测试一下python版本的dqrant向量数据库的效果，完整代码如下：
安装库
!pip install qdrant-client>=1.1.1
!pip install -U sentence-transformers
导入
from qdrant_client import models, QdrantClient
from sentence_transformers import SentenceTransformerencoder = SentenceTransformer("all-MiniLM-L6-v2", device="cuda")
准备测试数据集
documents = [{"name": "The Time Machine","description": "A man travels through time and witnesses the evolution of humanity."* 8,"author": "H.G. Wells","year": 1895,},{"name": "Ender's Game","description": "A young boy is trained to become a military leader in a war against an alien race."* 4,"author": "Orson Scott Card","year": 1985,},{"name": "Brave New World","description": "A dystopian society where people are genetically engineered and conditioned to conform to a strict social hierarchy."* 6,"author": "Aldous Huxley","year": 1932,},
] * 50000print(len(documents))
创建存储库
qdrant = QdrantClient(":memory:") # 内存中
# qdrant = QdrantClient(path='./qdrant') # 存储到本地
在数据库中创建一个collection（类似一个存储桶）
qdrant.recreate_collection(collection_name="my_books",vectors_config=models.VectorParams(size=encoder.get_sentence_embedding_dimension(), # Vector size is defined by used modeldistance=models.Distance.COSINE,),
)
对文档进行向量化
import hashlib
from tqdm import tqdmdef sha256(text):hash_object = hashlib.sha256()hash_object.update(text.encode("utf-8"))hash_value = hash_object.hexdigest()return hash_valuerecords = []
bs = 256
for i in tqdm(range(0, len(documents), bs)):docs = documents[i : i + bs]vectors = encoder.encode([doc["description"] for doc in docs], normalize_embeddings=True).tolist()record = [models.Record(id=idx, vector=vec, payload=doc) # sha256(doc['description'])for idx, vec, doc in zip(range(i, i + bs), vectors, docs)]records.extend(record)
上传到向量数据库中指定的collection
qdrant.upload_points(collection_name="my_books", points=records, batch_size=128, parallel=12
)
语义搜索
query = "Aliens attack our planet"
hits = qdrant.search(collection_name="my_books",query_vector=encoder.encode(query).tolist(),limit=6,
)
for hit in hits:print(hit.payload, "score:", hit.score)
条件搜索
search only for books from 21st century
hits = qdrant.search(collection_name="my_books",query_vector=encoder.encode("Tyranic society").tolist(),query_filter=models.Filter(must=[models.FieldCondition(key="year", range=models.Range(gte=1980))]),limit=3,
)
for hit in hits:print(hit.payload, "score:", hit.score)
参考官方GitHub
github
colab