Haystack 完整教程

from haystack import Pipeline
from haystack.components.retrievers import InMemoryBM25Retriever, InMemoryEmbeddingRetriever
from haystack.components.joiners import DocumentJoiner
from haystack.components.rankers import TransformersSimilarityRanker
from haystack.components.builders import PromptBuilder
from haystack.components.generators import OpenAIGenerator

# 创建Pipeline
pipeline = Pipeline()

# 添加双路检索器
pipeline.add_component("bm25_retriever", InMemoryBM25Retriever(document_store, top_k=10))
pipeline.add_component("embedding_retriever", InMemoryEmbeddingRetriever(document_store, top_k=10))

# 合并检索结果
pipeline.add_component("joiner", DocumentJoiner(join_mode="merge"))

# 重排序（可选但推荐）
pipeline.add_component("ranker", TransformersSimilarityRanker(model="cross-encoder/ms-marco-MiniLM-L-6-v2", top_k=5))

# 构建提示词
template = """
基于以下文档回答问题：

{% for doc in documents %}
文档 {{ loop.index }}:
{{ doc.content }}
---
{% endfor %}

问题：{{ query }}
答案：
"""
pipeline.add_component("prompt_builder", PromptBuilder(template=template))

# 生成答案
pipeline.add_component("llm", OpenAIGenerator(model="gpt-4"))

# 连接所有组件
pipeline.connect("bm25_retriever.documents", "joiner.documents")
pipeline.connect("embedding_retriever.documents", "joiner.documents")
pipeline.connect("joiner.documents", "ranker.documents")
pipeline.connect("ranker.documents", "prompt_builder.documents")
pipeline.connect("prompt_builder.prompt", "llm.prompt")

# 运行查询
result = pipeline.run({
    "bm25_retriever": {"query": "什么是Haystack"},
    "embedding_retriever": {"query": "什么是Haystack"},
    "ranker": {"query": "什么是Haystack"},
    "prompt_builder": {"query": "什么是Haystack"}
})

print(result["llm"]["replies"][0])