flask實現基於elasticsearch的關鍵詞搜索建議

1、實現效果

2、fuzzy模糊查詢和suggest查詢

• fuzzy模糊查詢

GET chaxun/job/_search
{
  "query": {
    "fuzzy": {
      "title": {
        "value": "pythn",
        "fuzziness": 2,
        "prefix_length": 2
      }
    }
  }
}

注釋：“fuzziness”為“編輯距離”，“編輯距離”是一種字符串之間相似程度的計算方法。即兩個字符串之間的編輯距離等於使一個字符串變成另外一個字符串而進行的插入、刪除、替換或相鄰字符交換位置而進行操作的最少次數。“prefix_length”，前綴相同長度。

• suggest查詢

POST chaxun/_search?pretty
{
    "suggest": {
        "suggest" : {
            "prefix" : "python",
            "completion" : {
                "field" : "suggest",
                "fuzzy" : {
                    "fuzziness" : 2
                }
            }
        }
    }
}

3、建立模型

from elasticsearch_dsl import Document, Completion, Text, Date, Keyword, Integer
from elasticsearch_dsl.analysis import CustomAnalyzer as _CustomAnalyzer

class CustomAnalyzer(_CustomAnalyzer):
    def get_analysis_definition(self):
        return {}

ik_analyzer = CustomAnalyzer("ik_max_word", filter=["lowercase"])

class BoleAtricle(Document):
    suggest = Completion(analyzer=ik_analyzer)
    title = Text(analyzer="ik_max_word")
    create_date = Date()
    url = Keyword()
    url_object_id = Keyword()
    front_image_url = Keyword()
    praise_nums = Integer()
    comments_nums = Integer()
    fav_nums = Integer()
    tags = Text(analyzer="ik_max_word")
    content = Text(analyzer="ik_max_word")

    class Index:
        name = "jobbole"

    class Meta:
        doc_type = "article"

4、視圖函數

from app.models import BoleAtricle
import json


@main.route("/suggest/<text>")
def get_suggest_phrase(text):
    s = BoleAtricle.search()
    suggest = s.suggest("suggest", text, completion={
        "field": "suggest", "fuzzy": {
            "fuzziness": 2
        },
        "size": 10
    })
    suggestions = suggest.execute()
    suggest_phrase = []
    li_suggest = suggestions.suggest.to_dict().get("suggest", [])
    if li_suggest:
        for item in li_suggest[0].get("options", []):
            suggest_phrase.append(item["_source"]["title"])
    return json.dumps(suggest_phrase)