# 识别结果后处理 #### 提取特定字段导入Excel 本示例调用身份证识别API，并从获取到的JSON结果中，提取所需的字段，填入至Excel。 - 前提条件 - [开通身份证识别](https://support.huaweicloud.com/usermanual-ocr/ocr_08_0002.html)。 - 参考[本地调用](https://support.huaweicloud.com/usermanual-ocr/ocr_08_0007.html)，安装OCR Python SDK。并执行**pip install xlsxwriter**命令安装依赖包。 - 登录"[我的凭证](https://console.huaweicloud.com/iam/#/myCredential) \> 访问密钥"页面，获取AK、SK。可以新增访问密钥，或使用已有的访问密钥。访问密钥为credentials.csv文件，包含AK/SK信息。 ![](https://support.huaweicloud.com/usermanual-ocr/zh-cn_image_0000001192206676.png "点击放大") - 代码示例 ``` # -*- coding: utf-8 -*- import base64 import xlsxwriter from huaweicloudsdkcore.auth.credentials import BasicCredentials from huaweicloudsdkocr.v1.region.ocr_region import OcrRegion from huaweicloudsdkcore.exceptions import exceptions from huaweicloudsdkocr.v1 import * from huaweicloudsdkcore.http.http_config import HttpConfig def recognize_id_card_request(): try: request = RecognizeIdCardRequest() request.body = IdCardRequestBody( image=image_base64 ) response = client.recognize_id_card(request) return response except exceptions.ClientRequestException as e: print(e.status_code) print(e.request_id) print(e.error_code) print(e.error_msg) def get_credential(): return BasicCredentials(ak, sk) def get_client(): config = HttpConfig.get_default_config() config.ignore_ssl_verification = True return OcrClient.new_builder(OcrClient) \ .with_credentials(credentials) \ .with_region(OcrRegion.CN_NORTH_4) \ .with_http_config(config) \ .build() def image_to_base64(imagepath): """ 将本地图片转化为base64编码 """ with open(imagepath, "rb") as bin_data: image_data = bin_data.read() base64_data = base64.b64encode(image_data).decode("utf-8") return base64_data def response_to_execl(save_file, data): """ :param save_file: 文件名 :param data: result data """ # 处理调用API返回的result数据 keys_list = list(data["result"].keys()) values_list = list(data["result"].values()) options = {'in_memory': True} with xlsxwriter.Workbook(save_file, options) as workbook: worksheet = workbook.add_worksheet() worksheet.set_column('A1:A20', 23) worksheet.set_column('B1:B20', 100) worksheet.write_column('A1', keys_list) worksheet.write_column('B1', values_list) workbook.close() if __name__ == '__main__': # 填写访问密钥AK、SK ak = "填写AK" sk = "填写SK" # Init Auth Info credentials = get_credential() # Create OcrClient client = get_client() image_base64 = image_to_base64(r"图片的路径，例如D:\local\test.png") # request id card service response = recognize_id_card_request().to_dict() # 数据保存在execl上 response_to_execl(r"excel路径，例如D:\local\test.xlsx", response) ``` #### 使用PDF进行文字识别 本示例将PDF文件转换为图片，并调用网络图片识别API，获取识别结果。 - 前提条件 - [开通网络图片识别](https://support.huaweicloud.com/usermanual-ocr/ocr_08_0002.html)。 - 参考[本地调用](https://support.huaweicloud.com/usermanual-ocr/ocr_08_0007.html)，安装OCR Python SDK。并执行**pip install fitz** 命令和**pip install PyMuPDF==1.18.0**命令安装依赖包。 - 登录"[我的凭证](https://console.huaweicloud.com/iam/#/myCredential) \> 访问密钥"页面，获取AK、SK。可以新增访问密钥，或使用已有的访问密钥。访问密钥为credentials.csv文件，包含AK/SK信息。 ![](https://support.huaweicloud.com/usermanual-ocr/zh-cn_image_0000001192526660.png "点击放大") - 代码示例 ``` # -*- coding: utf-8 -*- import os import base64 import fitz import io from PIL import Image from glob import glob from huaweicloudsdkcore.auth.credentials import BasicCredentials from huaweicloudsdkocr.v1.region.ocr_region import OcrRegion from huaweicloudsdkcore.exceptions import exceptions from huaweicloudsdkocr.v1 import * from huaweicloudsdkcore.http.http_config import HttpConfig class CovertPdfToJpg: def __init__(self, file_path, save_root): self.file_path = file_path self.save_root = save_root @staticmethod def open_pdf(file): return fitz.open(file) @staticmethod def get_trans(doc, page, min_side=0, max_side=0, rotate=0.0): """ 创建scale对象 """ region = doc[page].rect scale = 1 if max_side > min_side > 0: scale = min_side / min(region.width, region.height) if max(region.width, region.height) * scale > max_side: scale = max_side / max(region.width, region.height) trans = fitz.Matrix(scale, scale).preRotate(rotate) return trans def page2pix(self, doc, page, trans): """ 根据给定的参数，解析当前页作为图片数据 """ return doc[page].getPixmap(matrix=trans, alpha=False) def pdf_to_jpg(self, width=1024, height=1400): """ 将pdf格式图片，转化为jpg格式 """ doc = self.open_pdf(self.file_path) save_dir = os.path.join(self.save_root) if not os.path.exists(save_dir): os.makedirs(save_dir) print("document", len(doc), doc.pageCount) for i in range(len(doc)): trans = self.get_trans(doc, i, width, height, rotate=0) try: pdf = self.page2pix(doc, i, trans) except: continue image = pdf.getPNGData() image = Image.open(io.BytesIO(image)) print(os.path.join( save_dir, os.path.basename(self.file_path).replace('.pdf', '') + '_' + str(i + 1) + '.jpg')) image.save( os.path.join(save_dir, os.path.basename(self.file_path).replace('.pdf', '') + '_' + str(i + 1) + '.jpg')) return def recognize_general_text_request(): try: request = RecognizeGeneralTextRequest() request.body = GeneralTextRequestBody( image=image_base64 ) response = client.recognize_general_text(request) print(response) except exceptions.ClientRequestException as e: print(e.status_code) print(e.request_id) print(e.error_code) print(e.error_msg) def get_credential(): return BasicCredentials(ak, sk) def get_client(): config = HttpConfig.get_default_config() config.ignore_ssl_verification = True return OcrClient.new_builder(OcrClient) \ .with_credentials(credentials) \ .with_region(OcrRegion.CN_NORTH_4) \ .with_http_config(config) \ .build() def url_to_base64(imagepath): with open(imagepath, "rb") as bin_data: image_data = bin_data.read() base64_data = base64.b64encode(image_data).decode("utf-8") return base64_data if __name__ == '__main__': # 填写访问密钥AK、SK ak = "填写AK" sk = "填写SK" # Init Auth Info credentials = get_credential() # Create OcrClient client = get_client() df_path = r"本地pdf文件路径，例如D:\local\test.pdf" save_path = r"转换后的图片路径，例如D:\local" covert_pdf_to_jpg = CovertPdfToJpg(df_path, save_path) covert_pdf_to_jpg.pdf_to_jpg() jpgs = glob(os.path.join(save_path, "*.jpg")) for jpg in jpgs: image_base64 = url_to_base64(jpg) recognize_general_text_request() ``` ![](https://support.huaweicloud.com/usermanual-ocr/public_sys-resources/note_3.0-zh-cn.png) 如果运行过程出现"AttributeError: 'Document' object has no attribute 'pageCount'"报错，可以卸载原有的**PyMuPDF**依赖包，并安装1.18.0版本解决。