在本教程中，我们将向您展示如何基于向量数据库构建标准检索器，该检索器将通过 top-k 相似度获取节点。

我们使用 Pinecone 作为向量数据库。我们使用高级摄取抽象来加载节点（要了解如何从零开始构建，请参阅我们之前的教程！）。

我们将展示如何执行以下操作：

如何生成查询嵌入

1. 如何使用不同的搜索模式（密集、稀疏、混合）查询向量数据库
2. 如何将结果解析为节点集
3. 如何将其放入自定义检索器中
4. 设置¶

我们构建一个空的 Pinecone 索引，并定义必要的 LlamaIndex 封装器/抽象，以便我们可以开始将数据加载到 Pinecone 中。

如果您在 colab 上打开此笔记本，您可能需要安装 LlamaIndex 🦙。

输入 [ ]

已复制！

%pip install llama-index-readers-file pymupdf
%pip install llama-index-vector-stores-pinecone
%pip install llama-index-embeddings-openai

!pip install llama-index

已复制！

!pip install llama-index

构建 Pinecone 索引¶

from pinecone import Pinecone, Index, ServerlessSpec import os api_key = os.environ["PINECONE_API_KEY"] pc = Pinecone(api_key=api_key)

已复制！

from pinecone import Pinecone, Index, ServerlessSpec
import os

api_key = os.environ["PINECONE_API_KEY"]
pc = Pinecone(api_key=api_key)

# dimensions are for text-embedding-ada-002 dataset_name = "quickstart" if dataset_name not in pc.list_indexes().names(): pc.create_index( dataset_name, dimension=1536, metric="euclidean", spec=ServerlessSpec(cloud="aws", region="us-east-1"), )

已复制！

# dimensions are for text-embedding-ada-002
dataset_name = "quickstart"
if dataset_name not in pc.list_indexes().names():
    pc.create_index(
        dataset_name,
        dimension=1536,
        metric="euclidean",
        spec=ServerlessSpec(cloud="aws", region="us-east-1"),
    )

pinecone_index = pc.Index(dataset_name)

已复制！

pinecone_index = pc.Index(dataset_name)

# [Optional] drop contents in index pinecone_index.delete(deleteAll=True)

已复制！

# [Optional] drop contents in index
pinecone_index.delete(deleteAll=True)

创建 PineconeVectorStore¶

LlamaIndex 中使用的简单封装抽象。封装在 StorageContext 中，以便我们可以轻松加载节点。

from llama_index.vector_stores.pinecone import PineconeVectorStore

已复制！

from llama_index.vector_stores.pinecone import PineconeVectorStore

vector_store = PineconeVectorStore(pinecone_index=pinecone_index)

已复制！

vector_store = PineconeVectorStore(pinecone_index=pinecone_index)

加载文档¶

!mkdir data !wget --user-agent "Mozilla" "https://arxiv.org/pdf/2307.09288.pdf" -O "data/llama2.pdf"

已复制！

!mkdir data
!wget --user-agent "Mozilla" "https://arxiv.org/pdf/2307.09288.pdf" -O "data/llama2.pdf"

from pathlib import Path from llama_index.readers.file import PyMuPDFReader

已复制！

from pathlib import Path
from llama_index.readers.file import PyMuPDFReader

loader = PyMuPDFReader() documents = loader.load(file_path="./data/llama2.pdf")

已复制！

loader = PyMuPDFReader()
documents = loader.load(file_path="./data/llama2.pdf")

加载到向量存储¶

将文档加载到 PineconeVectorStore 中。

注意：我们在这里使用高级摄取抽象，通过 VectorStoreIndex.from_documents。本教程的其余部分我们将避免使用 VectorStoreIndex。

from llama_index.core import VectorStoreIndex from llama_index.core.node_parser import SentenceSplitter from llama_index.core import StorageContext

已复制！

from llama_index.core import VectorStoreIndex
from llama_index.core.node_parser import SentenceSplitter
from llama_index.core import StorageContext

splitter = SentenceSplitter(chunk_size=1024) storage_context = StorageContext.from_defaults(vector_store=vector_store) index = VectorStoreIndex.from_documents( documents, transformations=[splitter], storage_context=storage_context )

已复制！

splitter = SentenceSplitter(chunk_size=1024)
storage_context = StorageContext.from_defaults(vector_store=vector_store)
index = VectorStoreIndex.from_documents(
    documents, transformations=[splitter], storage_context=storage_context
)

定义向量检索器¶

现在我们准备好针对此向量存储定义我们的检索器，以检索一组节点。

我们将逐步展示过程，然后将其封装到一个函数中。

query_str = "Can you tell me about the key concepts for safety finetuning"

已复制！

query_str = "Can you tell me about the key concepts for safety finetuning"

1. 生成查询嵌入¶

from llama_index.embeddings.openai import OpenAIEmbedding embed_model = OpenAIEmbedding()

已复制！

from llama_index.embeddings.openai import OpenAIEmbedding

embed_model = OpenAIEmbedding()

query_embedding = embed_model.get_query_embedding(query_str)

已复制！

query_embedding = embed_model.get_query_embedding(query_str)

2. 查询向量数据库¶

我们展示如何使用不同的模式查询向量数据库：默认、稀疏和混合。

我们首先构建一个 VectorStoreQuery，然后查询向量数据库。

# construct vector store query from llama_index.core.vector_stores import VectorStoreQuery query_mode = "default" # query_mode = "sparse" # query_mode = "hybrid" vector_store_query = VectorStoreQuery( query_embedding=query_embedding, similarity_top_k=2, mode=query_mode )

已复制！

# construct vector store query
from llama_index.core.vector_stores import VectorStoreQuery

query_mode = "default"
# query_mode = "sparse"
# query_mode = "hybrid"

vector_store_query = VectorStoreQuery(
    query_embedding=query_embedding, similarity_top_k=2, mode=query_mode
)

# returns a VectorStoreQueryResult query_result = vector_store.query(vector_store_query) query_result

已复制！

# returns a VectorStoreQueryResult
query_result = vector_store.query(vector_store_query)
query_result

3. 将结果解析为节点集¶

VectorStoreQueryResult 返回节点集和相似度。我们用此构建一个 NodeWithScore 对象。

from llama_index.core.schema import NodeWithScore from typing import Optional nodes_with_scores = [] for index, node in enumerate(query_result.nodes): score: Optional[float] = None if query_result.similarities is not None: score = query_result.similarities[index] nodes_with_scores.append(NodeWithScore(node=node, score=score))

已复制！

from llama_index.core.schema import NodeWithScore
from typing import Optional

nodes_with_scores = []
for index, node in enumerate(query_result.nodes):
    score: Optional[float] = None
    if query_result.similarities is not None:
        score = query_result.similarities[index]
    nodes_with_scores.append(NodeWithScore(node=node, score=score))

from llama_index.core.response.notebook_utils import display_source_node for node in nodes_with_scores: display_source_node(node, source_length=1000)

已复制！

from llama_index.core.response.notebook_utils import display_source_node

for node in nodes_with_scores:
    display_source_node(node, source_length=1000)

4. 将其放入检索器中¶

让我们将其放入一个检索器子类中，以便它可以插入 LlamaIndex 工作流的其余部分！

from llama_index.core import QueryBundle from llama_index.core.retrievers import BaseRetriever from typing import Any, List class PineconeRetriever(BaseRetriever): """Retriever over a pinecone vector store.""" def __init__( self, vector_store: PineconeVectorStore, embed_model: Any, query_mode: str = "default", similarity_top_k: int = 2, ) -> None: """Init params.""" self._vector_store = vector_store self._embed_model = embed_model self._query_mode = query_mode self._similarity_top_k = similarity_top_k super().__init__() def _retrieve(self, query_bundle: QueryBundle) -> List[NodeWithScore]: """Retrieve.""" if query_bundle.embedding is None: query_embedding = self._embed_model.get_query_embedding( query_bundle.query_str ) else: query_embedding = query_bundle.embedding vector_store_query = VectorStoreQuery( query_embedding=query_embedding, similarity_top_k=self._similarity_top_k, mode=self._query_mode, ) query_result = self._vector_store.query(vector_store_query) nodes_with_scores = [] for index, node in enumerate(query_result.nodes): score: Optional[float] = None if query_result.similarities is not None: score = query_result.similarities[index] nodes_with_scores.append(NodeWithScore(node=node, score=score)) return nodes_with_scores

已复制！

from llama_index.core import QueryBundle
from llama_index.core.retrievers import BaseRetriever
from typing import Any, List


class PineconeRetriever(BaseRetriever):
    """Retriever over a pinecone vector store."""

    def __init__(
        self,
        vector_store: PineconeVectorStore,
        embed_model: Any,
        query_mode: str = "default",
        similarity_top_k: int = 2,
    ) -> None:
        """Init params."""
        self._vector_store = vector_store
        self._embed_model = embed_model
        self._query_mode = query_mode
        self._similarity_top_k = similarity_top_k
        super().__init__()

    def _retrieve(self, query_bundle: QueryBundle) -> List[NodeWithScore]:
        """Retrieve."""
        if query_bundle.embedding is None:
            query_embedding = self._embed_model.get_query_embedding(
                query_bundle.query_str
            )
        else:
            query_embedding = query_bundle.embedding

        vector_store_query = VectorStoreQuery(
            query_embedding=query_embedding,
            similarity_top_k=self._similarity_top_k,
            mode=self._query_mode,
        )
        query_result = self._vector_store.query(vector_store_query)

        nodes_with_scores = []
        for index, node in enumerate(query_result.nodes):
            score: Optional[float] = None
            if query_result.similarities is not None:
                score = query_result.similarities[index]
            nodes_with_scores.append(NodeWithScore(node=node, score=score))

        return nodes_with_scores

retriever = PineconeRetriever( vector_store, embed_model, query_mode="default", similarity_top_k=2 )

已复制！

retriever = PineconeRetriever(
    vector_store, embed_model, query_mode="default", similarity_top_k=2
)

retrieved_nodes = retriever.retrieve(query_str) for node in retrieved_nodes: display_source_node(node, source_length=1000)

已复制！

retrieved_nodes = retriever.retrieve(query_str)
for node in retrieved_nodes:
    display_source_node(node, source_length=1000)

将其插入我们的 RetrieverQueryEngine 以合成响应¶

注意：我们将在未来的教程中更详细地介绍如何从零开始构建响应合成！

from llama_index.core.query_engine import RetrieverQueryEngine query_engine = RetrieverQueryEngine.from_args(retriever)

已复制！

from llama_index.core.query_engine import RetrieverQueryEngine

query_engine = RetrieverQueryEngine.from_args(retriever)

response = query_engine.query(query_str)

已复制！

response = query_engine.query(query_str)

print(str(response))

已复制！

print(str(response))

返回顶部

The key concepts for safety fine-tuning include supervised safety fine-tuning, safety RLHF (Reinforcement Learning from Human Feedback), and safety context distillation. Supervised safety fine-tuning involves gathering adversarial prompts and safe demonstrations to train the model to align with safety guidelines. Safety RLHF integrates safety into the RLHF pipeline by training a safety-specific reward model and gathering challenging adversarial prompts for fine-tuning. Safety context distillation refines the RLHF pipeline by generating safer model responses using a safety preprompt and fine-tuning the model on these responses without the preprompt. These concepts are used to mitigate safety risks and improve the safety of the model's responses.