# 安装依赖pip install llama_index pyvis
# 导入包from llama_index import ( ObsidianReader, # 这个太赞了，可直接读obsidian文件 LLMPredictor, ServiceContext, KnowledgeGraphIndex,)
from llama_index.graph_stores import SimpleGraphStorefrom llama_index.storage.storage_context import StorageContextfrom llama_index.llms import OpenAI
filepath = '/path/to/your/obsidian/vault'
# 初始化contextuse_context = { "temperature": 0, "model": "选择一个model，huggingface上找一个就行", "chunk_size": 512}
documents = ObsidianReader(filepath).load_data()llm = OpenAI(temperature=use_context['temperature'], model=use_context['model'])service_context = ServiceContext.from_defaults(llm=llm,  chunk_size=use_context['chunk_size'])
graph_store = SimpleGraphStore()storage_context = StorageContext.from_defaults(graph_store=graph_store)
# 构建知识图谱indexindex = KnowledgeGraphIndex.from_documents( documents=documents, max_triplets_per_chunk=2, storage_context=storage_context, service_context=service_context, include_embeddings=True,)
# 搞定，你可以实验一下跑的结果query = "写一段提示词，可以跟你本地的markdown文件内容相关的问题"query_engine = index.as_query_engine( include_text=True, response_mode="tree_summarize", embedding_mode="hybrid", similarity_top_k=5,)
response = query_engine.query(query)