevinChang\\Desktop\\四家分公司影印件清单_去重匹配版.xlsx')
sheets = wb.sheetnames
sheet = wb.get_sheet_by_name(sheets[0])
shopid = sheet['B']
name = sheet['C']
strength = sheet['D']
mfrs = sheet['F']
'''
class ProductionCertificate(Tools):
def __init__(self, jsonpath, imgpath):
Tools.__init__(self)
self.jsonpath = jsonpath
self.imgpath = imgpath
self.logmgr = LogMgr()
def generatemd5(strid):
md5 = hashlib.md5()
md5.update(strid.encode('utf-8'))
return md5.hexdigest()
def subfiledata(self,direction, parameter, boundary, datas):
leftdata = []
rightdata = []
for data in datas:
if direction == 1 or direction == 2:
if data['location'][parameter] >= boundary:
# 此处有bug
leftdata.append(data)
else:
rightdata.append(data)
else:
if data['location'][parameter] <= boundary:
leftdata.append(data)
else:
rightdata.append(data)
return leftdata+(rightdata)
def _productionCertificate(self, datas, nums):
"""
识别生产许可证
"""
keylist = []
datadict = {}#这里做了一点小改动！！！！！！！！！！！！！！！！！！！1
i = 0
flag = 0
for (word, i) in zip(datas, range(0, nums)):
list_result = self._judge_keywords(word['words'])
if list_result != None:
if list_result[0] in datadict and keylist[-1][0] != list_result[0]:
datadict[list_result[0]] += list_result[1]
flag = 1
else:
datadict[list_result[0]] = list_result[1]
flag = 1
keylist.append([list_result[0], list_result[2]])
else:
flag = 1
j = i
while j >= 0:
if not keylist:
break
if ("分类码" in keylist[-1][0]):
if re.match(r'[a-zA-z]+', word['words']):
flag = 1
else:
break
elif "有效期至" in keylist[-1][0]:
if re.match(r'[0-9]+年?[0-9]+月?[0-9]+日?', word['words']):
flag = 1
else:
break
# # 字段追加问题
# if re.match(r'.?[:：]', word['words'][:10]) and not re.match(r'质*量受*权人*',word['words']):
# if not re.match(r'质*量受*权人*', word['words']):
# flag = 0
# break
if flag:
if keylist[-1][1] in datas[j]['words']:
datadict[keylist[-1][0]] += word['words']
break
j -= 1
# datadict[list_result[0]] = list_result[1]
# flag = 1
return datadict
def _judge_keywords(self, strword):
'''判断关键字'''
# re_coname = re.compile(r"企业*名称*|企*业名*称")
# re_cernum = re.compile(r"证书*编号*|证*书编*号")
# re_addr = re.compile(r"地址")
# re_cerscope = re.compile(r"认证*范围*|认*证范*围")
# re_valid = re.compile(r"有效期至*|有效*期至")
# re_liceauth = re.compile(r"发证*机关*|发*证机*关")
# re_licedate = re.compile(r"发证*日期*|发*证日*期")
re_entname = re.compile(r"企业*名称*|企*业名*称")
re_regAddr = re.compile(r"注册*地址|注册地*址")
re_uscc = re.compile(r"社会*信用社*代*码|社*会信用*社*代码*")
re_legalReps = re.compile(r"法定*代表*人|法*定代表人*")
re_entPrincipal = re.compile(r"企.负责人|.业负*责人|企.负.人")
re_qcPrincipal = re.compile(r"质*量负责*人|质*量负责人*|.量负.人*")
re_vld = re.compile(r"有*效期*至|有效*期至")
re_supervisionDEP = re.compile(r"日常*监管*机构*|日*常监*管机*构")
re_supervisor = re.compile(r"日常*监管*人员*|日*常监*管人*员")
re_supervisorCT = re.compile(r"监督*举报*电话*|监*督举*报电*话")
re_licNO = re.compile(r"编号|編号|号：|号:|号")
re_licNO2 = re.compile(r"号")
re_cateCode = re.compile(r"分*类码")
re_prodAddrScope = re.compile(r"生*产地*址和生产*范*围|生*产*地址和*生*产范*围|.产.址和.产.围|生产地址.生产范.")
re_issueOrg = re.compile(r"发证机.|发证.关")
re_issuer = re.compile(r"签发*人")
re_issueDate = re.compile(r"发证*日*期")
re_kindsOfEnterprise = re.compile(r"企业*类型*")
re_useLimit = re.compile(r"此*复印件*仅*限用*于*")
re_qcLegal = re.compile(r"质*量受*权人*")
re_NO = re.compile(r"NO|N0")
re_authorizedDEPT = re.compile(r"国*家*食品*药品*监督*管*理局制*|.家*食.药.监督*.理局.")
re_country = re.compile(r"中华人民共和国")
re_kindsOfDocument = re.compile(r"药品生产许可证")
#这里将提取关键字段的长度延长到了12个,尽可能的将由于印章等造成的干扰降低
if len(strword) >= 4:
index = 6
else:
index = len(strword)
if (re.match(r'.+?(?:\:)', strword[:index])):
if re_entname.search(strword[:index]):
return ['企业名称_许可证', strword[re_entname.search(strword).span()[1]+1:], re_entname.search(strword).group()]
elif re_regAddr.search(strword[:index]):
return ['注册地址', strword[re_regAddr.search(strword).span()[1] + 1:], re_regAddr.search(strword).group()]
elif re_uscc.search(strword[:9]):
return ['社会信用社代码', strword[re_uscc.search(strword).span()[1] + 1:], re_uscc.search(strword).group()]
elif re_legalReps.search(strword[:7]):
return ['法定代表人', strword[re_legalReps.search(strword).span()[1] + 1:], re_legalReps.search(strword).group()]
elif re_entPrincipal.search(strword[:7]):
return ['企业负责人', strword[re_entPrincipal.search(strword).span()[1] + 1:], re_entPrincipal.search(strword).group()]
elif re_qcPrincipal.search(strword[:7]):
return ['质量负责人', strword[re_qcPrincipal.search(strword).span()[1] + 1:], re_qcPrincipal.search(strword).group()]
elif re_vld.search(strword[:index]):
return ['有效期至', strword[re_vld.search(strword).span()[1] + 1:], re_vld.search(strword).group()]
elif re_supervisionDEP.search(strword[:8]):
return ['日常监管机构', strword[re_supervisionDEP.search(strword).span()[1] + 1:], re_supervisionDEP.search(strword).group()]
elif re_supervisionDEP.search(strword[:8]):
return ['日常监管机构', strword[re_supervisionDEP.search(strword).span()[1] + 1:], re_supervisionDEP.search(strword).group()]
elif re_supervisor.search(strword[:8]):
return ['日常监管人员', strword[re_supervisor.search(strword).span()[1] + 1:], re_supervisor.search(strword).group()]
elif re_supervisorCT.search(strword[:8]):
return ['监督举报电话', strword[re_supervisorCT.search(strword).span()[1] + 1:], re_supervisorCT.search(strword).group()]
elif re_licNO.search(strword[:3]):
return ['许可证编号', strword[re_licNO.search(strword).span()[1] + 1:], re_licNO.search(strword).group()]
elif re_licNO2.search(strword[:1]):
return ['许可证编号', strword[re_licNO2.search(strword).span()[1] + 1:], re_licNO2.search(strword).group()]
elif re_cateCode.search(strword[:5]):
return ['分类码', strword[re_cateCode.search(strword).span()[1] + 1:], re_cateCode.search(strword).group()]
elif re_prodAddrScope.search(strword[:11]):
return ['生产地址和生产范围', strword[re_prodAddrScope.search(strword).span()[1] + 1:], re_prodAddrScope.search(strword).group()]
elif re_issueOrg.search(strword[:index]):
return ['发证机关', strword[re_issueOrg.search(strword).span()[1] + 1:], re_issueOrg.search(strword).group()]
elif re_issuer.search(strword[:5]):
return ['签发人', strword[re_issuer.search(strword).span()[1] + 1:], re_issuer.search(strword).group()]
elif re_issueDate.search(strword[:index]):
return ['发证日期', strword[re_issueDate.search(strword).span()[1] + 1:],re_issueDate.search(strword).group()]
elif re_kindsOfEnterprise.search(strword[:index]):
return ['企业类型', strword[re_kindsOfEnterprise.search(strword).span()[1] + 1:],re_kindsOfEnterprise.search(strword).group()]
elif re_useLimit.search(strword[:10]):
return ['此复印件仅限用于', strword[re_useLimit.search(strword).span()[1] + 1:],re_useLimit.search(strword).group()]
# elif re_qcLegal.search(strword[:index]):
# return ['质量受权人', strword[re_qcLegal.search(strword).span()[1] + 1:],re_qcLegal.search(strword).group()]
elif re_NO.search(strword[:3]):
return ['NO', strword[re_NO.search(strword).span()[1] + 1:],re_NO.search(strword).group()]
elif re_authorizedDEPT.search(strword[:13]):
return ['国家食品药品监督管理局制', strword[re_authorizedDEPT.search(strword).span()[1]:],re_authorizedDEPT.search(strword).group()]
elif re_country.search(strword[:8]):
return ['中华人民共和国', strword[re_country.search(strword).span()[1]:],re_country.search(strword).group()]
elif re_kindsOfDocument.search(strword[:8]):
return ['药品生产许可证', strword[re_kindsOfDocument.search(strword).span()[1]:],re_kindsOfDocument.search(strword).group()]
else:
return None
else:
if re_entname.search(strword[:index]):
return ['企业名称_许可证', strword[re_entname.search(strword).span()[1]+1:], re_entname.search(strword).group()]
elif re_regAddr.search(strword[:index]):
return ['注册地址', strword[re_regAddr.search(strword).span()[1] + 1:], re_regAddr.search(strword).group()]
elif re_uscc.search(strword[:9]):
return ['社会信用社代码', strword[re_uscc.search(strword).span()[1] + 1:], re_uscc.search(strword).group()]
elif re_legalReps.search(strword[:7]):
return ['法定代表人', strword[re_legalReps.search(strword).span()[1] + 1:], re_legalReps.search(strword).group()]
elif re_entPrincipal.search(strword[:7]):
return ['企业负责人', strword[re_entPrincipal.search(strword).span()[1] + 1:], re_entPrincipal.search(strword).group()]
elif re_qcPrincipal.search(strword[:7]):
return ['质量负责人', strword[re_qcPrincipal.search(strword).span()[1] + 1:], re_qcPrincipal.search(strword).group()]
elif re_vld.search(strword[:index]):
return ['有效期至', strword[re_vld.search(strword).span()[1] + 1:], re_vld.search(strword).group()]
elif re_supervisionDEP.search(strword[:8]):
return ['日常监管机构', strword[re_supervisionDEP.search(strword).span()[1] + 1:], re_supervisionDEP.search(strword).group()]
elif re_supervisionDEP.search(strword[:8]):
return ['日常监管机构', strword[re_supervisionDEP.search(strword).span()[1] + 1:], re_supervisionDEP.search(strword).group()]
elif re_supervisor.search(strword[:8]):
return ['日常监管人员', strword[re_supervisor.search(strword).span()[1] + 1:], re_supervisor.search(strword).group()]
elif re_supervisorCT.search(strword[:8]):
return ['监督举报电话', strword[re_supervisorCT.search(strword).span()[1] + 1:], re_supervisorCT.search(strword).group()]
elif re_licNO.search(strword[:3]):
return ['许可证编号', strword[re_licNO.search(strword).span()[1] + 1:], re_licNO.search(strword).group()]
elif re_cateCode.search(strword[:5]):
return ['分类码', strword[re_cateCode.search(strword).span()[1] + 1:], re_cateCode.search(strword).group()]
elif re_prodAddrScope.search(strword[:11]):
return ['生产地址和生产范围', strword[re_prodAddrScope.search(strword).span()[1] + 1:], re_prodAddrScope.search(strword).group()]
elif re_issueOrg.search(strword[:index]):
return ['发证机关', strword[re_issueOrg.search(strword).span()[1] + 1:], re_issueOrg.search(strword).group()]
elif re_issuer.search(strword[:5]):
return ['签发人', strword[re_issuer.search(strword).span()[1] + 1:], re_issuer.search(strword).group()]
elif re_issueDate.search(strword[:index]):
return ['发证日期', strword[re_issueDate.search(strword).span()[1] + 1:],re_issueDate.search(strword).group()]
elif re_kindsOfEnterprise.search(strword[:index]):
return ['企业类型', strword[re_kindsOfEnterprise.search(strword).span()[1] + 1:],re_kindsOfEnterprise.search(strword).group()]
elif re_useLimit.search(strword[:10]):
return ['此复印件仅限用于', strword[re_useLimit.search(strword).span()[1] + 1:],re_useLimit.search(strword).group()]
# elif re_qcLegal.search(strword[:index]):
# return ['质量受权人', strword[re_qcLegal.search(strword).span()[1] + 1:], re_qcLegal.search(strword).group()]
elif re_NO.search(strword[:3]):
return ['NO', strword[re_NO.search(strword).span()[1]+1:], re_NO.search(strword).group()]
elif re_authorizedDEPT.search(strword[:13]):
return ['国家食品药品监督管理局制', strword[re_authorizedDEPT.search(strword).span()[1]:],re_authorizedDEPT.search(strword).group()]
elif re_country.search(strword[:8]):
return ['中华人民共和国', strword[re_country.search(strword).span()[1]:], re_country.search(strword).group()]
elif re_kindsOfDocument.search(strword[:8]):
return ['药品生产许可证', strword[re_kindsOfDocument.search(strword).span()[1]:],re_kindsOfDocument.search(strword).group()]
else:
return None
def recognize_deploy(self, imgs, id_code):
nums = 0
flag = 0
temp = ''
datas = []
for file in imgs:
#提取药品名称
id = file['imgpath'].split('/')[-2]
file_name = file['imgpath'].split('/')[-1]
if re.search(r'[\u4e00-\u9fa5]+', id):
dragname = re.search(r'[\u4e00-\u9fa5]+', id).group()
else:
dragname = re.search(r'[\u4e00-\u9fa5]+', file_name).group()
if 'error_code' in file['imgjson']:
self.logmgr.error(file['imgpath'] + ' : ' + 'Size Error!')
#判别是否是多栏
try:
kindict = hmc.kinds(file['imgpath'], file['imgjson'])
except Exception as e:
self.logmgr.error(file['imgpath'] + ' : ' + 'Size Error!')
continue
print('Current processing: {}'.format(file['imgpath']))
#提取关键信息
datatmp = file['imgjson']['words_result']
nums += file['imgjson']['words_result_num']
if kindict['kinds'] == 2:
datas += subfiledata(kindict['direction'], kindict['parameter'], kindict['boundary'][0], datatmp)
elif kindict['kinds'] == 1:
datas += datatmp
if len(datas) > 0 and nums > 0:
datadict = self._productionCertificate(datas, nums)
if '企业类型' in datadict:
del datadict['企业类型']
if '此复印件仅限于' in datadict:
del datadict['次复印件仅限于']
if 'NO' in datadict:
del datadict['NO']
if '国家食品药品监督管理局制' in datadict:
del datadict['国家食品药品监督管理局制']
if '中华人民共和国' in datadict:
del datadict['中华人民共和国']
if '药品许可证' in datadict:
del datadict['药品许可证']
######################################增加部分###########################################
datadict['ID_CODE'] = id_code
datadict['REMARK'] = ''
datadict['ADD_USER'] = '******'
datadict['JOB_ID'] = self._generatemd5(temp)
######################################增加部分###########################################
if not datadict:
nums = self._cleandata(datadict, datas, nums)
return datadict
return datadict
#try:
# #self._data_to_db('DRUGMFRSCERT', datadict)
# nums = self._cleandata(datadict, datas, nums)
#except Exception as e:
# #self.logmgr.error(file[0] + '\\' + file_name + "insert error!! : " + str(e))
# #self._update_item('OCRWORKFILE','JOB_ID', jobid,'IS_TO_DB','F')
# nums = self._cleandata(datadict, datas, nums)
# continue
def recognize(self, path, id_code):
flag = 0
page = 0
temp =''
jobdict = {}
for file in os.walk(path):#这里将原来imgpath换成了 jsonpath
for file_name in file[2]:
if '生产许可证' in file_name:
jsonname = file_name.split('.')[0]
curpath = file[0].split('data')[1]
index = jsonname.rfind('_')
id = curpath[curpath.rfind('\\') + 1:]
if re.search(r'[\u4e00-\u9fa5]+', id):
dragname = re.search(r'[\u4e00-\u9fa5]+', id).group()
else:
dragname = re.search(r'[\u4e00-\u9fa5]+', file_name).group()
if dragname.find('(') > 0:
dragname = dragname[:dragname.find('(')]
jsonPath = file[0] + '\\' + file_name
datajson = self._load_json(file[0] + '\\' + file_name)
source_img_path = self.imgpath + curpath + '\\' + jsonname[:index] + '.' + jsonname[index:].split('_')[1]
original_path = self.imgpath + '\\' + curpath + '\\' + jsonname[:index - 2] + '.' + 'pdf'
#服务器
jobdict['SER_IP'] = '10.67.28.8'
#job id
jobdict['JOB_ID'] = self._generatemd5(file[0] + jsonname)
jobid = jobdict['JOB_ID']
jobdict['SRC_FILE_NAME'] = jsonname[:index - 2] + '.' + 'pdf'
jobdict['SRC_FILE_PATH'] = original_path
#原文件
jobdict['CUT_FILE_NAME'] = jsonname[:index] + '.' + jsonname[index:].split('_')[1]
#原路径
jobdict['CUT_FILE_PATH'] = 'G:\\IMG' + '\\' + curpath
#时间
jobdict['HANDLE_TIME'] = time.strftime("%Y-%m-%d %X", time.localtime())
#药品名
jobdict['DRUG_NAME'] = dragname
#影像件类型
jobdict['FILE_TYPE'] = '药品生产许可证'
#同一套影像件识别码
jobdict['ID_CODE'] = id_code
#分公司
jobdict['SRC_CO'] = curpath.split('\\')[1]
#源文件相对路径
jobdict['FILE_REL_PATH'] = '\\' + jsonname[:index] + '.' + jsonname[index:].split('_')[1]
#文件服务器域名
jobdict['SYS_URL'] = '10.67.28.8'
#页数
jobdict['PAGE_NUM'] = page
#文件ocr解析识别状态 fk sysparams
jobdict['OCR_STATE'] = 'T'
#备注说明
jobdict['REMARK'] = ''
#创建用户
jobdict['ADD_USER'] = '******'
# 图片过大或者一些原因，没有识别出来就会有error_code字段
if 'error_code' in datajson:
jobdict['IS_TO_DB'] = 'F'
self.job.job_add(jobdict)
self.job.job_todb()
self.job.job_del()
self.logmgr.error(file[0] + '\\' + file_name + ": img size error!")
continue
#source_img_path = 'img\\'+jsonname+'.jpg' #由于需要增加分栏的程序所以，需要图片的路径，但是目前这里面的路径存在一定的问题
# source_img_path = file[0] + '\\' + file_name
# original_path = path_root + '\\' + curpath + '\\' + imgname[:index - 2] + '.' + 'pdf'
# FIXME:换工作环境这里也得改！
try:
kindict = hmc.kinds(source_img_path, jsonPath)
except Exception as e:
self.logmgr.error(file[0] + '\\' + file_name + ':' + str(e))
continue
#index = jsonname.rfind('.')
# print('Current processing: {}'.format(source_img_path + '\\' +
# '\\' + imgname[:index] +
# '.' + imgname[index:].split('.')[1],
# file[0] + '\\' + file_name))
datas = datajson['words_result']
nums = datajson['words_result_num']
if kindict['kinds'] == 2:
datas = self.subfiledata(kindict['direction'], kindict['parameter'], kindict['boundary'][0],datas)
elif kindict['kinds'] == 1 or kindict['kinds'] == 0:
datas = datas
flag = 1
page += 1
#中间文件
jobdict['MID_FILE_NAME'] = file_name
#中间文件路径
jobdict['MID_FILE_PATH'] = file[0]
#评分
jobdict['OCR_SCORE'] = int(self._getscore(datas, nums))
#影像件内容是否入库
if len(datas) > 0 and nums > 0:
jobdict['IS_TO_DB'] = 'T'
else:
jobdict['IS_TO_DB'] = 'F'
#文件文本内容
jobdict['FILE_TEXT'] = self._middict(datas, self.codepath + '\\middata\\' + curpath, jsonname)
###############
temp = jobdict['FILE_TEXT']
#jobdict['JOB_ID'] = self._generatemd5(jobdict['FILE_TEXT'])
###############
page += 1
self.job.job_add(jobdict)
self.job.job_todb()
self.job.job_del()
if flag:
if len(datas) > 0 and nums > 0:
datadict = self._productionCertificate(datas, nums)
if '企业类型' in datadict:
del datadict['企业类型']
if '此复印件仅限于' in datadict:
del datadict['次复印件仅限于']
if 'NO' in datadict:
del datadict['NO']
if '国家食品药品监督管理局制' in datadict:
del datadict['国家食品药品监督管理局制']
if '中华人民共和国' in datadict:
del datadict['中华人民共和国']
if '药品许可证' in datadict:
del datadict['药品许可证']
print(source_img_path)
######################################增加部分###########################################
datadict['ID_CODE'] = id_code
datadict['REMARK'] = ''
datadict['ADD_USER'] = '******'
datadict['JOB_ID'] = self._generatemd5(temp)
######################################增加部分###########################################
print(datadict)
if not datadict:
nums = self._cleandata(datadict, datas, nums)
continue
try:
self._data_to_db('DRUGMFRSCERT', datadict)
nums = self._cleandata(datadict, datas, nums)
except Exception as e:
print('Error: ', e)
self.logmgr.error(file[0] + '\\' + file_name + "insert error!! : " + str(e))
self._update_item('OCRWORKFILE','JOB_ID', jobid,'IS_TO_DB','F')
nums = self._cleandata(datadict, datas, nums)
continue
class GMP(Tools):
"""
GMP证书的识别
"""
def __init__(self, imgpath):
Tools.__init__(self)
self.imgpath = imgpath
self.logmgr = LogMgr()
def _recognize(self,datas, nums):
"""
识别GMP证书, 程序的主逻辑
"""
keylist = []
datadict = dict()
for (word, i) in zip(datas, range(0, nums)):
'''
循环读识别出的数据，然后根据judge_keywords函数是否提取到了关键信息；
若提取到了，则保存到datadict中。
若未提取到，list_result为空。有两种情况，
1.这段信息不是我们所需要的。
2.这段信息是上个关键字的值。
然后执行else，进行更精确的判别。若是需归到上个字段，则循环递减，根据
keylist[1],也就是list_reault[2]是否出现再上面的某个字段。若有则追加。
'''
list_result = self._judge_keywords(word['words'])
if list_result != None:
if list_result[0] in datadict and keylist[-1][0] != list_result[0]:
datadict[list_result[0]] += list_result[1]
flag = 1
else:
datadict[list_result[0]] = list_result[1]
flag = 1
#保存关键字段的信息，以及这段信息原本关键字段的信息
keylist.append([list_result[0],list_result[2]])
else:
j = i
while j > 0:
if not keylist:
break
#FIXMEED:逻辑问题 4/10 DONE
if re.match(r'\s[a-zA-Z]+', word['words']):
break
#提取"有效期至"与"发证日期"字段
if re.match(r'\d{4}|\d{2}', word['words']):
if len(word['words']) <= 4:
break
elif '/' in word['words']:
if keylist[-1][0] == '发证机关':
datadict['发证日期'] = word['words']
keylist.append(['杂', '杂'])
break
if '有效期至' in datadict:
if re.search(r'\d{4}|\d{2}', datadict['有效期至']):
break
else:
datadict['有效期至'] = word['words']
break
if flag:
if keylist[-1][0] == '地址':
if i + 1 >= nums:
break
is_scope = self._judge_keywords(datas[i + 1]['words'])
if is_scope != None and is_scope[0] == '认证范围':
datadict['认证范围'] = word['words']
break
if keylist[-1][0] == '有效期至':
break
if keylist[-1][1] in datas[j]['words']:
datadict[keylist[-1][0]] += word['words']
break
j -= 1
return datadict
def _judge_keywords(self, strword):
'''
判断关键字,若识别到关键字，返回一个包含关键字的list。
$resultlist[0] -----要入库的关键字
$resultlist[1] -----提取到内容
$resultlist[2] -----需判断的信息中本来的关键字
如:'证书编号:H12345',resultlist = ['证书编号', 'H12345', '证书编号']
'证书号:H123', resultlist = ['证书编号', 'H123', '证书号']
'''
re_coname = re.compile(r"企业*名称*|企*业名*称")
re_cernum = re.compile(r"证书*编号*|证*书编*号")
re_addr = re.compile(r"地址")
re_cerscope = re.compile(r"认证*范围*|认*证范*围")
re_valid = re.compile(r"有效期至*|有效*期至")
re_liceauth = re.compile(r"发证*机关*|发*证机*关")
re_licedate = re.compile(r"发证*日期*|发*证日*期")
re_abandon = re.compile(r"经审*查")
if len(strword) >= 8:
index = 6
else:
index = len(strword)
if(re.match(r'.+?(?:\:)', strword[:index])):
if re_coname.search(strword[:index]):
return ['企业名称_GMP', strword[re_coname.search(strword).span()[1]:], re_coname.search(strword).group()]
elif re_cernum.search(strword[:index]):
return ['证书编号' , strword[re_cernum.search(strword).span()[1] + 1:], re_cernum.search(strword).group()]
elif re_addr.search(strword[:self._sort_index(strword)]):
return ['地址' , strword[re_addr.search(strword).span()[1]:],re_addr.search(strword).group()]
elif re_cerscope.search(strword[:index]):
return ['认证范围' , strword[re_cerscope.search(strword).span()[1]:],re_cerscope.search(strword).group()]
elif re_valid.search(strword[:index]):
return ['有效期至' , strword[re_valid.search(strword).span()[1]:],re_valid.search(strword).group()]
elif re_liceauth.search(strword[:index]):
return ['发证机关' , strword[re_liceauth.search(strword).span()[1]:],re_liceauth.search(strword).group()]
elif re_licedate.search(strword[:index]):
return ['发证时间' , strword[re_licedate.search(strword).span()[1]:],re_licedate.search(strword).group()]
else:
return None
else:
if re_coname.search(strword[:index]):
return ['企业名称_GMP', strword[re_coname.search(strword).span()[1]:], re_coname.search(strword).group()]
elif re_cernum.search(strword[:index]):
return ['证书编号' , strword[re_cernum.search(strword).span()[1] + 1:], re_cernum.search(strword).group()]
elif re_addr.search(strword[:self._sort_index(strword)]):
return ['地址' , strword[re_addr.search(strword).span()[1]:],re_addr.search(strword).group()]
elif re_cerscope.search(strword[:index]):
return ['认证范围' , strword[re_cerscope.search(strword).span()[1]:],re_cerscope.search(strword).group()]
elif re_valid.search(strword[:index]):
return ['有效期至' , strword[re_valid.search(strword).span()[1]:],re_valid.search(strword).group()]
elif re_liceauth.search(strword[:index]):
return ['发证机关' , strword[re_liceauth.search(strword).span()[1]:],re_liceauth.search(strword).group()]
elif re_licedate.search(strword[:index]):
return ['发证时间' , strword[re_licedate.search(strword).span()[1]:],re_licedate.search(strword).group()]
elif re_abandon.search(strword[:index]):
return ['经审查', strword[re_abandon.search(strword).span()[1]:], re_abandon.search(strword).group()]
else:
return None
def gmp_delploy(self, imgs, idcode):
flag = 0
tmp = ''
#datas = []
for file in imgs:
file_name = file['imgpath'].split('/')[-1]
id = file['imgpath'].split('/')[-2]
if re.search(r'[\u4e00-\u9fa5]+', id):
dragname = re.search(r'[\u4e00-\u9fa5]+', id).group()
else:
dragname = re.search(r'[\u4e00-\u9fa5]+', file_name).group()
if dragname.find('(') > 0:
dragname = dragname[:dragname.find('(')]
if 'error_code' in file['imgjson']:
self.logmgr.error(file['imgpath'] + " : Img Size Error!")
continue
datas = file['imgjson']['words_result']
nums = file['imgjson']['words_result_num']
if len(datas) > 0 and nums > 0:
datadicttmp = self._recognize(datas, nums)
datadict = dict()
if '企业名称_GMP' in datadicttmp:
if re.match('[:：]',datadicttmp['企业名称_GMP']):
datadict['企业名称_GMP'] = datadicttmp['企业名称_GMP'][1:]
else:
datadict['企业名称_GMP'] = datadicttmp['企业名称_GMP']
if '证书编号' in datadicttmp:
if re.match('[:：]',datadicttmp['证书编号']):
datadict['证书编号'] = datadicttmp['证书编号'][1:]
else:
datadict['证书编号'] = datadicttmp['证书编号']
if '地址' in datadicttmp:
if re.match('[:：]',datadicttmp['地址']):
datadict['地址'] = datadicttmp['地址'][1:]
else:
datadict['地址'] = datadicttmp['地址']
if '认证范围' in datadicttmp:
if re.match('[:：]',datadicttmp['认证范围']):
datadict['认证范围'] = datadicttmp['认证范围'][1:]
else:
datadict['认证范围'] = datadicttmp['认证范围']
if '有效期至' in datadicttmp:
if re.match('[:：]',datadicttmp['有效期至']):
datadict['有效期至'] = datadicttmp['有效期至'][1:]
else:
datadict['有效期至'] = datadicttmp['有效期至']
if '发证机关' in datadicttmp:
if re.match('[:：]',datadicttmp['发证机关']):
datadict['发证机关'] = datadicttmp['发证机关'][1:]
else:
datadict['发证机关'] = datadicttmp['发证机关']
if '发证日期' in datadicttmp:
if re.match('[:：]',datadicttmp['发证日期']):
datadict['发证日期'] = datadicttmp['发证日期'][1:]
else:
datadict['发证日期'] = datadicttmp['发证日期']
if '地址' not in datadict:
datadict['地址'] = ''
if '企业名称_GMP' not in datadict:
datadict['企业名称_GMP'] = ''
if re.search(r'.+公司.+',datadict['企业名称_GMP']):
datadict['地址'] = datadict['地址']+datadict['企业名称_GMP'].split('公司')[1]
datadict['企业名称_GMP'] = datadict['企业名称_GMP'].split('公司')[0]+'公司'
if not datadict:
nums = self._cleandata(datadict, datas, nums)
return datadict
return datadict
#try:
# #self._data_to_db('GMPCERT', datadict)
# nums = self._cleandata(datadict, datas, nums)
#except Exception as e:
# print('Error: ', e)
# #self._update_item('OCRWORKFILE','JOB_ID', jobid,'IS_TO_DB','F')
# self.logmgr.error(file[0] + '\\' + file_name + "insert error!! : " + str(e))
# nums = self._cleandata(datadict, datas, nums)
# return 'None'
def gmp(self, datapath, id_code):
flag = 0
temp = ''
for file in os.walk(datapath):
jobdict = {}
for file_name in file[2]:
page = 1
if 'GMP证书' in file_name:
imgname = file_name.split('.')[0]
curpath = file[0].split('data')[1]
index = imgname.rfind('_')
id = curpath[curpath.rfind('\\') + 1:]
if re.search(r'[\u4e00-\u9fa5]+', id):
dragname = re.search(r'[\u4e00-\u9fa5]+', id).group()
else:
dragname = re.search(r'[\u4e00-\u9fa5]+', file_name).group()
if dragname.find('(') > 0:
dragname = dragname[:dragname.find('(')]
#id_code = id[name_index_e - 1:]
datajson = self._load_json(file[0] + '\\' + file_name)
original_path = self.imgpath + '\\' + curpath + '\\' + imgname[:index - 2] + '.' + 'pdf'
#服务器
jobdict['SER_IP'] = '10.67.28.8'
#job id
jobdict['JOB_ID'] = self._generatemd5(file[0] + imgname)
jobid = jobdict['JOB_ID']
jobdict['SRC_FILE_NAME'] = imgname[:index - 2] + '.' + 'pdf'
jobdict['SRC_FILE_PATH'] = original_path
# jobdict['JOB_ID'] = self._generatemd5(jobdict[])
#原文件
jobdict['CUT_FILE_NAME'] = imgname[:index] + '.' + imgname[index:].split('_')[1]
#原路径
jobdict['CUT_FILE_PATH'] = 'G:\\IMG' + '\\' + curpath
#时间
jobdict['HANDLE_TIME'] = time.strftime("%Y-%m-%d %X", time.localtime())
#药品名
jobdict['DRUG_NAME'] = dragname
#影像件类型
jobdict['FILE_TYPE'] = 'GMP证书'
#同一套影像件识别码
jobdict['ID_CODE'] = id_code
#分公司
jobdict['SRC_CO'] = curpath.split('\\')[1]
#源文件相对路径
jobdict['FILE_REL_PATH'] = '\\' + imgname[:index] + '.' + imgname[index:].split('_')[1]
#文件服务器域名
jobdict['SYS_URL'] = '10.67.28.8'
#页数
jobdict['PAGE_NUM'] = page
#文件ocr解析识别状态 fk sysparams
jobdict['OCR_STATE'] = 'T'
#备注说明
jobdict['REMARK'] = ''
#创建用户
jobdict['ADD_USER'] = '******'
#图片过大或者一些原因，没有识别出来就会有error_code字段
if 'error_code' in datajson:
jobdict['IS_TO_DB'] = 'F'
self.job.job_add(jobdict)
self.job.job_todb()
self.job.job_del()
self.logmgr.error(file[0] + '\\' + file_name + ": img size error!")
continue
#source_img_path = imgpaht_root_desktop + '\\' + curpath + '\\' + imgname[:index] + '.' + imgname[index:].split('_')[1]
#try:
# kindict = hmc.kinds(source_img_path, datajson)
#except Exception as e:
# logmgr.error(file[0] + '\\' + file_name + ':' + str(e))
# continue
#print('Current processing: {}'.format(imgpaht_root_desktop + '\\' + curpath +
# '\\' + imgname[:index] +
# '.' + imgname[index:].split('_')[1],
# file[0] + '\\' + file_name))
datas = datajson['words_result']
nums = datajson['words_result_num']
flag = 1
#中间文件
jobdict['MID_FILE_NAME'] = file_name
#中间文件路径
jobdict['MID_FILE_PATH'] = file[0]
#评分
jobdict['OCR_SCORE'] = int(self._getscore(datas, nums))
#影像件内容是否入库
if len(datas) > 0 and nums > 0:
jobdict['IS_TO_DB'] = 'T'
else:
jobdict['IS_TO_DB'] = 'F'
#文件文本内容
jobdict['FILE_TEXT'] = self._middict(datas, self.codepath + '\\middata\\' + curpath, imgname)
###############
temp = jobdict['FILE_TEXT']
#jobdict['JOB_ID'] = self._generatemd5(jobdict['FILE_TEXT'])
###############
try:
self.job.job_add(jobdict)
except Exception:
self.job.update_item('JOB_ID', jobid, 'IS_TO_DB', 'F')
self.job.job_todb()
self.job.job_del()
if flag:
if len(datas) > 0 and nums > 0:
datadicttmp = self._recognize(datas, nums)
datadict = dict()
if '企业名称_GMP' in datadicttmp:
if re.match('[:：]',datadicttmp['企业名称_GMP']):
datadict['企业名称_GMP'] = datadicttmp['企业名称_GMP'][1:]
else:
datadict['企业名称_GMP'] = datadicttmp['企业名称_GMP']
if '证书编号' in datadicttmp:
if re.match('[:：]',datadicttmp['证书编号']):
datadict['证书编号'] = datadicttmp['证书编号'][1:]
else:
datadict['证书编号'] = datadicttmp['证书编号']
if '地址' in datadicttmp:
if re.match('[:：]',datadicttmp['地址']):
datadict['地址'] = datadicttmp['地址'][1:]
else:
datadict['地址'] = datadicttmp['地址']
if '认证范围' in datadicttmp:
if re.match('[:：]',datadicttmp['认证范围']):
datadict['认证范围'] = datadicttmp['认证范围'][1:]
else:
datadict['认证范围'] = datadicttmp['认证范围']
if '有效期至' in datadicttmp:
if re.match('[:：]',datadicttmp['有效期至']):
datadict['有效期至'] = datadicttmp['有效期至'][1:]
else:
datadict['有效期至'] = datadicttmp['有效期至']
if '发证机关' in datadicttmp:
if re.match('[:：]',datadicttmp['发证机关']):
datadict['发证机关'] = datadicttmp['发证机关'][1:]
else:
datadict['发证机关'] = datadicttmp['发证机关']
if '发证日期' in datadicttmp:
if re.match('[:：]',datadicttmp['发证日期']):
datadict['发证日期'] = datadicttmp['发证日期'][1:]
else:
datadict['发证日期'] = datadicttmp['发证日期']
######################################增加部分###########################################
datadict['ID_CODE']=id_code
datadict['REMARK']=''
datadict['ADD_USER']='******'
datadict['JOB_ID'] = self._generatemd5(temp)
if '地址' not in datadict:
datadict['地址'] = ''
if '企业名称_GMP' not in datadict:
datadict['企业名称_GMP'] = ''
if re.search(r'.+公司.+',datadict['企业名称_GMP']):
datadict['地址'] = datadict['地址']+datadict['企业名称_GMP'].split('公司')[1]
datadict['企业名称_GMP'] = datadict['企业名称_GMP'].split('公司')[0]+'公司'
######################################增加部分###########################################
print(datadict)
if not datadict:
nums = self._cleandata(datadict, datas, nums)
continue
try:
self._data_to_db('GMPCERT', datadict)
nums = self._cleandata(datadict, datas, nums)
except Exception as e:
print('Error: ', e)
self._update_item('OCRWORKFILE','JOB_ID', jobid,'IS_TO_DB','F')
self.logmgr.error(file[0] + '\\' + file_name + "insert error!! : " + str(e))
nums = self._cleandata(datadict, datas, nums)
continue
class MyOcr(object):
"""
文字识别
@app_id @api_key @secret_key 为百度ai平台上申请的值
@typeid 精度的选择
1--调用通用文字识别
2--含位置信息的通用文字识别
3--高精度的文字识别
4--含位置信息的高精度文字识别
"""
def __init__(self, typeid, app_id = APP_ID, api_key = API_KEY, secret_key = SECRET_KEY):
self.client = AipOcr(app_id, api_key, secret_key)
#self.client = AipOcr(appid[1], apikey[1], secretkey[1])
self.typeid = typeid
self.codepath = os.path.dirname(__file__)
self.datapath = self.codepath + '\data'
os.makedirs(self.datapath, exist_ok=True)
self.log = LogMgr()
def _get_file_content(self, filePath):
"""读取图片"""
with open(filePath, 'rb') as fp:
return fp.read()
def _write_json_file(self, filepath, data):
"""写入json文件"""
with open(filepath, 'w', encoding = 'utf-8') as fw:
fw.write(json.dumps(data, ensure_ascii=False))
def _list_custom(self, path):
root = os.listdir(path)
return os.listdir(path + '\\' + root[0]), path + '\\' + root[0]
def ocr_deploy(self, rec_dict):
files = rec_dict['files']
#ocr所需的参数
options = {}
options["detect_direction"] = "true"
options["detect_language"] = "true"
options["probability"] = "true"
#dirlist = os.listdir(imgpath)
#dirlist, root = self._list_custom(imgpath)
for file in files:
if re.search(r'进口注册证|GMP|说明书|药品再注册批件|营业执照|生产许可证|进口药品许可证|进口药品注册证', file['type']):
for img in file['imgs']:
print('Current img: {}'.format(img['imgpath']))
try:
data = self.client.accurate(base64.b64decode(bytes(img['base64'], encoding='utf-8')), options)
except Exception as e:
print('Error: ', e)
self.log.error(img['imgpath'] + "Error! : " + str(e))
continue
img.update({"imgjson" : data})
return rec_dict
def _ocr(self, imgpath):
"""
识别img文件下的图片
@输出json数据，保存到data文件夹下
"""
#imgpath = self.codepath + '\IMG'+'\国控天星'
#FIXME:电脑环境不同，路径也不一样，切换环境的话要修改路径
#imgpath = 'F:\IMG'
#imgpath = r'D:\IMG'
options = {}
options["detect_direction"] = "true"
options["detect_language"] = "true"
options["probability"] = "true"
#FIXME:图片路径需改
dirlist = os.listdir(imgpath)
root = imgpath
#dirlist, root = self._list_custom(imgpath)
for file in os.walk(imgpath):
for file_name in file[2]:
if re.search(r'进口注册证|GMP|说明书|药品再注册批件|营业执照|生产许可证|进口药品许可证', file_name):
if '备案' in file_name:
continue
if os.path.isdir(file[0] + '\\' + file_name):
continue
if not re.match(r'[jJ][pP][gG]', file_name[-3:]):
continue
datafilepath = self.datapath + file[0].split('IMG')[1]
if not os.path.exists(datafilepath):
os.makedirs(datafilepath)
img = self._get_file_content(file[0] + '\\' + file_name)
if file_name[:-4].find('.'):
file_name = file_name[:-4].replace('.', '') + file_name[-4:]
try:
prefix,suffix = file_name.split('.')
except Exception as e:
print('split error: {}\ncurrent file: {}'.format(e, file[0] + '\\' + file_name))
self.log.error(file[0] + '\\' + file_name + " Error!! : " + str(e))
continue
#判断文件是否存在
if os.path.isfile((datafilepath +'\{}.json').format(prefix + '_' + suffix)):
continue
print('Current img: {}'.format(file[0] + '\\' + file_name))
#FIXME:
testdict = dict()
testdict['base64'] = str(base64.b64encode(img), 'utf-8')
#img_test = str.encode(testdict['base64'])
#self._write_json_file('F:\\IMG\\11A0015\\test.json', str(img))
try:
if self.typeid == 1:
data = self.client.basicGeneral(img, options)
elif self.typeid == 2:
data = self.client.general(img, options)
elif self.typeid == 3:
data = self.client.basicAccurate(base64.b64decode(bytes(testdict['base64'], encoding='utf-8')), options)
elif self.typeid == 4:
data = self.client.accurate(img, options)
except Exception as e:
print('Error: ', e)
self.log.error(file[0] + '\\' + file_name + " Error!! : " + str(e))
continue
self._write_json_file((datafilepath +'\{}.json').format(prefix + '_' + suffix), data)
def _write_dict(self):
files = os.listdir(self.datapath)
for file in files:
format_data = introduction.introduction(self.datapath + '\\' + file)
print(format_data)
def pdf2img(self):
"""pdf转jpg"""
file_dir = self.codepath + '/PDF/说明书/'
save_dir = self.codepath + '/IMG/图片/'
for files in os.walk(file_dir):
for file_name in files[2]:
file_path = file_dir
[file_name_prefix, file_name_suffix] = file_name.split('.')
file = file_dir + file_name
with(Image(filename=file, resolution=300)) as img:
images = img.sequence
pages = len(images)
for i in range(pages):
images[i].type = 'truecolor'
save_name = save_dir + file_name_prefix + str(i) + '.jpg'
Image(images[i]).save(filename=save_name)
def run(self, imgpath):
"""入口函数"""
print('********Start Identify********')
self._ocr(imgpath)
print('********End********')
# -*- coding : utf-8 -*-
import os
import json
from DatabaseToolsNew import cxOracle
import re
from FindKeyword import findImportWords
import HowManyColumn4 as hmc
#import openpyxl
import xlwings as xw
from log import LogMgr
logmgr = LogMgr()
'''
使用openpyxl太慢了,改用xlwings
wb = openpyxl.load_workbook('C:\\Users\\DevinChang\\Desktop\\四家分公司影印件清单_去重匹配版.xlsx')
sheets = wb.sheetnames
sheet = wb.get_sheet_by_name(sheets[0])
shopid = sheet['B']
name = sheet['C']
strength = sheet['D']
mfrs = sheet['F']
'''
def load_excel(excel):
wb = xw.Book(excel)
sheet = wb.sheets[0]
shopid = sheet['B:B'].value
name = sheet['C:C'].value
strength = sheet['D:D'].value