{"id":48464,"date":"2025-07-30T11:59:38","date_gmt":"2025-07-30T03:59:38","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/48464.html"},"modified":"2025-07-30T11:59:38","modified_gmt":"2025-07-30T03:59:38","slug":"4%e5%8d%a14090%e6%9c%8d%e5%8a%a1%e5%99%a8graphrag%e7%8e%af%e5%a2%83%e9%83%a8%e7%bd%b2%e3%80%81%e5%8a%a0%e8%bd%bdollama%e5%90%91%e9%87%8f%e6%a8%a1%e5%9e%8b%e5%8f%8a%e5%90%91%e9%87%8f%e7%9f%a5%e8%af%86","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/48464.html","title":{"rendered":"4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528"},"content":{"rendered":"

@4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528<\/p>\n

\u672c\u4eba\u5b89\u88c5\u7684\u786c\u4ef6\u73af\u5883\u5982\u4e0b:
\nCPU:intel\u81f3\u5f3a\u91d1\u724c6148*2\u4e3b\u98912.5G \u517140\u683880\u7ebf\u7a0b
\n\u5185\u5b58：\u4e09\u661f32G DDR4\u670d\u52a1\u5668\u5185\u5b582933 *4=128GB
\n\u663e\u5361：\u82f1\u4f1f\u8fbe4090\u6da1\u8f6e\u516c\u7248 *4 = 96G
\n\u64cd\u4f5c\u7cfb\u7edf:ubuntu 24.04 LTS
\nconda:Anaconda 24.11.3<\/p>\n

\u90e8\u7f72GraphRAG\u73af\u5883<\/h2>\n

1.\u521b\u5efa\u865a\u62df\u73af\u5883,\u672c\u9879\u76ee\u7531\u4e8e\u90e8\u7f72GraphRAG\u7a33\u5b9a\u7248\u672c，\u7ecf\u591a\u6b21\u6d4b\u8bd5python3.10\u7248\u672c\u76f8\u5bf9\u8f83\u7a33\u5b9a\u3002<\/p>\n

conda create -n<\/span> graphrag python<\/span>=<\/span>3.10<\/span> -y<\/span><\/p>\n

2.\u8fdb\u5165\u521b\u5efa\u7684GraphRAG\u73af\u5883<\/p>\n

conda activate graphrag <\/p>\n

\u865a\u62df\u73af\u5883\u521b\u5efa\u6210\u529f\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n3.\u5b89\u88c5graphrag\u73af\u5883，\u672c\u9879\u76ee\u5b89\u88c5\u7684\u7a33\u5b9a\u76840.5.0\u7248\u672c<\/p>\n

pip install<\/span> graphrag<\/span>==<\/span>0.5<\/span>.0<\/p>\n

\u5b89\u88c5\u6210\u529f\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"<\/p>\n

4.\u5b89\u88c5openai\u6807\u51c6\u63a5\u5165\u5927\u6a21\u578b\u7684API\u534f\u8bae\u4f9d\u8d56\u9879<\/p>\n

pip install<\/span> openai<\/p>\n

\u5b89\u88c5\u6210\u529f\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n5.\u5b89\u88c5\u8c03\u7528ollama\u670d\u52a1\u7684\u4f9d\u8d56\u9879<\/p>\n

pip install<\/span> ollama<\/p>\n

\u5b89\u88c5\u6210\u529f\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n6.\u7531\u4e8egraphrag\u53d1\u5e03\u662f\u57fa\u4e8eopenai\u73af\u5883\u7684，\u56e0\u6b64\u672c\u5730\u5316\u90e8\u7f72\u53ca\u8c03\u7528\u672c\u5730\u63a8\u7406\u5927\u6a21\u578b，\u9700\u8981\u4fee\u6539graphrag\u5f00\u6e90\u7684\u914d\u7f6e\u6587\u4ef6，\u6ee1\u8db3\u9879\u76ee\u6587\u4ef6\u8c03\u7528\u672c\u5730\u63a8\u7406\u6a21\u578b，\u652f\u6301\u6570\u636e\u4fdd\u5bc6\u76f8\u5173\u4e1a\u52a1\u8fd0\u884c；\u5171\u9700\u8981\u4fee\u6539\u4e09\u4e2aGraphrag\u7684python\u7684\u6e90\u4ee3\u7801\u6587\u4ef6，\u4fee\u6539\u7b80\u5355，\u4f46\u662f\u4e5f\u9700\u8981\u4f1apython\u7f16\u7a0b\u6280\u672f，\u9632\u6b62\u4fee\u6539\u540e\u683c\u5f0f\u95ee\u9898\u62a5\u9519\u3002
\n6.1.\u4fee\u6539llm\/openai\/openai_embeddings_llm.py\u7684GraphRAG\u6e90\u4ee3\u7801\u6587\u4ef6，\u67e5\u627e\u8def\u5f84\u5728conda\u7cfb\u7edf\u76ee\u5f55\u4e0b\u7684env->\u5f53\u524d\u5b89\u88c5\u73af\u5883,\u6211\u5f53\u524d\u9879\u76ee\u540d\u662fgraphrag->lib\/python3.10\/site-packages\/graphrag\u76ee\u5f55\u4e0b，\u5f53\u524d\u6211\u7684\u9879\u76ee\u5177\u4f53\u76ee\u5f55\u662f<\/p>\n

\/home\/allanfan\/anaconda3\/envs\/graphrag\/lib\/python3.10\/site-packages\/graphrag<\/p>\n

\u5f53\u524d\u9879\u76ee\u627e\u5230\u7684graphrag\u6e90\u4ee3\u7801\u4f4d\u7f6e\u76ee\u5f55\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"<\/p>\n

\u5907\u6ce8：\u5927\u5bb6\u7684\u9879\u76ee\u6839\u636e\u5b89\u88c5\u7684conda\u73af\u5883\u76ee\u5f55，\u6309\u6211\u4e0a\u9762\u4e00\u76f4\u627e\u4e0b\u53bb\u5373\u53ef，\u5f53\u524dpython\u73af\u5883\u662f3.10<\/span>，\u5176\u5b83\u66f4\u9ad8\u7248\u672c\u6211\u8bd5\u8fc7，\u4e0d\u592a\u597d\u7528\u3002\u5982\u679c\u5927\u5bb6\u6709\u65f6\u95f4\u53ef\u4ee5\u8bd5\u4e00\u4e0b\u66f4\u9ad8\u7248\u672c，\u4f46\u662fgraphrag 0.5<\/span>.<\/span>0<\/span>\u4ee5\u4e0a\u7248\u672c\u5b8c\u5168\u57fa\u4e8eopenai\u534f\u8bae，\u4e2a\u4eba\u5c1d\u8bd5\u5f88\u591a\u6b21\u65e0\u6cd5\u672c\u5730\u5316\u63a5\u63a8\u7406\u6a21\u578b\u3002<\/p>\n

\u901a\u8fc7\u4e0a\u9762\u76ee\u5f55\u5411\u4e0b\u627e\u5230llm\/openai\/openai_embeddings_llm.py\u6e90\u6587\u4ef6，\u7528sudo nano\u6253\u5f00\u6587\u4ef6\u6ce8\u91ca\u6389\u4e0b\u9762\u4ee3\u7801，\u7136\u540e\u6dfb\u52a0\u4ee5\u4e0b\u65b0\u4ee3\u7801\u3002<\/p>\n

sudo<\/span> nano<\/span> openai_embeddings_llm.py<\/p>\n

\u6ce8\u91ca36\u884c\u81f341\u884c，\u6ce8\u91ca\u5177\u4f53\u4ee3\u7801\u5982\u4e0b\u6240\u793a，\u884c\u6570\u4ec5\u4f9b\u53c2\u8003，\u5177\u4f53\u662f\u4ee5\u4e0b\u4ee3\u7801\u3002<\/p>\n

'''embedding = await self.client.embeddings.create(
\n input=input,
\n **args,
\n)
\nreturn [d.embedding for d in embedding.data]'''<\/span><\/p>\n

\u6ce8\u91ca\u4ee3\u7801\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u6dfb\u52a0\u4ee3\u7801\u5982\u4e0b\u6240\u793a\u3002\u5176\u4e2d，\u5411\u91cf\u6a21\u578b\u540d\u79f0\u6839\u636e\u81ea\u5df1\u4e0b\u8f7d\u7684\u6a21\u578b\u547d\u540d\u4fee\u6539，\u672c\u9879\u76ee\u6d4b\u8bd5\u7528\u7684\u662fnomic-embed-text\u6a21\u578b\u3002
\n\u6587\u4ef6\u5934\u90e8\u6dfb\u52a0\u5f15\u7528\u672c\u5730ollama\u4f9d\u8d56\u9879<\/p>\n

import<\/span> ollama<\/p>\n

\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"<\/p>\n

\u4e0a\u9762\u6ce8\u91ca\u540e\u9762\u6dfb\u52a0\u4ee5\u4e0b\u4ee3\u7801<\/p>\n

embedding_list =<\/span> [<\/span>]<\/span>
\nfor<\/span> inp in<\/span> input<\/span>:<\/span>
\n embedding =<\/span> ollama.<\/span>embeddings(<\/span>model=<\/span>"nomic-embed-text"<\/span>,<\/span>prompt=<\/span>inp)<\/span>
\n embedding_list.<\/span>append(<\/span>embedding[<\/span>"embedding"<\/span>]<\/span>)<\/span>
\nreturn<\/span> embedding_list<\/p>\n

\u6dfb\u52a0\u65b0\u4ee3\u7801\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\nnano\u4fdd\u5b58\u6587\u4ef6\u6309ctrl+o,\u4fdd\u5b58\u540e\u9000\u51fa\u6309ctrl+c\u3002\u4fee\u6539\u540e\u6574\u4f53\u4ee3\u7801\u53c2\u8003\u5982\u4e0b\u6240\u793a\u3002<\/p>\n

# Copyright (c) 2024 Microsoft Corporation.<\/span>
\n# Licensed under the MIT License<\/span><\/p>\n

"""The EmbeddingsLLM class."""<\/span>
\nimport<\/span> ollama<\/p>\n

from<\/span> typing_extensions import<\/span> Unpack<\/p>\n

from<\/span> graphrag.<\/span>llm.<\/span>base import<\/span> BaseLLM
\nfrom<\/span> graphrag.<\/span>llm.<\/span>types import<\/span> (<\/span>
\n EmbeddingInput,<\/span>
\n EmbeddingOutput,<\/span>
\n LLMInput,<\/span>
\n)<\/span><\/p>\n

from<\/span> .<\/span>openai_configuration import<\/span> OpenAIConfiguration
\nfrom<\/span> .<\/span>types import<\/span> OpenAIClientTypes<\/p>\n

class<\/span> OpenAIEmbeddingsLLM<\/span>(<\/span>BaseLLM[<\/span>EmbeddingInput,<\/span> EmbeddingOutput]<\/span>)<\/span>:<\/span>
\n """A text-embedding generator LLM."""<\/span><\/p>\n

_client:<\/span> OpenAIClientTypes
\n _configuration:<\/span> OpenAIConfiguration<\/p>\n

def<\/span> __init__<\/span>(<\/span>self,<\/span> client:<\/span> OpenAIClientTypes,<\/span> configuration:<\/span> OpenAIConfiguration)<\/span>:<\/span>
\n self.<\/span>client =<\/span> client
\n self.<\/span>configuration =<\/span> configuration<\/p>\n

async<\/span> def<\/span> _execute_llm<\/span>(<\/span>
\n self,<\/span> input<\/span>:<\/span> EmbeddingInput,<\/span> **<\/span>kwargs:<\/span> Unpack[<\/span>LLMInput]<\/span>
\n )<\/span> –<\/span>><\/span> EmbeddingOutput |<\/span> None<\/span>:<\/span>
\n args =<\/span> {<\/span>
\n "model"<\/span>:<\/span> self.<\/span>configuration.<\/span>model,<\/span>
\n **<\/span>(<\/span>kwargs.<\/span>get(<\/span>"model_parameters"<\/span>)<\/span> or<\/span> {<\/span>}<\/span>)<\/span>,<\/span>
\n }<\/span>
\n """embedding = await self.client.embeddings.create(
\n input=input,
\n **args,
\n ) return [d.embedding for d in embedding.data]"""<\/span>
\n embedding_list =<\/span> [<\/span>]<\/span>
\n for<\/span> inp in<\/span> input<\/span>:<\/span>
\n embedding =<\/span> ollama.<\/span>embeddings(<\/span>model=<\/span>"nomic-embed-text"<\/span>,<\/span>prompt=<\/span>inp)<\/span>
\n embedding_list.<\/span>append(<\/span>embedding[<\/span>"embedding"<\/span>]<\/span>)<\/span>
\n return<\/span> embedding_list<\/p>\n

6.2.\u4fee\u6539qutry\/llm\/oai\/embedding.py\u6e90\u4ee3\u7801\u6587\u4ef6，\u8fd8\u5728\u5f53\u524d\u6211\u7684\u9879\u76ee\u4e0b\u67e5\u627e\u8be5\u6e90\u4ee3\u7801\u6587\u4ef6，\u4e0b\u9762\u5f53\u524d\u6211\u9879\u76ee\u7684graphrag\u76ee\u5f55.<\/p>\n

\/home\/allanfan\/anaconda3\/envs\/graphrag\/lib\/python3.10\/site-packages\/graphrag<\/p>\n

\u627e\u5230qutry\/llm\/oai\/embedding.py，\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u7528nano\u547d\u4ee4\u6253\u5f00\u6587\u4ef6，\u6ce8\u91ca\u53ca\u6dfb\u52a0\u4ee5\u4e0b\u4ee3\u7801\u3002<\/p>\n

sudo<\/span> nano<\/span> embedding.py<\/p>\n

\u6587\u4ef6\u5934\u90e8\u6dfb\u52a0\u5f15\u7528\u672c\u5730ollama\u4f9d\u8d56\u9879<\/p>\n

import<\/span> ollama<\/p>\n

\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u627e\u5230\u6e90\u4ee3\u7801\u4e2d\u4e24\u5904\u4ee5\u4e0b\u4ee3\u7801，\u5982\u4e0b\u6240\u793a
\n\u7b2c\u4e00\u5904：<\/p>\n

embedding,<\/span> chunk_len =<\/span> self.<\/span>_embed_with_retry(<\/span>chunk,<\/span> **<\/span>kwargs)<\/span><\/p>\n

\u7b2c\u4e8c\u5904：<\/p>\n

chunk_embeddings =<\/span> np.<\/span>average(<\/span>chunk_embeddings,<\/span> axis=<\/span>0<\/span>,<\/span> weights=<\/span>chunk_lens)<\/span>
\nchunk_embeddings =<\/span> chunk_embeddings \/<\/span> np.<\/span>linalg.<\/span>norm(<\/span>chunk_embeddings)<\/span>
\nreturn<\/span> chunk_embeddings.<\/span>tolist(<\/span>)<\/span><\/p>\n

\u5c06\u4e0a\u9762\u4e24\u5904\u4ee3\u7801\u6dfb\u52a0#，\u6ce8\u91ca\u6389，\u64cd\u4f5c\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u5728\u4e0a\u9762\u4e24\u5904\u6ce8\u91ca\u540e\u9762\u6dfb\u52a0\u4ee5\u4e0b\u4e24\u6bb5\u65b0\u4ee3\u7801，\u5176\u4e2d，\u7b2c\u4e00\u5904\u4ee3\u7801\u6a21\u578b\u540d\u79f0\u6839\u636e\u5b9e\u9645\u6a21\u578b\u540d\u79f0\u4fee\u6539，\u672c\u9879\u76ee\u5411\u91cf\u6a21\u578b\u4f7f\u7528\u7684\u662fnomic-embed-text\u3002
\n\u7b2c\u4e00\u5904：<\/p>\n

embedding =<\/span> ollama.<\/span>embeddings(<\/span>model=<\/span>'nomic-embed-text'<\/span>,<\/span> prompt=<\/span>chunk)<\/span>[<\/span>'embedding'<\/span>]<\/span><\/p>\n

\u7b2c\u4e8c\u5904：<\/p>\n

return<\/span> chunk_embeddings<\/p>\n

\u5177\u4f53\u6dfb\u52a0\u4ee3\u7801\u540e，\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u6dfb\u52a0\u597d\u65b0\u4ee3\u7801\u540e\u6309ctrl+o\u4fdd\u5b58，\u4fdd\u5b58\u540e\u6309ctrl+c\u9000\u51fa\u5373\u53ef\u3002<\/p>\n

\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n6.3.\u4fee\u6539qutry\/llm\/text_utils.py\u6e90\u4ee3\u7801\u6587\u4ef6，，\u8fd8\u5728\u5f53\u524d\u6211\u7684\u9879\u76ee\u4e0b\u67e5\u627e\u8be5\u6e90\u4ee3\u7801\u6587\u4ef6，\u4e0b\u9762\u5f53\u524d\u6211\u9879\u76ee\u7684graphrag\u76ee\u5f55.<\/p>\n

\/home\/allanfan\/anaconda3\/envs\/graphrag\/lib\/python3.10\/site-packages\/graphrag<\/p>\n

\u627e\u5230qutry\/llm\/text_utils.py，\u5982\u4e0b\u56fe\u6240\u793a\u3002<\/p>\n

\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u4f7f\u7528sudo nano\u6253\u5f00\u6e90\u4ee3\u7801\u6587\u4ef6，\u6dfb\u52a0\u4ee5\u4e0b\u4e00\u884c\u4ee3\u7801\u5373\u53ef，\u8be5\u4ee3\u7801\u662f\u5c06tokens\u89e3\u7801\u6210\u5b57\u7b26\u4e32\u3002<\/p>\n

tokens =<\/span> token_encoder.<\/span>decode(<\/span>tokens)<\/span> <\/p>\n

\u6dfb\u52a0\u4e0a\u9762\u4ee3\u7801\u4f4d\u7f6e\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u6dfb\u52a0\u597d\u65b0\u4ee3\u7801\u540e\u6309ctrl+o\u4fdd\u5b58，\u4fdd\u5b58\u540e\u6309ctrl+c\u9000\u51fa\u5373\u53ef\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u81f3\u6b64GraphRAG\u73af\u5883\u90e8\u7f72\u5df2\u5b8c\u6210，\u63a5\u4e0b\u6765\u5f00\u59cbGraphRAG\u7684\u77e5\u8bc6\u5e93\u7d22\u5f15\u6784\u5efa\u3001\u77e5\u8bc6\u5e93\u67e5\u8be2\u670d\u52a1\u7b49\u3002<\/p>\n

GraphRAG\u914d\u7f6eollama\u5411\u91cf\u6a21\u578b\u3001\u914d\u7f6e\u672c\u5730vllm\u7684DeepSeek 32B\u63a8\u7406\u6a21\u578b\u3001\u77e5\u8bc6\u5e93\u6784\u5efa\u3001\u793a\u4f8b\u77e5\u8bc6\u5e93\u7684\u67e5\u8be2<\/h2>\n

1.\u521b\u5efa\u77e5\u8bc6\u5e93\u76ee\u5f55，\u672c\u9879\u76ee\u793a\u4f8b\u6c11\u653f\u90e8\u76f8\u5173\u6587\u4ef6\u6784\u5efa\u77e5\u8bc6\u5e93，\u4e2a\u4eba\u5148\u5728\u6c11\u653f\u90e8\u4e0b\u8f7d\u201c\u6c11\u653f\u90e82024\u5e7411\u6708\u6587\u544a\u201d\u7684PDF\u7248，\u901a\u8fc7python\u7a0b\u5e8f\u5c06pdf\u8f6c\u5316\u6210word,\u8f6c\u5316\u7a0b\u5e8f\u5927\u5bb6\u53ef\u4ee5\u91c7\u7528\u5728\u7ebf\u8f6c\u6362，\u6216\u8005WPS\u8f6c\u6362，\u8fd9\u91cc\u5c31\u4e0d\u505a\u7b80\u5355\u7684python\u8f6c\u5316\u7a0b\u5e8f\u5206\u4eab；\u5728\u5f53\u524d\u9879\u76ee\u6839\u76ee\u5f55\u4e0b，\u672c\u4eba\u521b\u5efa\u7684\u77e5\u8bc6\u76ee\u5f55\u540d\u79f0\u4e3aminzheng,\u5177\u4f53\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n2.\u5728\u9879\u76ee\u6839\u636e\u4e0b，\u5bf9minzheng\u76ee\u5f55\u521d\u59cb\u6784\u5efa\u77e5\u8bc6\u5e93\u7d22\u5f15，\u6267\u884c\u547d\u4ee4\u5982\u4e0b\u3002\u6ce8\u610f：\u6267\u884c\u73af\u5883\u4e00\u5b9a\u8981\u5728\u4e0a\u9762\u81ea\u5df1\u547d\u540d\u7684CONDA\u73af\u5883\u4e0b，\u672c\u9879\u76eeCONDA\u73af\u5883\u540d\u79f0\u4e3agraphrag,\u5426\u5219\u6267\u884c\u62a5\u9519\u3002<\/p>\n

graphrag init –root<\/span> .\/minzheng<\/p>\n

\u6267\u884c\u5b8c\u6210\u4e4b\u540e，\u8fdb\u5165minzheng\u76ee\u5f55，\u4f1a\u770b\u5230graphrag\u521d\u59cb\u5316\u521b\u5efa\u7684\u76ee\u5f55\u53ca\u6587\u4ef6，.env\u9879\u76ee\u914d\u7f6e\u6587\u4ef6; settings.yaml\u9879\u76ee\u8fd0\u884c\u914d\u7f6e\u53ca\u4f9d\u8d56\u6587\u4ef6prompts\u9879\u76ee\u63d0\u793a\u8bcd\u5de5\u7a0b，\u8fd9\u4e2a\u6587\u4ef6\u5f88\u91cd\u8981，\u4ed6\u662f\u77e5\u8bc6\u56fe\u8c31\u6700\u6838\u5fc3\u5730\u65b9，\u6784\u5efa\u9879\u76ee\u7b80\u5355，\u4f46\u662f\u9488\u5bf9\u4e1a\u52a1\u6784\u5efa\u7b26\u5408\u4e1a\u52a1\u7684\u5de5\u4f5c\u5168\u5728\u8fd9\u91cc，\u4e00\u4e2a\u597d\u7684\u77e5\u8bc6\u5e93\u5c31\u770b\u4f60\u5199\u7684\u63d0\u793a\u8bcd\u4e86\u3002<\/p>\n

\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n3.\u4fee\u6539\u5f53\u524d\u9879\u76ee\u914d\u7f6e\u6587\u4ef6.env,\u5f53\u524d\u9879\u76ee\u91c7\u7528\u672c\u5730ollama\u90e8\u7f72\u7684\u63a8\u7406\u6a21\u578b\u8fdb\u884c\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u53ca\u67e5\u8be2，\u90e8\u7f72\u73af\u5883\u5df2\u4fee\u6539\u6e90\u4ee3\u7801\u4e3a\u8c03\u7528\u672c\u5730，\u5982\u679c\u4e0d\u4fee\u6539，\u672c\u5730\u9700\u8981\u5728openai\u7533\u8bf7API\u5bc6\u94a5，\u5728\u8be5\u6587\u4ef6\u914d\u7f6e\u5373\u53ef，\u7531\u4e8e\u8c03\u7528\u672c\u5730vllm\u90e8\u7f72\u7684DeepSeek 32B\u63a8\u7406\u6a21\u578b，\u540c\u65f6\u672c\u5730\u5316\u8c03\u7528vllm\u542f\u7528api_key,\u56e0\u6b64\u672c\u914d\u7f6e\u6587\u4ef6\u53ef\u4fee\u6539\u4e3a\u6211\u8bbe\u7f6e\u7684api_key\u5373\u53ef，\u6211\u8bbe\u7f6e\u672c\u5730\u90e8\u7f72\u7684DeepSeek 32B\u7684api_key\u662fallanfan\u3002
\n\u4e3a\u4e86\u4e2a\u4eba\u66f4\u597d\u8bc6\u522b\u662f\u8c03\u7528\u672c\u5730，\u5efa\u8bae\u4fee\u6539\u914d\u7f6e\u4e3a<\/p>\n

GRAPHRAG_API_KEY=allanfan<\/p>\n

\u5177\u4f53\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u4fee\u6539\u597d\u540e\u6309ctrl+o\u4fdd\u5b58，\u4fdd\u5b58\u540e\u6309ctrl+c\u9000\u51fa\u5373\u53ef\u3002<\/p>\n

4.\u4fee\u6539\u5f53\u524d\u9879\u76ee\u914d\u7f6e\u6587\u4ef6settings.yaml,\u8be5\u6587\u4ef6\u539f\u5185\u5bb9\u5982\u4e0b\u3002<\/p>\n

### This config file contains required core defaults that must be set, along with a handful of common optional settings.
\n### For a full list of available settings, see https:\/\/microsoft.github.io\/graphrag\/config\/yaml\/<\/p>\n

### LLM settings ###
\n## There are a number of settings to tune the threading and token limits for LLM calls – check the docs.<\/p>\n

encoding_model: cl100k_base # this needs to be matched to your model!<\/p>\n

llm:
\n api_key: ${GRAPHRAG_API_KEY} # set this in the generated .env file
\n type: openai_chat # or azure_openai_chat
\n model: gpt-4-turbo-preview
\n model_supports_json: true # recommended if this is available for your model.
\n # audience: "https:\/\/cognitiveservices.azure.com\/.default"
\n # api_base: https:\/\/<<\/span>instance<\/span>><\/span><\/span>.openai.azure.com
\n # api_version: 2024-02-15-preview
\n # organization: <<\/span>organization_id<\/span>><\/span><\/span>
\n # deployment_name: <<\/span>azure_model_deployment_name<\/span>><\/span><\/span><\/p>\n

parallelization:
\n stagger: 0.3
\n # num_threads: 50<\/p>\n

async_mode: threaded # or asyncio<\/p>\n

embeddings:
\n async_mode: threaded # or asyncio
\n vector_store:
\n type: lancedb
\n db_uri: 'output\/lancedb'
\n container_name: default
\n overwrite: true
\n llm:
\n api_key: ${GRAPHRAG_API_KEY}
\n type: openai_embedding # or azure_openai_embedding
\n model: text-embedding-3-small
\n # api_base: https:\/\/<<\/span>instance<\/span>><\/span><\/span>.openai.azure.com
\n # api_version: 2024-02-15-preview
\n # audience: "https:\/\/cognitiveservices.azure.com\/.default"
\n # organization: <<\/span>organization_id<\/span>><\/span><\/span>
\n # deployment_name: <<\/span>azure_model_deployment_name<\/span>><\/span><\/span><\/p>\n

### Input settings ###<\/p>\n

input:
\n type: file # or blob
\n file_type: text # or csv
\n base_dir: "input"
\n file_encoding: utf-8
\n file_pattern: ".*\\\\\\\\.txt$"<\/p>\n

chunks:
\n size: 1200
\n overlap: 100
\n group_by_columns: [id]<\/p>\n

### Storage settings ###
\n## If blob storage is specified in the following four sections,
\n## connection_string and container_name must be provided<\/p>\n

cache:
\n type: file # or blob
\n base_dir: "cache"<\/p>\n

reporting:
\n type: file # or console, blob
\n base_dir: "logs"<\/p>\n

storage:
\n type: file # or blob
\n base_dir: "output"<\/p>\n

## only turn this on if running `graphrag index` with custom settings
\n## we normally use `graphrag update` with the defaults
\nupdate_index_storage:
\n # type: file # or blob
\n # base_dir: "update_output"<\/p>\n

### Workflow settings ###<\/p>\n

skip_workflows: []<\/p>\n

entity_extraction:
\n prompt: "prompts\/entity_extraction.txt"
\n entity_types: [organization,person,geo,event]
\n max_gleanings: 1<\/p>\n

summarize_descriptions:
\n prompt: "prompts\/summarize_descriptions.txt"
\n max_length: 500<\/p>\n

claim_extraction:
\n enabled: false
\n prompt: "prompts\/claim_extraction.txt"
\n description: "Any claims or facts that could be relevant to information discovery."
\n max_gleanings: 1<\/p>\n

community_reports:
\n prompt: "prompts\/community_report.txt"
\n max_length: 2000
\n max_input_length: 8000<\/p>\n

cluster_graph:
\n max_cluster_size: 10<\/p>\n

embed_graph:
\n enabled: false # if true, will generate node2vec embeddings for nodes<\/p>\n

umap:
\n enabled: false # if true, will generate UMAP embeddings for nodes<\/p>\n

snapshots:
\n graphml: false
\n raw_entities: false
\n top_level_nodes: false
\n embeddings: false
\n transient: false<\/p>\n

### Query settings ###
\n## The prompt locations are required here, but each search method has a number of optional knobs that can be tuned.
\n## See the config docs: https:\/\/microsoft.github.io\/graphrag\/config\/yaml\/#query<\/p>\n

local_search:
\n prompt: "prompts\/local_search_system_prompt.txt"<\/p>\n

global_search:
\n map_prompt: "prompts\/global_search_map_system_prompt.txt"
\n reduce_prompt: "prompts\/global_search_reduce_system_prompt.txt"
\n knowledge_prompt: "prompts\/global_search_knowledge_system_prompt.txt"<\/p>\n

drift_search:
\n prompt: "prompts\/drift_search_system_prompt.txt"<\/p>\n

\u4e0a\u9762\u6587\u4ef6\u6211\u4eec\u9700\u8981\u5173\u6ce8\u51e0\u4e2a\u5730\u65b9，\u4e00\u4e2a\u662fllm\u914d\u7f6e\u9879，\u4e00\u4e2a\u662fembeddings\u914d\u7f6e\u9879\u3002<\/p>\n

\u9996\u5148\u9488\u5bf9settings.yaml\u91cc\u9762\u7684llm\u914d\u7f6e\u4fee\u6539\u5982\u4e0b：
\n\u539f\u6587\u4ef6llm\u4ee3\u7801<\/p>\n

llm:
\n api_key: ${GRAPHRAG_API_KEY} # set this in the generated .env file
\n type: openai_chat # or azure_openai_chat
\n model: gpt-4-turbo-preview
\n model_supports_json: true # recommended if this is available for your model.
\n # audience: "https:\/\/cognitiveservices.azure.com\/.default"
\n # api_base: https:\/\/<<\/span>instance<\/span>><\/span><\/span>.openai.azure.com
\n # api_version: 2024-02-15-preview
\n # organization: <<\/span>organization_id<\/span>><\/span><\/span>
\n # deployment_name: <<\/span>azure_model_deployment_name<\/span>><\/span><\/span><\/p>\n

\u4fee\u6539\u540ellm\u4ee3\u7801<\/p>\n

llm:
\n api_key: ${GRAPHRAG_API_KEY} # \u53d6\u914d\u7f6e\u6587\u4ef6.env\u7684APIKey，\u4e0d\u8fdb\u884c\u4fee\u6539
\n type: openai_chat # \u5f53\u524d\u53d6\u672c\u5730\u79c1\u6709\u5316\u90e8\u7f72\u7684DeepSeek 32B,DeepSeek\u5ef6\u7528openai\u534f\u8bae
\n model: DeepSeek-R1-Distill-Qwen-32B # \u5f53\u524d\u6a21\u578b\u540d\u79f0\u66f4\u6362\u4e3a\u672c\u5730\u79c1\u5316\u90e8\u7f72\u7684deepseek 32B\u6a21\u578b\u540d\u79f0
\n model_supports_json: true # recommended if this is available for your model.
\n api_base: http:\/\/localhost:8001\/v1 # \u6dfb\u52a0\u672c\u5730\u5316\u6a21\u578b\u5bf9\u5916\u53d1\u5e03\u8c03\u7528\u7684\u5730\u5740,\u8be5\u5730\u5740\u586b\u5199\u81ea\u5df1\u90e8\u7f72\u7684\u5730\u5740\u5373\u53ef
\n # audience: "https:\/\/cognitiveservices.azure.com\/.default"
\n # api_base: https:\/\/<<\/span>instance<\/span>><\/span><\/span>.openai.azure.com
\n # api_version: 2024-02-15-preview
\n # organization: <<\/span>organization_id<\/span>><\/span><\/span>
\n # deployment_name: <<\/span>azure_model_deployment_name<\/span>><\/span><\/span><\/p>\n

\u4fee\u6539\u540e\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"<\/p>\n

\u7b2c\u4e8c\u6b65\u9488\u5bf9settings.yaml\u91cc\u9762\u7684embeddings\u914d\u7f6e\u4fee\u6539\u5982\u4e0b：
\n\u539f\u6587\u4ef6embeddings\u7684\u4ee3\u7801<\/p>\n

embeddings:
\n async_mode: threaded # or asyncio
\n vector_store:
\n type: lancedb
\n db_uri: 'output\/lancedb'
\n container_name: default
\n overwrite: true
\n llm:
\n api_key: ${GRAPHRAG_API_KEY}
\n type: openai_embedding # or azure_openai_embedding
\n model: text-embedding-3-small
\n # api_base: https:\/\/<<\/span>instance<\/span>><\/span><\/span>.openai.azure.com
\n # api_version: 2024-02-15-preview
\n # audience: "https:\/\/cognitiveservices.azure.com\/.default"
\n # organization: <<\/span>organization_id<\/span>><\/span><\/span>
\n # deployment_name: <<\/span>azure_model_deployment_name<\/span>><\/span><\/span><\/p>\n

\u4fee\u6539\u540eembeddings\u7684\u4ee3\u7801\u5982\u4e0b<\/p>\n

embeddings:
\n async_mode: threaded # or asyncio
\n vector_store:
\n type: lancedb
\n db_uri: 'output\/lancedb'
\n container_name: default
\n overwrite: true
\n llm:
\n api_key: ollama # \u5f53\u524d\u5411\u91cf\u6a21\u578b\u4f7f\u7528\u7684\u662follama\u90e8\u7f72\u7684,ollama\u672a\u542f\u7528api_key,\u65b9\u4fbf\u8ba4\u77e5\u662f\u8c03\u7528\u672c\u5730ollama\u8fd0\u884c\u7684\u5411\u91cf\u6a21\u578b，\u8be5\u5904\u586b\u5199ollama
\n type: openai_embedding # \u6a21\u578b\u5ef6\u7528openai\u7684\u5411\u91cf\u6a21\u578b\u534f\u8bae，\u8be5\u5904\u4e0d\u8fdb\u884c\u4fee\u6539
\n model: nomic-embed-text # \u672c\u9879\u76ee\u91c7\u7528\u7684\u5411\u91cf\u6a21\u578b\u540d\u79f0\u662fnomic-embed-text，\u8be5\u5904\u4fee\u6539\u4e3anomic-embed-text
\n api_base: http:\/\/localhost:11434\/api # \u672c\u9879\u76ee\u8c03\u7528\u662f\u672c\u5730\u90e8\u7f72\u7684\u5411\u91cf\u6a21\u578b，\u8be5\u5730\u5740\u5730\u5740\u4fee\u6539\u6210\u672c\u5730\u90e8\u7f72\u63a8\u7406\u7684\u5730\u5740\u5373\u53ef
\n # api_version: 2024-02-15-preview
\n # audience: "https:\/\/cognitiveservices.azure.com\/.default"
\n # organization: <<\/span>organization_id<\/span>><\/span><\/span>
\n # deployment_name: <<\/span>azure_model_deployment_name<\/span>><\/span><\/span><\/p>\n

\u4fee\u6539\u540e\u5982\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u4fee\u6539\u597d\u914d\u7f6e\u540e\u6309ctrl+o\u4fdd\u5b58，\u4fdd\u5b58\u540e\u6309ctrl+c\u9000\u51fa\u5373\u53ef\u3002<\/p>\n

5.\u5728minzheng\u76ee\u5f55\u4e0b\u521b\u5efainput\u6587\u4ef6\u5939，\u7528\u4e8e\u6784\u5efa\u77e5\u8bc6\u7684\u6587\u672c\u5185\u5bb9，\u672c\u9879\u76ee\u793a\u4f8b\u4e0a\u4f20\u5728\u6c11\u653f\u90e8\u4e0b\u8f7d\u7684\u201c\u6c11\u653f\u90e82024\u5e7411\u6708\u6587\u544a\u201d，\u7528\u4e8e\u6784\u5efa\u77e5\u8bc6\u5e93\u5185\u5bb9\u3002\u5177\u4f53\u8be6\u89c1\u4e0b\u56fe\u6240\u793a\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n6.\u6784\u5efaGraphRAG\u7d22\u5f15，\u5728\u5f53\u524dgraphrag\u53ca\u4e0a\u9762\u5b89\u88c5\u7684\u73af\u5883\u4e0b\u6267\u884c\u6784\u5efa\u7d22\u5f15\u547d\u4ee4，\u7a0b\u5e8f\u4f1a\u6839\u636e\u4e0a\u9762\u914d\u7f6e\u6587\u4ef6\u81ea\u52a8\u8c03\u7528\u672c\u5730ollama\u53ca\u672c\u5730vllm\u7684\u5411\u91cf\u6a21\u578b\u53ca\u63a8\u7406\u6a21\u578b\u3002<\/p>\n

graphrag index –root<\/span> .\/minzheng –reporter<\/span> rich<\/p>\n

\u6267\u884c\u6548\u679c\u5982\u4e0b\u56fe\u6240\u793a\u3002<\/p>\n

\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u6211\u63d0\u4f9b\u7684\u6570\u636e\u91cf\u4e0d\u5927，\u6267\u884c\u6574\u4e2a\u8fc7\u7a0b10\u5206\u949f\u4ee5\u5185\u5b8c\u6210\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u4e0a\u56fe\u662f\u672c\u5730\u63a8\u7406\u6a21\u578b\u8fdb\u884c\u5b9e\u4f53\u53ca\u5173\u7cfb\u7684\u63d0\u53d6\u3002<\/p>\n

7.\u6784\u5efa\u5b8c\u6210\u7684GraphRAG\u7684\u793a\u4f8b\u9879\u76ee\u5185\u5bb9\u67e5\u8be2，\u4f8b\u5982：\u6211\u67e5\u8be2\u201c\u56f0\u5883\u513f\u7ae5\u201d\u53ca\u201c\u4e2d\u56fd\u798f\u5229\u5f69\u7968\u201d，\u770b\u77e5\u8bc6\u56fe\u8c31\u5982\u4f55\u51c6\u786e\u7684\u56de\u7b54\u6211\u3002
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"<\/p>\n

\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u5269\u4e0b\u5927\u5bb6\u81ea\u5df1\u90e8\u7f72\u53ca\u6d4b\u8bd5\u5427！\u5bf9\u4e86，\u5982\u679c\u4f60\u7684\u670d\u52a1\u5668GPU\u8d44\u6e90\u4e0d\u591f，\u53ef\u4ee5\u5728\u914d\u7f6e\u6587\u4ef6\u4e2d\u4fee\u6539\u6210\u4f60\u5728\u7f51\u4e0a\u7533\u8bf7DEEPSEEK\u7684API\u5730\u5740\u53caKEY\u5373\u53ef，\u5982\u4e0b\u6240\u793a\u3002<\/p>\n

llm:
\n api_key: API_KEY # \u76f4\u63a5\u586b\u5199\u7533\u8bf7\u7684API_KEY,\u6216\u8005\u5728.env\u91cc\u9762\u4fee\u6539\u4e5f\u53ef\u4ee5
\n type: openai_chat # \u5f53\u524d\u53d6\u672c\u5730\u79c1\u6709\u5316\u90e8\u7f72\u7684DeepSeek 32B,DeepSeek\u5ef6\u7528openai\u534f\u8bae
\n model: deepseek-chat # DEEPSEEK\u4f7f\u7528\u6a21\u578b\u540d\u79f0 deepseek-chat
\n model_supports_json: true # recommended if this is available for your model.
\n api_base: https:\/\/api.deepseek.com # Deepseek\u8c03\u7528\u7684\u5730\u5740,deepseek\u662f\u5ef6\u7528openai\u5730\u5740\u534f\u8bae，\u540e\u9762\u7684v1\/chat\/completions\u9ed8\u8ba4\u81ea\u52a8\u8865\u9f50
\n # audience: "https:\/\/cognitiveservices.azure.com\/.default"
\n # api_base: https:\/\/<<\/span>instance<\/span>><\/span><\/span>.openai.azure.com
\n # api_version: 2024-02-15-preview
\n # organization: <<\/span>organization_id<\/span>><\/span><\/span>
\n # deployment_name: <<\/span>azure_model_deployment_name<\/span>><\/span><\/span><\/p>\n

8.\u4f7f\u7528APP\u6784\u5efaGraphRAG\u5bf9\u5916\u7684\u63a5\u53e3\u4ee3\u7801\u53ca\u793a\u4f8b\u3002
\n8.1.\u5bf9\u5916\u63a5\u53e3\u8def\u7531\u4ee3\u7801<\/p>\n

#===================================================================#<\/span>
\n#===================================================================#<\/span>
\n# GraphRAG\u5bf9\u5916\u63a5\u53e3\u5e94\u7528V1.0-Allan.Fan 20250525<\/span>
\n# \u6a21\u578b\u67e5\u8be2\u6784\u5efa\u65b9\u6cd5\u5b9e\u73b0<\/span>
\n#===================================================================#<\/span>
\n#===================================================================#<\/span>
\nimport<\/span> json
\nfrom<\/span> datetime import<\/span> datetime
\nfrom<\/span> flask import<\/span> Blueprint,<\/span> request,<\/span> g
\nfrom<\/span> services.<\/span>appquerymodeloutput_service import<\/span> get_appquerymodeloutput
\nfrom<\/span> flask_restx import<\/span> Api,<\/span>Namespace,<\/span> Resource,<\/span> reqparse<\/p>\n

appquerymodeloutput_bp =<\/span> Blueprint(<\/span>'querymodel_bp'<\/span>,<\/span> __name__,<\/span> url_prefix=<\/span>"\/app\/query\/model"<\/span>)<\/span>
\napi =<\/span> Api(<\/span>appquerymodeloutput_bp,<\/span> doc=<\/span>'\/doc'<\/span>,<\/span> version=<\/span>'1.0'<\/span>,<\/span> title=<\/span>'\u6a21\u578b\u67e5\u8be2\u63a5\u53e3\u6587\u6863'<\/span>,<\/span> description=<\/span>'\u5173\u4e8e\u4f01\u4e1a\u6a21\u578b\u6587\u6863\u8bf4\u660e'<\/span>)<\/span><\/p>\n

# ————————————————————<\/span>
\n# \u521b\u5efa\u9ed8\u8ba4\u547d\u540d\u7a7a\u95f4<\/span>
\n# ————————————————————<\/span>
\ndefault_ns =<\/span> Namespace(<\/span>
\n name=<\/span>""<\/span>,<\/span>
\n path=<\/span>""<\/span>,<\/span>
\n description=<\/span>"\u6a21\u578b\u67e5\u8be2\u63a5\u53e3\u6587\u6863"<\/span>
\n)<\/span>
\napi.<\/span>add_namespace(<\/span>default_ns)<\/span><\/p>\n

# ———————————————-<\/span>
\n# \u53c2\u6570\u89e3\u6790\u5668<\/span>
\n# ———————————————-<\/span>
\nparser =<\/span> reqparse.<\/span>RequestParser(<\/span>)<\/span>
\nparser.<\/span>add_argument(<\/span>'query'<\/span>,<\/span> type<\/span>=<\/span>str<\/span>,<\/span> default=<\/span>''<\/span>,<\/span> required=<\/span>True<\/span>,<\/span>
\n help<\/span>=<\/span>'\u5411\u4f01\u4e1a\u6a21\u578b\u63d0\u51fa\u95ee\u9898'<\/span>,<\/span> location=<\/span>'args'<\/span>)<\/span><\/p>\n

@appquerymodeloutput_bp.<\/span>before_request<\/span>
\ndef<\/span> before_request<\/span>(<\/span>)<\/span>:<\/span>
\n g.<\/span>client_ip =<\/span> request.<\/span>headers.<\/span>get(<\/span>'X-Forwarded-For'<\/span>,<\/span> request.<\/span>remote_addr)<\/span> <\/p>\n

@default_ns.<\/span>route<\/span>(<\/span>'\/output'<\/span>,<\/span>methods=<\/span>[<\/span>'GET'<\/span>]<\/span>)<\/span>
\nclass<\/span> AppQueryModelOutput<\/span>(<\/span>Resource)<\/span>:<\/span>
\n @default_ns.<\/span>expect<\/span>(<\/span>parser)<\/span>
\n def<\/span> get<\/span>(<\/span>self)<\/span>:<\/span>
\n """\u4f01\u4e1a\u6a21\u578b\u5e94\u7528"""<\/span>
\n args =<\/span> parser.<\/span>parse_args(<\/span>)<\/span>
\n if<\/span> args[<\/span>'query'<\/span>]<\/span> ==<\/span> None<\/span>:<\/span>
\n return<\/span> {<\/span>'code'<\/span>:<\/span> 400<\/span>,<\/span> 'msg'<\/span>:<\/span> '\u63d0\u51fa\u95ee\u9898\u4e0d\u80fd\u4e3a\u7a7a'<\/span>}<\/span><\/p>\n

query =<\/span> args[<\/span>'query'<\/span>]<\/span>
\n print<\/span> (<\/span>f"<\/span>{<\/span>datetime.<\/span>now(<\/span>)<\/span>.<\/span>strftime(<\/span>'%Y-%m-%d %H:%M:%S'<\/span>)<\/span>}<\/span><\/span> query: <\/span>{<\/span>query}<\/span><\/span>"<\/span><\/span>)<\/span><\/p>\n

if<\/span> not<\/span> query:<\/span>
\n return<\/span> {<\/span>"success"<\/span>:<\/span> False<\/span>,<\/span> "message"<\/span>:<\/span> "\u63d0\u51fa\u95ee\u9898\u4e0d\u80fd\u4e3a\u7a7a!"<\/span>}<\/span>,<\/span> 400<\/span>
\n try<\/span>:<\/span>
\n istrue,<\/span>querymodeloutput_data =<\/span> get_appquerymodeloutput(<\/span>query)<\/span>
\n if<\/span> istrue:<\/span>
\n return<\/span> {<\/span>"success"<\/span>:<\/span> istrue,<\/span>"message"<\/span>:<\/span> "\u6a21\u578b\u8bf7\u6c42\u5904\u7406\u6210\u529f"<\/span>,<\/span>"data"<\/span>:<\/span> querymodeloutput_data}<\/span>,<\/span> 200<\/span>
\n else<\/span>:<\/span>
\n return<\/span> {<\/span>"success"<\/span>:<\/span> istrue,<\/span> "message"<\/span>:<\/span> f"\u6a21\u578b\u8bf7\u6c42\u5904\u7406\u5931\u8d25!"<\/span><\/span>,<\/span> "data"<\/span>:<\/span> querymodeloutput_data}<\/span>,<\/span> 200<\/span>
\n except<\/span> Exception as<\/span> e:<\/span>
\n return<\/span> {<\/span>"success"<\/span>:<\/span> False<\/span>,<\/span> "message"<\/span>:<\/span> str<\/span>(<\/span>e)<\/span>}<\/span>,<\/span> 500<\/span> <\/p>\n

8.2.GraphRAG\u83b7\u53d6\u67e5\u8be2\u95ee\u9898\u540e\u8fdb\u884c\u68c0\u7d22\u7684\u4e1a\u52a1\u4ee3\u7801\u3002<\/p>\n

#===================================================================#<\/span>
\n#===================================================================#<\/span>
\n# GraphRAG\u5bf9\u5916\u63a5\u53e3\u5e94\u7528V1.0-Allan.Fan 20250525<\/span>
\n# \u6a21\u578b\u67e5\u8be2\u76f8\u5173API\u65b9\u6cd5\u5b9e\u73b0<\/span>
\n#===================================================================#<\/span>
\n#===================================================================#<\/span>
\nimport<\/span> subprocess
\ndef<\/span> get_appquerymodeloutput<\/span>(<\/span>query)<\/span>:<\/span>
\n result =<\/span> [<\/span>]<\/span>
\n taskoneoutput =<\/span> {<\/span>
\n "msg"<\/span>:<\/span>None<\/span>,<\/span>
\n "count"<\/span>:<\/span>0<\/span>,<\/span>
\n "data"<\/span>:<\/span>[<\/span>]<\/span>
\n }<\/span>
\n try<\/span>:<\/span>
\n command =<\/span> [<\/span>'graphrag'<\/span>,<\/span>'query'<\/span>,<\/span>'–root'<\/span>,<\/span> '\/home\/allanfan\/code\/graphrag\/minzheng'<\/span>,<\/span>'–method'<\/span>,<\/span> 'global'<\/span>,<\/span>'–query'<\/span>,<\/span> query]<\/span>
\n result =<\/span> subprocess.<\/span>run(<\/span>command,<\/span> capture_output=<\/span>True<\/span>,<\/span> text=<\/span>True<\/span>)<\/span>
\n taskoneoutput[<\/span>"data"<\/span>]<\/span> =<\/span> result.<\/span>stdout
\n return<\/span> True<\/span>,<\/span>taskoneoutput
\n except<\/span> Exception as<\/span> e:<\/span>
\n return<\/span> False<\/span>,<\/span>str<\/span>(<\/span>e)<\/span><\/p>\n

\u5907\u6ce8：\/home\/allanfan\/code\/graphrag\/minzheng\u8fd9\u4e2a\u5730\u5740\u662f\u4e0a\u9762\u6784\u5efa\u77e5\u8bc6\u56fe\u8c31\u5185\u5bb9\u7684\u5728\u670d\u52a1\u5668\u7684\u7edd\u5bf9\u5730\u5740，\u8bf7\u4fee\u6539\u6210\u81ea\u5df1\u7684\u670d\u52a1\u5668\u6784\u5efa\u7684\u9879\u76ee\u7edd\u5bf9\u5730\u5740\u5373\u53ef，\u4e0a\u9762\u662f\u6211\u793a\u4f8b\u9879\u76ee\u5728\u670d\u52a1\u5668\u7684\u7edd\u5bf9\u5730\u5740\u4f4d\u7f6e\u3002<\/p>\n

\u4ee5\u4e0a\u5c31\u662fGraphRAG\u4ece\u73af\u5883\u90e8\u7f72，\u52a0\u8f7d\u672c\u5730\u5411\u91cf\u6a21\u578b\u3001\u672c\u5730\u63a8\u7406\u6a21\u578b\u3001\u7b80\u5355\u7684\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u7b80\u5355\u7684\u5bf9\u5916API\u63a5\u53e3\u5f00\u53d1；\u5269\u4e0b\u6700\u590d\u6742\u7684\u5c31\u662f\u77e5\u8bc6\u56fe\u8c31\u7684\u5b9e\u4f53\u3001\u5173\u7cfb\u5982\u4f55\u6839\u636e\u4e1a\u52a1\u6784\u5efa，\u4ee5\u53ca\u68c0\u7d22\u7684\u6570\u636e\u53ec\u56de\u7387\u7ba1\u7406，\u68c0\u7d22\u6570\u636e\u901f\u5ea6\u53ca\u51c6\u786e\u7387\u7684\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa，\u8fd9\u4e2a\u662f\u9664\u6280\u672f\u4e4b\u5916\u8fd8\u8981\u7ed3\u5408\u4e1a\u52a1\u53bb\u6574\u5408\u53d1\u5c55\u3002\u5e0c\u671b\u4e0a\u9762\u7684GraphRAG\u80fd\u591f\u5728\u4e1a\u52a1\u4e2d\u5e2e\u52a9\u81f3\u5927\u5bb6；\u5982\u679c\u80fd\u591f\u5e2e\u52a9\u5230\u4f60\u4e00\u70b9，\u4e5f\u5e0c\u671b\u4f60\u591a\u652f\u6301\u652f\u6301\u6211，\u7ed9\u6211\u591a\u70b9\u51fb\u53ca\u6536\u85cf；\u4e5f\u671f\u671b\u4f60\u7684\u5173\u6ce8\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"

\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb1k\u6b21\uff0c\u70b9\u8d5e36\u6b21\uff0c\u6536\u85cf17\u6b21\u3002\u672c\u6587\u4ecb\u7ecd\u4e86\u57284\u53614090\u670d\u52a1\u5668\u4e0a\u90e8\u7f72GraphRAG\u73af\u5883\u7684\u8fc7\u7a0b\u3002\u786c\u4ef6\u914d\u7f6e\u5305\u62ec\u53cc\u8def\u81f3\u5f3a\u91d1\u724c6148\u5904\u7406\u5668\u3001128GB\u5185\u5b58\u548c4\u57574090\u663e\u5361\u3002\u521b\u5efaPython3.10\u865a\u62df\u73af\u5883\u3001\u5b89\u88c5GraphRAG 0.5.0\u53ca\u76f8\u5173\u4f9d\u8d56\u5305\uff0c\u4ee5\u53ca\u4fee\u6539\u6e90\u4ee3\u7801\u4ee5\u652f\u6301\u672c\u5730ollama\u5411\u91cf\u6a21\u578b\u548cvllm\u63a8\u7406\u6a21\u578b\u3002\u91cd\u70b9\u8bf4\u660e\u4e86\u5982\u4f55\u4fee\u6539openai_embeddings_llm.py\u548cembedding.py\u6587\u4ef6\uff0c\u5c06\u539fOpenAI\u63a5\u53e3\u66ff\u6362\u4e3a\u672c\u5730ollama\u63a5\u53e3\uff0c\u5b9e\u73b0\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u548c\u672c\u5730\u5927\u6a21\u578b\u63a8\u7406\u7684\u65e0\u7f1d\u63a5\u5165\u3002\u9002\u7528\u4e8e\u9700\u8981\u6570\u636e\u4fdd\u5bc6\u7684\u4f01\u4e1a\u73af\u5883._graphrag 0.5.0<\/p>\n","protected":false},"author":2,"featured_media":48444,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[66,3651,346,347,126,4119,50,43,613],"topic":[],"class_list":["post-48464","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-server","tag-ai","tag-deep-learning","tag-embedding","tag-llama","tag-ollama","tag-rag","tag-50","tag-43","tag-613"],"yoast_head":"\n4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/48464.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb1k\u6b21\uff0c\u70b9\u8d5e36\u6b21\uff0c\u6536\u85cf17\u6b21\u3002\u672c\u6587\u4ecb\u7ecd\u4e86\u57284\u53614090\u670d\u52a1\u5668\u4e0a\u90e8\u7f72GraphRAG\u73af\u5883\u7684\u8fc7\u7a0b\u3002\u786c\u4ef6\u914d\u7f6e\u5305\u62ec\u53cc\u8def\u81f3\u5f3a\u91d1\u724c6148\u5904\u7406\u5668\u3001128GB\u5185\u5b58\u548c4\u57574090\u663e\u5361\u3002\u521b\u5efaPython3.10\u865a\u62df\u73af\u5883\u3001\u5b89\u88c5GraphRAG 0.5.0\u53ca\u76f8\u5173\u4f9d\u8d56\u5305\uff0c\u4ee5\u53ca\u4fee\u6539\u6e90\u4ee3\u7801\u4ee5\u652f\u6301\u672c\u5730ollama\u5411\u91cf\u6a21\u578b\u548cvllm\u63a8\u7406\u6a21\u578b\u3002\u91cd\u70b9\u8bf4\u660e\u4e86\u5982\u4f55\u4fee\u6539openai_embeddings_llm.py\u548cembedding.py\u6587\u4ef6\uff0c\u5c06\u539fOpenAI\u63a5\u53e3\u66ff\u6362\u4e3a\u672c\u5730ollama\u63a5\u53e3\uff0c\u5b9e\u73b0\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u548c\u672c\u5730\u5927\u6a21\u578b\u63a8\u7406\u7684\u65e0\u7f1d\u63a5\u5165\u3002\u9002\u7528\u4e8e\u9700\u8981\u6570\u636e\u4fdd\u5bc6\u7684\u4f01\u4e1a\u73af\u5883._graphrag 0.5.0\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/48464.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2025-07-30T03:59:38+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730035918-6889989651256.png\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"9 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/48464.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/48464.html\",\"name\":\"4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2025-07-30T03:59:38+00:00\",\"dateModified\":\"2025-07-30T03:59:38+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/48464.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/48464.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/48464.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/48464.html","og_locale":"zh_CN","og_type":"article","og_title":"4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb1k\u6b21\uff0c\u70b9\u8d5e36\u6b21\uff0c\u6536\u85cf17\u6b21\u3002\u672c\u6587\u4ecb\u7ecd\u4e86\u57284\u53614090\u670d\u52a1\u5668\u4e0a\u90e8\u7f72GraphRAG\u73af\u5883\u7684\u8fc7\u7a0b\u3002\u786c\u4ef6\u914d\u7f6e\u5305\u62ec\u53cc\u8def\u81f3\u5f3a\u91d1\u724c6148\u5904\u7406\u5668\u3001128GB\u5185\u5b58\u548c4\u57574090\u663e\u5361\u3002\u521b\u5efaPython3.10\u865a\u62df\u73af\u5883\u3001\u5b89\u88c5GraphRAG 0.5.0\u53ca\u76f8\u5173\u4f9d\u8d56\u5305\uff0c\u4ee5\u53ca\u4fee\u6539\u6e90\u4ee3\u7801\u4ee5\u652f\u6301\u672c\u5730ollama\u5411\u91cf\u6a21\u578b\u548cvllm\u63a8\u7406\u6a21\u578b\u3002\u91cd\u70b9\u8bf4\u660e\u4e86\u5982\u4f55\u4fee\u6539openai_embeddings_llm.py\u548cembedding.py\u6587\u4ef6\uff0c\u5c06\u539fOpenAI\u63a5\u53e3\u66ff\u6362\u4e3a\u672c\u5730ollama\u63a5\u53e3\uff0c\u5b9e\u73b0\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u548c\u672c\u5730\u5927\u6a21\u578b\u63a8\u7406\u7684\u65e0\u7f1d\u63a5\u5165\u3002\u9002\u7528\u4e8e\u9700\u8981\u6570\u636e\u4fdd\u5bc6\u7684\u4f01\u4e1a\u73af\u5883._graphrag 0.5.0","og_url":"https:\/\/www.wsisp.com\/helps\/48464.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2025-07-30T03:59:38+00:00","og_image":[{"url":"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2025\/07\/20250730035918-6889989651256.png"}],"author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"9 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/48464.html","url":"https:\/\/www.wsisp.com\/helps\/48464.html","name":"4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2025-07-30T03:59:38+00:00","dateModified":"2025-07-30T03:59:38+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/48464.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/48464.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/48464.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"4\u53614090\u670d\u52a1\u5668GraphRAG\u73af\u5883\u90e8\u7f72\u3001\u52a0\u8f7dollama\u5411\u91cf\u6a21\u578b\u53ca\u5411\u91cf\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u3001\u52a0\u8f7dvllm\u672c\u5730DeepSeek 32B\u63a8\u7406\u6a21\u578b\u65e0\u7f1d\u63a5\u5165\u53caAPI\u63a5\u53e3\u8c03\u7528"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/48464","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=48464"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/48464\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media\/48444"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=48464"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=48464"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=48464"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=48464"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}