elasticsearch安装与使用(2)-基于term匹配的简单搜索引擎搭建

把一篇pdf论文解析后，放入es数据库中，建立倒排索引表，并实现简单搜索。

1、pdf论文解析(英文)

安装pdf解析包

pip install pdfminer.six

def extract_text_from_pdf(filename, page_numbers=None, min_line_length=1):
    '''
    从pdf文件中提取文字
    :param filename: pdf文件
    :param page_numbers: 指定页码，list
    :param min_line_length: 文本最小分隔长度
    :return:
    '''
    paragraphs = []
    buffer = ''
    full_text = ''
    # 提取全部文本
    for i, page_layout in enumerate(extract_pages(filename)):
        if page_numbers is not None and i not in page_numbers:
            continue
        for element in page_layout:
            if isinstance(element, LTTextContainer):
                full_text += element.get_text() + '\n'
        # 按空行分隔，将文本重新组织成段落
        lines = full_text.split('\n')
        for text in lines:
            if len(text) >= min_line_length:
                buffer += (' ' + text) if not text.endswith('-') else text.strip('-')
            elif buffer:
                paragraphs.append(buffer)
                buffer = ''
        if buffer:
            paragraphs.append(buffer)

    return paragraphs

2、关键字提取

nltk安装见nltk安装与使用

import re
from nltk.stem import PorterStemmer
from nltk.tokenize import word_tokenize
from nltk.corpus import stopwords


def keywords(text):
    '''
    提取文本关键字(简化版)
    :param text: 一段字符串
    :return:
    '''
    # 提取所有字母数字，并替换所有非字母数字的字符为空格
    no_symbols = re.sub(r'[^a-zA-Z0-9\s]', ' ', text)

    # 分词
    word_tokens = word_tokenize(no_symbols)

    # 去停用词
    stop_words = set(stopwords.words('english'))
    filtered_words = [w for w in word_tokens if not w.lower() in stop_words]

    # 取词根
    ps = PorterStemmer()
    key_words = [ps.stem(w) for w in filtered_words]

    return ' '.join(key_words)

3、创建Elasticsearch连接，注意添加证书

Elasticsearch安装见elasticsearch安装与使用(1)-使用docker安装Elasticsearch

from elasticsearch import Elasticsearch, helpers
# 创建Elasticsearch连接
es = Elasticsearch(
    hosts=['https://localhost:9200'],  # 服务地址与端口
    basic_auth=("elastic", "N-sf6R*O0Ur344otTfzc"),  # 用户名，密码
    ca_certs="/Users/sunwenjun/data/elastic8/http_ca.crt"  # 证书
)

4、建立倒排索引库

def add_data_to_es(index_name="test_index", text_list=["text1","text2"]):
    '''
    建立索引，并往索引里添加数据
    :param index_name: 定义索引名称
    param paragraphs: 需要检索的文本列表  
    :return:
    '''

    # 如果索引已存在，删除它
    if es.indices.exists(index=index_name):
        es.indices.delete(index=index_name)

    # 创建索引
    es.indices.create(index=index_name)

    # 灌库指令
    actions = []
    for text in text_list:
        action = {
            "_index": index_name,
            "_source": {
                "keywords": keywords(text),
                "text": text
            }
        }
        actions.append(action)

    # 文本灌库
    res = helpers.bulk(es, actions)
    return res

# 往es里添加数据
index_name = "index_test"
add_data_to_es(index_name, paragraphs)

索引库可视化见elasticsearch安装与使用(3)-索引库可视化

5、实现搜索

def search(index_name, query, top_n=3):
    '''
    查询
    :param index_name:
    :param query:
    :param top_n:
    :return:
    '''
    search_query = {
        "match":
            {"keywords": keywords(query)}
    }
    search_res = es.search(index=index_name, query=search_query, size=top_n)

    results = [hit["_source"]["text"] for hit in search_res["hits"]["hits"]]
    return results

query = "retrieval "
results = search(index_name, query, 5)
for res in results:
   	print(res)

5、完整代码

esdemo-01

参考

无需重新学习，使用 Kibana 查询/可视化 SLS 数据