wordpress自带站内搜索功能洛阳seo博客

通过PaddleOCR识别图片中的文字，将识别结果报存到es中，利用es查询语句返回结果图片。

技术逻辑

• PaddleOCR部署、es部署
• 创建mapping
• 将PaddleOCR识别结果保存至es
• 通过查询，返回结果

前期准备

PaddleOCR、es部署请参考https://blog.csdn.net/zhanghan11366/article/details/137026144?spm=1001.2014.3001.5502

创建mapping

from elasticsearch import Elasticsearch# 连接Elasticsearch
es_client = Elasticsearch("http://0.0.0.0:9200/", basic_auth=("elastic", "ZargEZ7NmJRkXLFlEqgE"))# 创建新的ES index
mapping = {'properties': {'description': {'type': 'text','analyzer': 'ik_smart','search_analyzer': 'ik_smart'},"insert_time": {"type": "date","format": "yyyy-MM-dd HH:mm:ss"},"image_path":{'type': 'text'}}
}es_client.indices.create(index='wechat_search_ocr', ignore=400)
result = es_client.indices.put_mapping(index='wechat_search_ocr', body=mapping)
print(result)

将PaddleOCR识别结果保存至es

核心代码展示

def image_ocr(image_dir):files = os.listdir(image_dir)image_files = [file for file in files if file.endswith(('jpg', 'jpeg', 'png', 'gif'))]for image_file in image_files:image_path = os.path.join(image_dir, image_file)if not os.path.isfile(image_path):print(f"文件不存在：{image_path}")continueimage = cv2.imread(image_path)if image is None:print(f"无法读取图像：{image_path}")continueimage_base64 = cv2_to_base64(image)data = {'images': [image_base64]}headers = {"Content-type": "application/json"}url = "http://192.168.30.71:8866/predict/ch_pp-ocrv3"try:r = requests.post(url=url, headers=headers, data=json.dumps(data))r.raise_for_status()  # 检查请求是否成功ocr_results = r.json().get("results", [])if ocr_results:description = "\n".join([ocr_record["text"].strip() for ocr_record in ocr_results[0]["data"]])doc = {"description": description,"insert_time": dt.now().strftime("%Y-%m-%d %H:%M:%S"),"image_path": image_file}es_client.index(index="wechat_search_ocr", body=doc)print("成功插入到 Elasticsearch 中！")else:print("OCR 服务返回结果为空！")except Exception as e:print(f"处理图像 {image_path} 时发生错误：{str(e)}")

通过查询，返回结果

核心代码展示

def image_search_by_text(query_str):result = []# 对query进行全文搜索queries = query_str.split()dsl = {"query": {"bool": {"must": [{"match": {"description": _}} for _ in queries]}},"size": 5}search_result = es_client.search(index='wechat_search_ocr', body=dsl)return search_resultdef image_search_interface(query_str):# 查询图像search_results = image_search_by_text(query_str)# 构建结果images=[]for hit in search_results['hits']['hits']:image_filename = hit['_source']['image_path']image_path = os.path.join('./data', image_filename)image = Image.open(image_path).convert('RGB')images.append(image)if len(images) >= 3:images = images[:3]else:for _ in range(3 - len(images)):images.append(None)return images[0], images[1], images[2]

结果如下：