pip install llama_index

pip install llama_index.llms.ollama

pip install langchain

export HF_ENDPOINT=https://hf-mirror.com

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, Settingsfrom llama_index.llms.ollama import Ollama
# 配置ollama的LLM模型，这里我们用gemma:7bSettings.llm = Ollama(model="gemma", request_timeout=600.0)
# 配置HuggingFaceEmbeddings嵌入模型，这里我们用BAAI/bge-small-zh-v1.5from langchain_community.embeddings import HuggingFaceEmbeddingsSettings.embed_model = HuggingFaceEmbeddings(model_name="BAAI/bge-small-zh-v1.5")
# 将data文件夹下的文档建立索引documents = SimpleDirectoryReader("data").load_data()index = VectorStoreIndex.from_documents(documents)
# 创建问答引擎，并提一个简单的问题query_engine = index.as_query_engine()response = query_engine.query("流程有哪三个特征？")print(response)

python app.py

import loggingimport sys# 增加日志信息logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)logging.getLogger().addHandler(logging.StreamHandler(stream=sys.stdout))

# 使用moonshot的apifrom langchain_openai import ChatOpenAIllm_langchain = ChatOpenAI(openai_api_base="https://api.moonshot.cn/v1/",     openai_api_key="input your key here",model_name="moonshot-v1-32k",)Settings.llm = llm_langchain

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, ServiceContext, Documentfrom llama_index.llms.ollama import Ollama
# 配置ollama的LLM模型，这里我们用gemma:7bllm = Ollama(model="gemma", request_timeout=600.0)
# 配置HuggingFaceEmbeddings嵌入模型，这里我们用BAAI/bge-small-zh-v1.5from langchain_community.embeddings import HuggingFaceEmbeddingsembed_model = HuggingFaceEmbeddings(model_name="BAAI/bge-small-zh-v1.5")
# 配置Streamlit Web应用import streamlit as st
st.set_page_config(page_title="本地大模型知识库RAG应用", page_icon="?", layout="centered", initial_sidebar_state="auto", menu_items=None)st.title("本地大模型知识库RAG应用")st.info("By 大卫", icon="?")
if "messages" not in st.session_state.keys(): # 初始化聊天历史记录st.session_state.messages = [{"role": "assistant", "content": "关于文档里的内容，请随便问"}]
@st.cache_resource(show_spinner=False)def load_data():with st.spinner(text="Loading and indexing the Streamlit docs – hang tight! This should take 1-2 minutes."):reader = SimpleDirectoryReader(input_dir="./data", recursive=True) # 将data文件夹下的文档建立索引docs = reader.load_data()service_context = ServiceContext.from_defaults(llm=llm,embed_model=embed_model) # 使用上面配置的模型index = VectorStoreIndex.from_documents(docs, service_context=service_context)return index
index = load_data()
if "chat_engine" not in st.session_state.keys(): # 初始化聊天引擎st.session_state.chat_engine = index.as_chat_engine(chat_mode="condense_question", verbose=True)
if prompt := st.chat_input("Your question"): # 提示用户输入问题，并将问题添加到消息历史记录st.session_state.messages.append({"role": "user", "content": prompt})
for message in st.session_state.messages: # 显示此前的问答记录with st.chat_message(message["role"]):st.write(message["content"])
if st.session_state.messages[-1]["role"] != "assistant": # 生成回答with st.chat_message("assistant"):with st.spinner("Thinking..."):response = st.session_state.chat_engine.chat(prompt)st.write(response.response)message = {"role": "assistant", "content": response.response}st.session_state.messages.append(message) # 将回答加入到消息历史记录

streamlit run app.py