Skip to content

周伟奇 / bmw-ocr

Go to a project
Toggle navigation
Toggle navigation pinning
24ed4007 by 周伟奇
Options

Merge branch 'feature/slice4' into feature/1119

2 parents ebc29c21 a10caee4
Inline Side-by-side
Showing 1 changed file with 67 additions and 4 deletions
src/apps/doc/management/commands/ocr_process.py
...@@ -64,6 +64,7 @@ class Command(BaseCommand, LoggerMixin): ...@@ -64,6 +64,7 @@ class Command(BaseCommand, LoggerMixin):
64 # ocr相关 64 # ocr相关
65 self.ocr_1_urls = conf.get_namespace('OCR_URL_1_') 65 self.ocr_1_urls = conf.get_namespace('OCR_URL_1_')
66 self.ocr_url_2 = conf.OCR_URL_2 66 self.ocr_url_2 = conf.OCR_URL_2
67 self.ocr_url_2_bc = conf.OCR_URL_2_BC
67 self.ocr_url_3 = conf.BC_URL 68 self.ocr_url_3 = conf.BC_URL
68 self.ocr_url_4 = conf.IC_URL 69 self.ocr_url_4 = conf.IC_URL
69 # EDMS web_service_api 70 # EDMS web_service_api
...@@ -1212,11 +1213,12 @@ class Command(BaseCommand, LoggerMixin): ...@@ -1212,11 +1213,12 @@ class Command(BaseCommand, LoggerMixin):
1212 "pid": str(pid), 1213 "pid": str(pid),
1213 "filedata": file_data 1214 "filedata": file_data
1214 } 1215 }
1216 ocr_url_2 = self.ocr_url_2_bc if classify == consts.BC_CLASSIFY else self.ocr_url_2
1215 1217
1216 for times in range(consts.RETRY_TIMES): 1218 for times in range(consts.RETRY_TIMES):
1217 try: 1219 try:
1218 start_time = time.time() 1220 start_time = time.time()
1219 ocr_2_response = requests.post(self.ocr_url_2, data=json_data_2) 1221 ocr_2_response = requests.post(ocr_url_2, data=json_data_2)
1220 if ocr_2_response.status_code != 200: 1222 if ocr_2_response.status_code != 200:
1221 raise OCR2Exception('ocr_2 status code: {0}'.format(ocr_2_response.status_code)) 1223 raise OCR2Exception('ocr_2 status code: {0}'.format(ocr_2_response.status_code))
1222 except Exception as e: 1224 except Exception as e:
...@@ -1232,10 +1234,69 @@ class Command(BaseCommand, LoggerMixin): ...@@ -1232,10 +1234,69 @@ class Command(BaseCommand, LoggerMixin):
1232 self.log_base, img_path, speed_time)) 1234 self.log_base, img_path, speed_time))
1233 1235
1234 if classify == consts.BC_CLASSIFY: 1236 if classify == consts.BC_CLASSIFY:
1237 new_ocr_2_res = {}
1238 position_dict = {}
1239 get_info = False
1240 bc_image_data = ''
1241 position = {}
1242 angle = 0
1243 for page_info in ocr_2_res.get('PageInfo', []):
1244 if get_info:
1245 break
1246 if page_info.get('ErrorCode', 1) in consts.SUCCESS_CODE_SET:
1247 for bc_res_all in page_info.get('Result', []):
1248 if get_info:
1249 break
1250 if bc_res_all.get('ErrorCode', 1) in consts.SUCCESS_CODE_SET:
1251 for bc_res in bc_res_all.get('ResultList', []):
1252 if get_info:
1253 break
1254 get_info = True
1255 bc_image_data = bc_res.get('image_data', '')
1256 position = bc_res.get('position', {})
1257 angle = bc_res.get('angle', 0)
1258 for field_info in bc_res.get('FieldList', []):
1259 new_ocr_2_res[field_info.get('key', '')] = field_info.get('value', '')
1260 position_dict[field_info.get('key', '')] = {
1261 consts.FIELD_POSITION_KEY: field_info.get('position', {}),
1262 consts.FIELD_QUAD_KEY: field_info.get('quad', []),
1263 }
1264
1265 if get_info:
1266 pre, suf = os.path.splitext(img_path)
1267 if len(bc_image_data) > 0:
1268 position = {}
1269 angle = 0
1270 section_img_path = '{0}_{1}_0{2}'.format(pre, part_idx, suf)
1271 try:
1272 with open(section_img_path, "wb") as fh:
1273 fh.write(base64.b64decode(bc_image_data.encode()))
1274 except Exception as e:
1275 self.online_log.warn(
1276 '{0} [bc section img save failed] [img_path={1}]'
1277 ' [part_idx={2}]]'.format(self.log_base, img_path, part_idx))
1278 else:
1279 section_img_path = img_path if ocr_data.get('section_img') is None else '{0}_{1}{2}'.format(pre, part_idx, suf)
1280 if ocr_data.get('section_img') is not None:
1281 try:
1282 with open(section_img_path, "wb") as fh:
1283 fh.write(base64.b64decode(ocr_data.get('section_img').encode()))
1284 except Exception as e:
1285 self.online_log.warn(
1286 '{0} [bc section img save failed] [img_path={1}]'
1287 ' [part_idx={2}]'.format(self.log_base,
1288 img_path, part_idx))
1289
1290 new_ocr_2_res['ErrorCode'] = 0
1291 position_dict[consts.POSITION_KEY] = position
1292 position_dict[consts.ANGLE_KEY] = angle
1293 new_ocr_2_res[consts.SECTION_IMG_PATH_KEY] = section_img_path
1294 new_ocr_2_res[consts.ALL_POSITION_KEY] = position_dict
1295
1235 name = '有' 1296 name = '有'
1236 json_data_3 = { 1297 json_data_3 = {
1237 "file": file_data, 1298 "file": file_data,
1238 'card_res': ocr_2_res 1299 'card_res': new_ocr_2_res
1239 } 1300 }
1240 card_name_response = requests.post(self.ocr_url_3, json_data_3) 1301 card_name_response = requests.post(self.ocr_url_3, json_data_3)
1241 if card_name_response.status_code == 200: 1302 if card_name_response.status_code == 200:
...@@ -1243,9 +1304,11 @@ class Command(BaseCommand, LoggerMixin): ...@@ -1243,9 +1304,11 @@ class Command(BaseCommand, LoggerMixin):
1243 if isinstance(card_name_res, dict) and \ 1304 if isinstance(card_name_res, dict) and \
1244 card_name_res.get('data', {}).get('is_exists_name') == 0: 1305 card_name_res.get('data', {}).get('is_exists_name') == 0:
1245 name = '无' 1306 name = '无'
1246 ocr_2_res['Name'] = name 1307 new_ocr_2_res['Name'] = name
1308 else:
1309 new_ocr_2_res = ocr_2_res
1247 1310
1248 self.license2_process(ocr_2_res, license_summary, pid, classify, 1311 self.license2_process(new_ocr_2_res, license_summary, pid, classify,
1249 res_list, pno, ino, part_idx, img_path, 1312 res_list, pno, ino, part_idx, img_path,
1250 do_dda, dda_id_bc_mapping, file_data=ocr_data.get('section_img')) 1313 do_dda, dda_id_bc_mapping, file_data=ocr_data.get('section_img'))
1251 break 1314 break
......
Styling with Markdown is supported
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!