使用 LlamaChat 和 Excel 构建一个简单的聊天机器人]

%pip install -q unstructured langchain
%pip install -q "unstructured[all-docs]"

import pandas as pd

excel_path = "book2.xlsx"
if excel_path:
    df = pd.read_excel(excel_path)
    data = df.to_string(index=false)
else:
    print("upload an excel file")

from langchain_text_splitters import recursivecharactertextsplitter
from langchain_community.embeddings import ollamaembeddings
from langchain_community.vectorstores import chroma

text_splitter = recursivecharactertextsplitter(chunk_size=7500, chunk_overlap=100)
chunks = text_splitter.split_text(data)

embedding_model = ollamaembeddings(model="nomic-embed-text", show_progress=false)
vector_db = chroma.from_texts(
    texts=chunks, 
    embedding=embedding_model,
    collection_name="local-rag"
)

from langchain_community.chat_models import chatollama

local_model = "llama2"
llm = chatollama(model=local_model)

from langchain.prompts import prompttemplate

query_prompt = prompttemplate(
    input_variables=["question"],
    template="""you are an ai assistant. answer the user's questions based on the column names: 
    id, order_id, name, sales, refund, and status. original question: {question}"""
)

from langchain.retrievers.multi_query import multiqueryretriever

retriever = multiqueryretriever.from_llm(
    vector_db.as_retriever(), 
    llm,
    prompt=query_prompt
)

from langchain.prompts import chatprompttemplate
from langchain_core.runnables import runnablepassthrough
from langchain_core.output_parsers import stroutputparser

template = """answer the question based only on the following context:
{context}
question: {question}
"""

prompt = chatprompttemplate.from_template(template)

chain = (
    {"context": retriever, "question": runnablepassthrough()}
    | prompt
    | llm
    | stroutputparser()
)

raw_result = chain.invoke("how many rows are there?")
final_result = f"{raw_result}\n\nif you have more questions, feel free to ask!"
print(final_result)

Based on the provided context, there are 10 rows in the table.
If you have more questions, feel free to ask!