app.py

# -*- coding: utf-8 -*-
# 財政部財政資訊中心 江信宗

import os
from dotenv import load_dotenv
load_dotenv()
os.environ["LANGCHAIN_COMMUNITY__USER_AGENT"] = "Taiwan_Tax_Knowledge-base"
from langchain_community.utils import user_agent
from langchain_groq import ChatGroq
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_huggingface import HuggingFaceEmbeddings
from langchain_community.vectorstores import Chroma
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain.chains import RetrievalQA
from langchain_community.document_loaders import WebBaseLoader, TextLoader
from langchain.prompts import PromptTemplate
from langchain.schema import Document
import gradio as gr
import re

def initialize_llm(api_key):
    os.environ["GROQ_API_KEY"] = api_key
    return ChatGroq(
        groq_api_key=api_key,
        model_name='llama-3.1-70b-versatile'
    )
print(f"成功初始化大型語言模型（LLM）")

print(f"開始載入知識檔案", end="")
def load_documents(sources):
    documents = []
    for source in sources:
        try:
            if isinstance(source, str):
                if source.startswith('http'):
                    loader = WebBaseLoader(source)
                else:
                    loader = TextLoader(source)
                documents.extend(loader.load())
            elif isinstance(source, dict):
                documents.append(Document(page_content=source['content'], metadata=source.get('metadata', {})))
            print(".", end="")
        except Exception as e:
            print(f"Error loading source {source}: {str(e)}")
    return documents

sources = [
    "TaxQADataSet_Slim1.txt",
    "TaxQADataSet_Slim2.txt",
    "TaxQADataSet_Slim3.txt",
    "TaxQADataSet_Slim4.txt",
    "TaxQADataSet_Slim5.txt",
    "TaxQADataSet_Slim6.txt",
    "TaxQADataSet_ntpc1.txt",
    "TaxQADataSet_ntpc2.txt",
    "TaxQADataSet_kctax.txt",
    "TaxQADataSet_chutax.txt",
    "HouseTaxAct1130103.txt",
    "VehicleLicenseTaxAct1101230.txt",
    "TaxCollectionAct1101217.txt",
    "LandTaxAct1100623.txt",
    "AmusementTaxAct960523.txt",
    "StampTaxAct910515.txt",
    "DeedTaxAct990505.txt",
    "ProgressiveHouseTaxRates1130701.txt"
]

documents = load_documents(sources)
print(f"\n成功載入 {len(documents)} 個網址或檔案")

text_splitter = RecursiveCharacterTextSplitter(
    chunk_size=512,
    chunk_overlap=50,
    length_function=len,
    separators=["\n\n\n","\n\n", "\n", "。"]
)

split_docs = text_splitter.split_documents(documents)
print(f"分割後的文件數量：{len(split_docs)}")

embeddings = HuggingFaceEmbeddings(model_name="BAAI/bge-m3")
print(f"\n成功初始化嵌入模型")

print(f"開始建立向量資料庫")
vectorstore = Chroma.from_documents(split_docs, embeddings, persist_directory="./Knowledge-base")
print(f"成功建立 Chroma 向量資料庫")

retriever = vectorstore.as_retriever(search_kwargs={"k": 3})

template = """Let's work this out in a step by step way to be sure we have the right answer. Must reply to me in Taiwanese Traditional Chinese.
在回答之前，請仔細分析檢索到的上下文，確保你的回答準確完整反映了上下文中的訊息，而不是依賴先前的知識，在回應的答案中不要提到是根據上下文回答。
如果檢索到的多個上下文之間存在聯繫，請整合這些訊息以提供全面的回答，但要避免過度推斷。
如果檢索到的上下文不包含足夠回答問題的訊息，請誠實的說明，不要試圖編造答案。

上下文： {context}

問題： {question}

答案："""

PROMPT = PromptTemplate(
    template=template, input_variables=["context", "question"]
)
print(f"成功定義 Prompt Template")

def create_chain(llm):
    return RetrievalQA.from_chain_type(
        llm=llm,
        chain_type="stuff",
        retriever=retriever,
        return_source_documents=True,
        chain_type_kwargs={"prompt": PROMPT}
    )
print(f"成功建立 RAG Chain")

def generate_insight_questions(answer, api_key):
    llm = initialize_llm(api_key)
    prompt = f"""Let's work this out in a step by step way to be sure we have the right answer. Must reply to me in Taiwanese Traditional Chinese.
    根據以下回答，生成3個相關的洞察問題：

    回答： {answer}

    請提供3個簡短但有深度的問題，這些問題應該符合：
    1. 與原始回答緊密相關
    2. 能夠引導更深入的討論
    3. 涵蓋不同的面向或角度

    請直接列出這3個問題，每個問題一行，不要添加編號或其他文字。
    """
    try:
        response = llm.invoke(prompt)
        if hasattr(response, 'content'):
            questions = response.content.split('\n')
        else:
            questions = str(response).split('\n')
        while len(questions) < 3:
            questions.append("提供更多地方稅資訊")
        return questions[:3]
    except Exception as e:
        print(f"Error generating insight questions: {str(e)}")
        return ["提供更多地方稅資訊", "提供其他地方稅問題", "還想了解什麼地方稅目"]

def answer_question(query, api_key):
    try:
        llm = initialize_llm(api_key)
        chain = create_chain(llm)
        result = chain({"query": query})
        answer = result["result"]
        insight_questions = generate_insight_questions(answer, api_key)
        while len(insight_questions) < 3:
            insight_questions.append("提供更多地方稅資訊")
        return answer, insight_questions[:3]
    except Exception as e:
        return f"抱歉，處理您的問題時發生錯誤：{str(e)}", []

def split_questions(query):
    questions = re.split(r'[？！。　]', query)
    return [q.strip() for q in questions if q.strip()]

def answer_multiple_questions(query, api_key):
    questions = split_questions(query)
    all_answers = []
    all_insight_questions = []
    for question in questions:
        answer, insight_questions = answer_question(question, api_key)
        if len(questions) > 1:
            all_answers.append(f"【問題】{question}\n答案：{answer}")
        else:
            all_answers.append(answer)
        all_insight_questions.extend(insight_questions)
    if len(questions) > 1:
      combined_answer = "\n\n\n".join(all_answers)
    else:
      combined_answer = "\n".join(all_answers)
    selected_insight_questions = all_insight_questions[:3]
    return combined_answer, selected_insight_questions

def convert_punctuation(text):
    return text.replace('?', '？').replace(',', '，').replace('!', '！').replace(' ', '　')

def handle_interaction(query, api_key, state):
    if state is None:
        state = {"history": []}
    query = convert_punctuation(query)
    answer, insight_questions = answer_multiple_questions(query, api_key)
    state["history"].append((query, answer))
    while len(insight_questions) < 3:
        insight_questions.append("提供更多地方稅資訊")
    return answer, insight_questions[0], insight_questions[1], insight_questions[2], state, query

custom_css = """
body {
    background-color: #e8f5e9;
}
#answer-box textarea, #query-input textarea {
    font-size: 18px !important;
    background-color: #ffffff;
    border: 1px solid #81c784;
    border-radius: 8px;
}
.center-text {
    text-align: center !important;
    color: #2e7d32 !important;
}
.gradio-container {
    background-color: #c8e6c9 !important;
    border-radius: 15px !important;
    padding: 20px !important;
    box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1) !important;
}
.gr-button {
    color: white !important;
    border: none !important;
    border-radius: 20px !important;
    transition: all 0.3s ease !important;
    font-weight: bold !important;
}
.gr-button:hover {
    transform: translateY(-2px) !important;
    box-shadow: 0 4px 8px rgba(0, 0, 0, 0.2) !important;
}
#submit-btn {
    background-color: #ff4081 !important;
}
#submit-btn:hover {
    background-color: #f50057 !important;
}
.insight-btn {
    background-color: #00bcd4 !important;
}
.insight-btn:hover {
    background-color: #00acc1 !important;
}
.gr-form {
    background-color: #e8f5e9 !important;
    padding: 15px !important;
    border-radius: 10px !important;
}
.api-key-input {
    max-width: 512px !important;
}
"""

with gr.Blocks(theme=gr.themes.Monochrome(), css=custom_css) as iface:
    gr.Markdown("# 地方稅知識庫系統 - 財政部財政資訊中心", elem_classes=["center-text"])
    gr.Markdown("※ RAG-based 系統建置：江信宗，LLM：Llama-3.1-70B，以地方稅極少知識資料示範，僅供參考，準確資訊請依地方稅稽徵機關回覆為準。", elem_classes=["center-text"])
    with gr.Row():
        query_input = gr.Textbox(lines=2, placeholder="請輸入您的問題（支援同時輸入多個問題）......", label="輸入您的問題，系統將基於學習到的知識資料提供相關答案。", elem_id="query-input")
        api_key_input = gr.Textbox(type="password", placeholder="請輸入您的 API Key", label="API authentication key for large language models", elem_classes="api-key-input")
    answer_output = gr.Textbox(lines=6, label="答案：", elem_id="answer-box")
    with gr.Row():
        insight_q1 = gr.Button("洞察問題 1", visible=False, elem_classes=["insight-btn"])
        insight_q2 = gr.Button("洞察問題 2", visible=False, elem_classes=["insight-btn"])
        insight_q3 = gr.Button("洞察問題 3", visible=False, elem_classes=["insight-btn"])
    state = gr.State()
    current_question = gr.Textbox(lines=2, label="當前問題", visible=False)
    submit_btn = gr.Button("傳送", elem_id="submit-btn")
    def update_ui(answer, q1, q2, q3, state, current_q):
        return [
            answer,
            gr.update(value=q1, visible=bool(q1)),
            gr.update(value=q2, visible=bool(q2)),
            gr.update(value=q3, visible=bool(q3)),
            state,
            current_q
        ]
    submit_btn.click(
        fn=handle_interaction,
        inputs=[query_input, api_key_input, state],
        outputs=[answer_output, insight_q1, insight_q2, insight_q3, state, current_question]
    ).then(
        fn=update_ui,
        inputs=[answer_output, insight_q1, insight_q2, insight_q3, state, current_question],
        outputs=[answer_output, insight_q1, insight_q2, insight_q3, state, current_question]
    )
    for btn in [insight_q1, insight_q2, insight_q3]:
        btn.click(
            lambda x: x,
            inputs=[btn],
            outputs=[query_input]
        )

if __name__ == "__main__":
    iface.launch(share=True, debug=True)