@componentclass QueryExpander:
def __init__(self, prompt: Optional[str] = None, model: str = "gpt-3.5-turbo"):
self.query_expansion_prompt = promptself.model = modelif prompt == None:self.query_expansion_prompt = """You are part of an information system that processes users queries.You expand a given query into {{ number }} queries that are similar in meaning.
Structure:Follow the structure shown below in examples to generate expanded queries.Examples:1. Example Query 1: "climate change effects"Example Expanded Queries: ["impact of climate change", "consequences of global warming", "effects of environmental changes"]
2. Example Query 2: ""machine learning algorithms""Example Expanded Queries: ["neural networks", "clustering", "supervised learning", "deep learning"]
Your Task:Query: "{{query}}"Example Expanded Queries:"""builder = PromptBuilder(self.query_expansion_prompt)llm = OpenAIGenerator(model = self.model)self.pipeline = Pipeline()self.pipeline.add_component(name="builder", instance=builder)self.pipeline.add_component(name="llm", instance=llm)self.pipeline.connect("builder", "llm")
@component.output_types(queries=List[str])def run(self, query: str, number: int = 5):result = self.pipeline.run({'builder': {'query': query, 'number': number}})expanded_query = json.loads(result['llm']['replies'][0]) + [query]return {"queries": list(expanded_query)}

expander = QueryExpander()expander.run(query="开源 NLP 框架", number=4)

{'queries': ['自然语言处理工具','免费 NLP 库','开源语言处理平台','带有开源代码的 NLP 软件','开源 NLP 框架']}

documents = [Document(content="气候的影响有很多，包括生物多样性的丧失。"),Document(content="气候变化的影响在极地冰盖的融化中是显而易见的。"),    Document(content="全球气候变化的后果变暖的包括的上升海平面。"),Document(content="环境变化的影响之一是天气模式的变化。"),Document(content="全球都在呼吁减少人们的航空旅行次数。"),Document(content="航空旅行是造成气候变化的主要因素之一。"),Document(content="预计土耳其夏季气候会变暖。"),]

'航空旅行是造成气候变化的主要因素之一。''气候变化的影响在极地冰盖的融化中是显而易见的。''气候的影响有很多，包括生物多样性的丧失。'

query_expander = QueryExpander()retriever = MultiQueryInMemoryBM25Retriever(InMemoryBM25Retriever(document_store=doc_store))
expanded_retrieval_pipeline = Pipeline()expanded_retrieval_pipeline.add_component("expander", query_expander)expanded_retrieval_pipeline.add_component("keyword_retriever", retriever)
expanded_retrieval_pipeline.connect("expander.queries", "keyword_retriever.queries")

expanded_retrieval_pipeline.run({"expander": {"query": "climate change"}},include_outputs_from=["expander"])

'expander': {'queries': ['全球变暖的后果', '气候变化对环境的影响', '气候变率的影响', '气候危机的影响', '温室气体排放的后果', '气候变化']}}

'全球气候变化的后果变暖的包括的上升海平面。''气候变化的影响在极地冰盖的融化中是显而易见的。''全球都在呼吁减少人们的航空旅行次数。''气候的影响有很多，包括生物多样性的丧失。''环境变化的影响之一是天气模式的变化。''航空旅行是造成气候变化的主要因素之一。'

"Electric_vehicle", "Dam", "Electric_battery", "Tree", "Solar_panel", "Nuclear_power","Wind_power", "Hydroelectricity", "Coal", "Natural_gas", "Greenhouse_gas", "Renewable_energy", "Fossil_fuel"

template = """You are part of an information system that summarises related documents.You answer a query using the textual content from the documents retrieved for thefollowing query.You build the summary answer based only on quoting information from the documents.You should reference the documents you used to support your answer.###Original Query: "{{query}}"Retrieved Documents: {{documents}}Summary Answer:"""query_expander = QueryExpander()retriever = MultiQueryInMemoryBM25Retriever(InMemoryBM25Retriever(document_store=doc_store))prompt_builder = PromptBuilder(template = template)llm = OpenAIGenerator()
query_expanded_rag_pipeline = Pipeline()query_expanded_rag_pipeline.add_component("expander", query_expander)query_expanded_rag_pipeline.add_component("keyword_retriever", retriever)query_expanded_rag_pipeline.add_component("prompt", prompt_builder)query_expanded_rag_pipeline.add_component("llm", llm)
query_expanded_rag_pipeline.connect("expander.queries", "keyword_retriever.queries")query_expanded_rag_pipeline.connect("keyword_retriever.documents", "prompt.documents")query_expanded_rag_pipeline.connect("prompt", "llm")