import os
import re
import time
import shutil
import base64
import signal
import requests
import traceback
from datetime import datetime
from django.core.management import BaseCommand
from multiprocessing import Process
from settings import conf
from common.mixins import LoggerMixin
from common.tools.pdf_to_img import PDFHandler
from apps.doc import consts
from apps.doc.exceptions import OCR1Exception
from apps.doc.ocr.wb import BSWorkbook
class Command(BaseCommand, LoggerMixin):
def __init__(self):
super().__init__()
self.log_base = '[folder ocr process]'
# 处理文件开关
self.switch = True
# 睡眠时间
self.sleep_time = float(conf.SLEEP_SECOND_FOLDER)
# input folder
self.input_dirs = conf.get_namespace('INPUT_DIR_')
# ocr相关
self.ocr_url = conf.OCR_URL_FOLDER
# 优雅退出信号：15
signal.signal(signal.SIGTERM, self.signal_handler)
def signal_handler(self, sig, frame):
self.switch = False # 停止处理文件
def license1_process(self, ocr_data, license_summary, classify, res_list, pno, ino, part_idx):
# 类别：'0'身份证， '1'居住证
license_data = ocr_data.get('data', [])
if not license_data:
res_list.append((pno, ino, part_idx, consts.RES_SUCCESS_EMPTY))
return
res_list.append((pno, ino, part_idx, consts.RES_SUCCESS))
license_summary.setdefault(classify, []).extend(license_data)
@staticmethod
def parse_img_path(img_path):
# 'page_{0}_img_{1}.{2}'.format(pno, img_index, ext)
img_name, _ = os.path.splitext(os.path.basename(img_path))
if re.match(r'page_\d+_img_\d+', img_name):
part_list = img_name.split('_')
return img_name, int(part_list[1])+1, int(part_list[3])+1
else:
return img_name, 1, 1
@staticmethod
def get_path(name, img_output_dir, wb_output_dir, pdf_output_dir):
time_stamp = datetime.now().strftime('%Y-%m-%d_%H:%M:%S')
new_name = '{0}_{1}'.format(time_stamp, name)
img_save_path = os.path.join(img_output_dir, new_name)
pdf_save_path = os.path.join(pdf_output_dir, new_name)
excel_name = '{0}.xlsx'.format(os.path.splitext(new_name)[0])
excel_path = os.path.join(wb_output_dir, excel_name)
return img_save_path, excel_path, pdf_save_path
def res_process(self, all_res, classify, excel_path):
try:
license_summary = {}
res_list = []
if not all_res:
return
else:
for img_path, ocr_res in all_res.items():
img_name, pno, ino = self.parse_img_path(img_path)
part_idx = 1
if isinstance(ocr_res, dict):
if ocr_res.get('code') == 1:
data_list = ocr_res.get('data', [])
if isinstance(data_list, list):
for part_idx, ocr_data in enumerate(data_list):
part_idx = part_idx + 1
self.license1_process(ocr_data, license_summary, classify, res_list, pno, ino, part_idx)
else:
res_list.append((pno, ino, part_idx, consts.RES_FAILED_3))
else:
res_list.append((pno, ino, part_idx, consts.RES_FAILED))
else:
res_list.append((pno, ino, part_idx, consts.RES_FAILED))
wb = BSWorkbook(set(), set(), set(), set())
wb.simple_license_rebuild(license_summary, consts.DOC_SCHEME_LIST[0])
wb.save(excel_path)
except Exception as e:
self.folder_log.error('{0} [wb build error] [path={1}] [error={2}]'.format(
self.log_base, excel_path, traceback.format_exc()))
def ocr_process(self, img_path, classify):
if os.path.exists(img_path):
# TODO 图片验证
with open(img_path, 'rb') as f:
base64_data = base64.b64encode(f.read())
# 获取解码后的base64值
file_data = base64_data.decode()
json_data = {
"file": file_data,
"classify": classify
}
for times in range(consts.RETRY_TIMES):
try:
start_time = time.time()
ocr_response = requests.post(self.ocr_url, json=json_data)
if ocr_response.status_code != 200:
raise OCR1Exception('{0} ocr status code: {0}'.format(self.log_base, ocr_response.status_code))
except Exception as e:
self.folder_log.warn('{0} [ocr failed] [times={1}] [img_path={2}] [error={3}]'.format(
self.log_base, times, img_path, traceback.format_exc()))
else:
ocr_res = ocr_response.json()
end_time = time.time()
speed_time = int(end_time - start_time)
self.folder_log.info('{0} [ocr success] [img={1}] [res={2}] [speed_time={3}]'.format(
self.log_base, img_path, ocr_res, speed_time))
return ocr_res
else:
self.folder_log.warn('{0} [ocr failed] [img_path={1}]'.format(self.log_base, img_path))
def pdf_process(self, name, path, classify, img_output_dir, wb_output_dir, pdf_output_dir):
if os.path.exists(path):
try:
img_save_path, excel_path, pdf_save_path = self.get_path(name, img_output_dir, wb_output_dir, pdf_output_dir)
self.folder_log.info('{0} [pdf to img start] [path={1}]'.format(self.log_base, path))
pdf_handler = PDFHandler(path, img_save_path)
pdf_handler.extract_image()
self.folder_log.info('{0} [pdf to img end] [path={1}]'.format(self.log_base, path))
except Exception as e:
self.folder_log.error('{0} [pdf to img error] [path={1}] [error={2}]'.format(
self.log_base, path, traceback.format_exc()))
else:
all_res = {}
for img_path in pdf_handler.img_path_list:
ocr_res = self.ocr_process(img_path, classify)
all_res[img_path] = ocr_res
self.res_process(all_res, classify, excel_path)
shutil.move(path, pdf_save_path)
def img_process(self, name, path, classify, wb_output_dir, img_output_dir, pdf_output_dir):
ocr_res = self.ocr_process(path, classify)
all_res = {path: ocr_res}
try:
img_save_path, excel_path, _ = self.get_path(name, img_output_dir, wb_output_dir, pdf_output_dir)
except Exception as e:
self.folder_log.error('{0} [get path error] [path={1}] [error={2}]'.format(
self.log_base, path, traceback.format_exc()))
else:
self.res_process(all_res, classify, excel_path)
shutil.move(path, img_save_path)
def folder_process(self, input_dir, classify):
while not os.path.isdir(input_dir):
self.folder_log.info('{0} [input dir is not dir] [input_dir={1}]'.format(self.log_base, input_dir))
print(self.switch)
if self.switch:
time.sleep(self.sleep_time)
continue
else:
return
output_dir = os.path.join(os.path.dirname(input_dir), 'Output')
img_output_dir = os.path.join(output_dir, 'image')
wb_output_dir = os.path.join(output_dir, 'excel')
pdf_output_dir = os.path.join(output_dir, 'pdf')
failed_output_dir = os.path.join(output_dir, 'failed')
os.makedirs(output_dir, exist_ok=True)
os.makedirs(img_output_dir, exist_ok=True)
os.makedirs(wb_output_dir, exist_ok=True)
os.makedirs(pdf_output_dir, exist_ok=True)
os.makedirs(failed_output_dir, exist_ok=True)
while self.switch:
if not os.path.isdir(input_dir):
self.folder_log.info('{0} [input dir is not dir] [input_dir={1}]'.format(self.log_base, input_dir))
time.sleep(self.sleep_time)
continue
# 1. 从input dir获取pdf or image
list_dir = os.listdir(input_dir)
if not list_dir:
self.folder_log.info('{0} [input dir empty] [input_dir={1}]'.format(self.log_base, input_dir))
time.sleep(self.sleep_time)
continue
for name in list_dir:
try:
path = os.path.join(input_dir, name)
if os.path.isfile(path):
self.folder_log.info('{0} [file start] [path={1}]'.format(self.log_base, path))
if name.endswith('.pdf'):
self.pdf_process(name, path, classify, img_output_dir, wb_output_dir, pdf_output_dir)
else:
self.img_process(name, path, classify, wb_output_dir, img_output_dir, pdf_output_dir)
self.folder_log.info('{0} [file end] [path={1}]'.format(self.log_base, path))
except Exception as e:
try:
path = os.path.join(input_dir, name)
self.folder_log.error('{0} [file error] [path={1}] [error={2}]'.format(self.log_base, path,
traceback.format_exc()))
shutil.move(path, failed_output_dir)
continue
except Exception as e:
self.folder_log.error('{0} [file error] [error={1}]'.format(
self.log_base, traceback.format_exc()))
continue
def handle(self, *args, **kwargs):
process_list = []
for classify_idx, input_dir in self.input_dirs.items():
classify = int(classify_idx.split('_')[0])
process = Process(target=self.folder_process, args=(input_dir, classify))
process_list.append(process)
for p in process_list:
p.start()
for p in process_list:
p.join()
self.folder_log.info('{0} [stop safely]'.format(self.log_base))