python-base--/vector_test/vector_util.py

from qdrant_client import QdrantClient
from qdrant_client.models import Distance, VectorParams

from qdrant_client.models import PointStruct

from qdrant_client.models import Filter, FieldCondition, MatchValue
from text2vec import SentenceModel


class MyVector:
    def __init__(self, ip, port=6333, grpc_port=6334, api_key=None, embedding_path=None):
        self.url = "http://" + ip
        self.port = port
        self.grpc_port = grpc_port
        self.api_key = api_key
        self.client = QdrantClient(
            url=self.url,
            port=self.port,
            grpc_port=self.grpc_port,
            api_key=self.api_key
        )
        self.model = SentenceModel(embedding_path)

    def create_collection(self, collection_name="test_collection", size=16):
        self.client.create_collection(
            # 设置索引的名称
            collection_name=collection_name,
            # 设置索引中输入向量的长度
            # 参数size是数据维度
            # 参数distance是计算的方法，主要有COSINE（余弦），EUCLID（欧氏距离）、DOT（点积），MANHATTAN（曼哈顿距离）
            vectors_config=VectorParams(size=size, distance=Distance.COSINE),
        )

    def delete_collection(self, collection_name="test_collection"):
        self.client.delete_collection(
            collection_name=collection_name,
        )

    def create_document(self, collection_name="test_collection", doc_id=None, vector=None, payload=None):
        operation_info = self.client.upsert(
            collection_name=collection_name,
            wait=True,
            points=[
                PointStruct(id=doc_id, vector=vector, payload=payload),
            ],
        )
        # 返回值
        # operation_id=0 status=<UpdateStatus.COMPLETED: 'completed'>
        print(operation_info)

    def delete_documentById(self, collection_name="test_collection", doc_id=None):
        operation_info = self.client.delete(
            collection_name=collection_name,
            wait=True,
            points_selector=Filter(
                should=[
                    FieldCondition(
                        key="id",
                        match=MatchValue(value=doc_id)
                    )
                ]
            ),
        )
        # 返回值
        # operation_id=0 status=<UpdateStatus.COMPLETED: 'completed'>
        print(operation_info)

    def query_data(self, collection_name="test_collection", query_vector=None, limit=10, filter_map=None):
        if filter_map is not None:
            # 遍历 filter_map 的 key和value
            query_filter = Filter(
                must=[
                    FieldCondition(
                        key=key,
                        match=MatchValue(value=value)
                    )
                    for key, value in filter_map.items()
                ]
            )
        else:
            query_filter = None

        search_result = self.client.search(
            # 设置索引
            collection_name=collection_name,
            # 查询向量
            query_vector=query_vector,
            # 限制返回值的数量
            limit=limit,
            query_filter=query_filter,
        )
        # 返回值
        # [ScoredPoint(id=4, version=0, score=1.362, payload={'city': 'New York'}, vector=None, shard_key=None), ScoredPoint(id=1, version=0, score=1.273, payload={'city': 'Berlin'}, vector=None, shard_key=None), ScoredPoint(id=3, version=0, score=1.208, payload={'city': 'Moscow'}, vector=None, shard_key=None)]
        return search_result


if __name__ == '__main__':
    client = MyVector("101.43.195.48", port=6333, grpc_port=6334, api_key="tianyunperfect123456", embedding_path='/Users/alvin/IdeaProjects/python-base/vector_test/models--shibing624--text2vec-base-chinese')
    # 1 创建索引
    # client.create_collection('memo_collection', size=768)

    # 2 添加数据
    # add_data()
    # encode = client.model.encode("胡辉地址：中国江苏省南京市雨花台区安德门大街23号金地威新B栋8楼")
    # client.create_document(
    #     collection_name='memo_collection',
    #     doc_id=1,
    #     vector=encode,
    # )
    client.delete_collection('memo_collection')
    client.create_collection('memo_collection', size=768)

    # 3 查询数据
    data = client.query_data(collection_name='memo_collection', query_vector=client.model.encode("你知道小青的女儿吗？"), limit=10)
    print(data)