Bagel

Bagel（AI 的开放推理平台）就像 AI 数据的 GitHub。这是一个协作平台，用户可以在其中创建、共享和管理推理数据集。它可以支持独立开发人员的私有项目、企业的内部协作以及数据 DAO 的公开贡献。

安装和设置

pip install bagelML langchain-community

从文本创建 VectorStore

from langchain_community.vectorstores import Bagel

texts = ["hello bagel", "hello langchain", "I love salad", "my car", "a dog"]
# create cluster and add texts
cluster = Bagel.from_texts(cluster_name="testing", texts=texts)

API 参考：Bagel

# similarity search
cluster.similarity_search("bagel", k=3)

[Document(page_content='hello bagel', metadata={}),
 Document(page_content='my car', metadata={}),
 Document(page_content='I love salad', metadata={})]

# the score is a distance metric, so lower is better
cluster.similarity_search_with_score("bagel", k=3)

[(Document(page_content='hello bagel', metadata={}), 0.27392977476119995),
 (Document(page_content='my car', metadata={}), 1.4783176183700562),
 (Document(page_content='I love salad', metadata={}), 1.5342965126037598)]

# delete the cluster
cluster.delete_cluster()

从文档创建 VectorStore

from langchain_community.document_loaders import TextLoader
from langchain_text_splitters import CharacterTextSplitter

loader = TextLoader("../../how_to/state_of_the_union.txt")
documents = loader.load()
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
docs = text_splitter.split_documents(documents)[:10]

API 参考：TextLoader | CharacterTextSplitter

# create cluster with docs
cluster = Bagel.from_documents(cluster_name="testing_with_docs", documents=docs)

# similarity search
query = "What did the president say about Ketanji Brown Jackson"
docs = cluster.similarity_search(query)
print(docs[0].page_content[:102])

Madam Speaker, Madam Vice President, our First Lady and Second Gentleman. Members of Congress and the

从集群获取所有文本/文档

texts = ["hello bagel", "this is langchain"]
cluster = Bagel.from_texts(cluster_name="testing", texts=texts)
cluster_data = cluster.get()

# all keys
cluster_data.keys()

dict_keys(['ids', 'embeddings', 'metadatas', 'documents'])

# all values and keys
cluster_data

{'ids': ['578c6d24-3763-11ee-a8ab-b7b7b34f99ba',
  '578c6d25-3763-11ee-a8ab-b7b7b34f99ba',
  'fb2fc7d8-3762-11ee-a8ab-b7b7b34f99ba',
  'fb2fc7d9-3762-11ee-a8ab-b7b7b34f99ba',
  '6b40881a-3762-11ee-a8ab-b7b7b34f99ba',
  '6b40881b-3762-11ee-a8ab-b7b7b34f99ba',
  '581e691e-3762-11ee-a8ab-b7b7b34f99ba',
  '581e691f-3762-11ee-a8ab-b7b7b34f99ba'],
 'embeddings': None,
 'metadatas': [{}, {}, {}, {}, {}, {}, {}, {}],
 'documents': ['hello bagel',
  'this is langchain',
  'hello bagel',
  'this is langchain',
  'hello bagel',
  'this is langchain',
  'hello bagel',
  'this is langchain']}

cluster.delete_cluster()

使用元数据创建集群并进行过滤

texts = ["hello bagel", "this is langchain"]
metadatas = [{"source": "notion"}, {"source": "google"}]

cluster = Bagel.from_texts(cluster_name="testing", texts=texts, metadatas=metadatas)
cluster.similarity_search_with_score("hello bagel", where={"source": "notion"})

[(Document(page_content='hello bagel', metadata={'source': 'notion'}), 0.0)]

# delete the cluster
cluster.delete_cluster()

相关

• 向量存储 概念指南
• 向量存储 操作指南