Skip to content

Nebulagraph query engine

NebulaGraphQueryEnginePack #

Bases: BaseLlamaPack

NebulaGraph查询引擎包。

Source code in llama_index/packs/nebulagraph_query_engine/base.py
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
class NebulaGraphQueryEnginePack(BaseLlamaPack):
    """NebulaGraph查询引擎包。"""

    def __init__(
        self,
        username: str,
        password: str,
        ip_and_port: str,
        space_name: str,
        edge_types: str,
        rel_prop_names: str,
        tags: str,
        max_triplets_per_chunk: int,
        docs: List[Document],
        query_engine_type: Optional[NebulaGraphQueryEngineType] = None,
        **kwargs: Any,
    ) -> None:
        """初始化参数。"""
        os.environ["GRAPHD_HOST"] = "127.0.0.1"
        os.environ["NEBULA_USER"] = username
        os.environ["NEBULA_PASSWORD"] = password
        os.environ[
            "NEBULA_ADDRESS"
        ] = ip_and_port  # such as "127.0.0.1:9669" for local instance

        nebulagraph_graph_store = NebulaGraphStore(
            space_name=space_name,
            edge_types=edge_types,
            rel_prop_names=rel_prop_names,
            tags=tags,
        )

        nebulagraph_storage_context = StorageContext.from_defaults(
            graph_store=nebulagraph_graph_store
        )

        # define LLM
        self.llm = OpenAI(temperature=0.1, model="gpt-3.5-turbo")
        self.service_context = ServiceContext.from_defaults(llm=self.llm)

        nebulagraph_index = KnowledgeGraphIndex.from_documents(
            documents=docs,
            storage_context=nebulagraph_storage_context,
            max_triplets_per_chunk=max_triplets_per_chunk,
            service_context=self.service_context,
            space_name=space_name,
            edge_types=edge_types,
            rel_prop_names=rel_prop_names,
            tags=tags,
            include_embeddings=True,
        )

        # create index
        vector_index = VectorStoreIndex.from_documents(docs)

        if query_engine_type == NebulaGraphQueryEngineType.KG_KEYWORD:
            # KG keyword-based entity retrieval
            self.query_engine = nebulagraph_index.as_query_engine(
                # setting to false uses the raw triplets instead of adding the text from the corresponding nodes
                include_text=False,
                retriever_mode="keyword",
                response_mode="tree_summarize",
            )

        elif query_engine_type == NebulaGraphQueryEngineType.KG_HYBRID:
            # KG hybrid entity retrieval
            self.query_engine = nebulagraph_index.as_query_engine(
                include_text=True,
                response_mode="tree_summarize",
                embedding_mode="hybrid",
                similarity_top_k=3,
                explore_global_knowledge=True,
            )

        elif query_engine_type == NebulaGraphQueryEngineType.RAW_VECTOR:
            # Raw vector index retrieval
            self.query_engine = vector_index.as_query_engine()

        elif query_engine_type == NebulaGraphQueryEngineType.RAW_VECTOR_KG_COMBO:
            from llama_index.core.query_engine import RetrieverQueryEngine

            # create custom retriever
            nebulagraph_vector_retriever = VectorIndexRetriever(index=vector_index)
            nebulagraph_kg_retriever = KGTableRetriever(
                index=nebulagraph_index, retriever_mode="keyword", include_text=False
            )
            nebulagraph_custom_retriever = CustomRetriever(
                nebulagraph_vector_retriever, nebulagraph_kg_retriever
            )

            # create response synthesizer
            nebulagraph_response_synthesizer = get_response_synthesizer(
                service_context=self.service_context,
                response_mode="tree_summarize",
            )

            # Custom combo query engine
            self.query_engine = RetrieverQueryEngine(
                retriever=nebulagraph_custom_retriever,
                response_synthesizer=nebulagraph_response_synthesizer,
            )

        elif query_engine_type == NebulaGraphQueryEngineType.KG_QE:
            # using KnowledgeGraphQueryEngine
            from llama_index.core.query_engine import KnowledgeGraphQueryEngine

            self.query_engine = KnowledgeGraphQueryEngine(
                storage_context=nebulagraph_storage_context,
                service_context=self.service_context,
                llm=self.llm,
                verbose=True,
            )

        elif query_engine_type == NebulaGraphQueryEngineType.KG_RAG_RETRIEVER:
            # using KnowledgeGraphRAGRetriever
            from llama_index.core.query_engine import RetrieverQueryEngine
            from llama_index.core.retrievers import KnowledgeGraphRAGRetriever

            nebulagraph_graph_rag_retriever = KnowledgeGraphRAGRetriever(
                storage_context=nebulagraph_storage_context,
                service_context=self.service_context,
                llm=self.llm,
                verbose=True,
            )

            self.query_engine = RetrieverQueryEngine.from_args(
                nebulagraph_graph_rag_retriever, service_context=self.service_context
            )

        else:
            # KG vector-based entity retrieval
            self.query_engine = nebulagraph_index.as_query_engine()

    def get_modules(self) -> Dict[str, Any]:
        """获取模块。"""
        return {
            "llm": self.llm,
            "service_context": self.service_context,
            "query_engine": self.query_engine,
        }

    def run(self, *args: Any, **kwargs: Any) -> Any:
        """运行流水线。"""
        return self.query_engine.query(*args, **kwargs)

get_modules #

get_modules() -> Dict[str, Any]

获取模块。

Source code in llama_index/packs/nebulagraph_query_engine/base.py
170
171
172
173
174
175
176
def get_modules(self) -> Dict[str, Any]:
    """获取模块。"""
    return {
        "llm": self.llm,
        "service_context": self.service_context,
        "query_engine": self.query_engine,
    }

run #

run(*args: Any, **kwargs: Any) -> Any

运行流水线。

Source code in llama_index/packs/nebulagraph_query_engine/base.py
178
179
180
def run(self, *args: Any, **kwargs: Any) -> Any:
    """运行流水线。"""
    return self.query_engine.query(*args, **kwargs)