import os import re import time import shutil import base64 import signal import requests import traceback from PIL import Image 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, OCR4Exception from apps.doc.ocr.wb import BSWorkbook class TIFFHandler: def __init__(self, path, img_save_path): self.path = path self.img_save_path = img_save_path self.img_path_list = [] def extract_image(self): os.makedirs(self.img_save_path, exist_ok=True) tiff = Image.open(self.path) tiff.load() for i in range(tiff.n_frames): try: save_path = os.path.join(self.img_save_path, 'page_{0}.jpeg'.format(i)) tiff.seek(i) tiff.save(save_path) self.img_path_list.append(save_path) except EOFError: break 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 self.ocr_url_4 = conf.IC_URL # 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, img_path): # 类别:'0'身份证, '1'居住证 license_data = ocr_data.get('data', []) if not license_data: return if classify == consts.MVC_CLASSIFY: # 车辆登记证 3/4页结果整合 for mvc_dict in license_data: try: mvc_page = mvc_dict.pop('page') except Exception as e: pass else: if mvc_page == 'VehicleRegArea': mvc_res = mvc_dict.pop('results', {}) mvc_dict['机动车登记证书编号'] = mvc_res.get('register_no', {}).get('words', '') for register_info in mvc_res.get('register_info', []): for detail_dict in register_info.get('details', {}).values(): mvc_dict.setdefault(detail_dict.get('chinese_key', '未知'), []).append( detail_dict.get('words', '')) del mvc_res if classify == consts.IC_CLASSIFY: for id_card_dict in license_data: try: base64_img = id_card_dict.pop('base64_img') except Exception as e: continue else: card_type = -1 json_data_4 = { 'mode': 1, 'user_info': { 'image_content': base64_img, }, 'options': { 'distinguish_type': 1, 'auto_rotate': True, }, } for times in range(consts.RETRY_TIMES): try: start_time = time.time() ocr_4_response = requests.post(self.ocr_url_4, json=json_data_4) if ocr_4_response.status_code != 200: raise OCR4Exception('ocr_4 status code: {0}'.format(ocr_4_response.status_code)) except Exception as e: self.folder_log.warn( '{0} [ocr_4 failed] [times={1}] [img_path={2}] [error={3}]'.format( self.log_base, times, img_path, traceback.format_exc())) else: ocr_4_res = ocr_4_response.json() end_time = time.time() speed_time = int(end_time - start_time) if ocr_4_res.get('code') == 0 and ocr_4_res.get('result', {}).get('rtn') == 0: card_type = ocr_4_res.get('result', {}).get( 'idcard_distinguish_result', {}).get('result', -1) self.folder_log.info( '{0} [ocr_4 success] [img_path={1}] [speed_time={2}]'.format( self.log_base, img_path, speed_time)) break else: self.folder_log.warn( '{0} [ocr_4 failed] [img_path={1}]'.format(self.log_base, img_path)) id_card_dict[consts.IC_TURE_OR_FALSE] = consts.IC_RES_MAPPING.get(card_type) 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 = {} 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 ocr_data in data_list: # part_idx = part_idx + 1 self.license1_process(ocr_data, license_summary, classify, img_path) wb = BSWorkbook(set(), set(), set(), set(), set()) wb.simple_license_rebuild(license_summary, consts.DOC_SCHEME_LIST[0]) wb.remove_base_sheet() 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: {1}'.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 images_process(self, img_path_list, classify, excel_path): all_res = {} for img_path in 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) 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())) raise e else: self.images_process(pdf_handler.img_path_list, classify, excel_path) shutil.move(path, pdf_save_path) def tif_process(self, name, path, classify, img_output_dir, wb_output_dir, tiff_output_dir): if os.path.exists(path): try: img_save_path, excel_path, tiff_save_path = self.get_path(name, img_output_dir, wb_output_dir, tiff_output_dir) self.folder_log.info('{0} [tiff to img start] [path={1}]'.format(self.log_base, path)) tiff_handler = TIFFHandler(path, img_save_path) tiff_handler.extract_image() self.folder_log.info('{0} [tiff to img end] [path={1}]'.format(self.log_base, path)) except Exception as e: self.folder_log.error('{0} [tiff to img error] [path={1}] [error={2}]'.format( self.log_base, path, traceback.format_exc())) raise e else: self.images_process(tiff_handler.img_path_list, classify, excel_path) shutil.move(path, tiff_save_path) def img_process(self, name, path, classify, wb_output_dir, img_output_dir, pdf_output_dir): 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: ocr_res = self.ocr_process(path, classify) all_res = {path: ocr_res} 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)) 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') tiff_output_dir = os.path.join(output_dir, 'tiff') 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(tiff_output_dir, exist_ok=True) os.makedirs(failed_output_dir, exist_ok=True) os_error_filename_set = set() 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 and len(os_error_filename_set) == 0: self.folder_log.info('{0} [input dir empty] [input_dir={1}]'.format(self.log_base, input_dir)) time.sleep(self.sleep_time) continue all_file_set = set(list_dir) true_file_set = all_file_set - os_error_filename_set if len(true_file_set) == 0 and len(os_error_filename_set) > 0: true_file_set.add(os_error_filename_set.pop()) for name in true_file_set: path = os.path.join(input_dir, name) try: if os.path.isfile(path): self.folder_log.info('{0} [file start] [path={1}]'.format(self.log_base, path)) if name.endswith('.pdf') or name.endswith('.PDF'): self.pdf_process(name, path, classify, img_output_dir, wb_output_dir, pdf_output_dir) elif name.endswith('.tif') or name.endswith('.TIF'): self.tif_process(name, path, classify, img_output_dir, wb_output_dir, tiff_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)) else: self.folder_log.info('{0} [path is dir] [path={1}]'.format(self.log_base, input_dir)) failed_path = os.path.join(failed_output_dir, '{0}_{1}'.format(time.time(), name)) shutil.move(path, failed_path) except OSError: os_error_filename_set.add(name) self.folder_log.error('{0} [os error] [path={1}] [error={2}]'.format( self.log_base, path, traceback.format_exc())) except Exception as e: try: self.folder_log.error('{0} [file error] [path={1}] [error={2}]'.format(self.log_base, path, traceback.format_exc())) failed_path = os.path.join(failed_output_dir, '{0}_{1}'.format(time.time(), name)) shutil.move(path, failed_path) except Exception as e: os_error_filename_set.add(name) self.folder_log.error('{0} [file move error] [path={1}] [error={2}]'.format( self.log_base, path, traceback.format_exc())) 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))