app.py

import gradio as gr
import os
from openai import OpenAI
from retriever import (
    load_collection, load_encoder, encode_query, retrieve_docs,
    query_rerank, expand_with_neighbors, dedup_by_chapter_event
)
from sentence_transformers import CrossEncoder

QWEN_MODEL="qwen/qwen3-235b-a22b:free"
DEEPSEEK_MODEL="deepseek/deepseek-chat-v3.1:free"
GPT_OSS_MODEL="openai/gpt-oss-20b:free"
api_key = os.getenv("OPENROUTER_API_KEY")
#deepseek_key = os.getenv("DEEPSEEK_API_KEY")
client = OpenAI(base_url="https://openrouter.ai/api/v1", api_key=api_key)

# open ai method
#api_key = os.getenv("OPENAI_API_KEY")
#client = OpenAI(api_key=api_key)
#GPT_MODEL="gpt-4o"



collection = load_collection()
encoder = load_encoder()
reranker = CrossEncoder("BAAI/bge-reranker-large")

def reformulate_query(user_question, model_name=QWEN_MODEL):
    prompt = f"""你是一个BangDream知识检索助手。请把用户的问题扩写或转写为适合知识库语义检索的检索语句，涵盖所有可能的提问方式或同义关键词。
    用户问题：{user_question}
    """
    resp = client.chat.completions.create(
        model=model_name,
        messages=[{"role": "user", "content": prompt}],
        temperature=0.1,
        max_tokens=4096,
    )
    return resp.choices[0].message.content.strip()

def build_rag_prompt(query, context, system_message):
    prompt = f"""{system_message}
你将获得多个独立的资料片段，请充分查阅每一条资料.
已知资料如下：
{context}

用户提问：{query}

规则:
1. 请参考所有已知资料, 并结合资料内容，简明、准确地回答问题。
2. 如果有多个相关答案或不同观点，可以考虑是否全部分点列出
3. 如果只能在部分资料里找到答案，也请说明是参考哪些资料内容
4. 如果不能确定答案，请如实说明理由，不要凭空编造。
"""
    return prompt


def respond(
        message,
        history: list[dict[str, str]],
        system_message,
        max_tokens,
        temperature,
        top_p,
):
    """
    message: 当前输入内容
    history: [{"role": "user", "content": ...}, {"role": "assistant", "content": ...}, ...]
    system_message: 自定义 System Prompt
    """
    default_system_message = "你是BangDream知识问答助手, 也就是邦学家. 只能基于提供的资料内容作答。"
    system_msg = (system_message or default_system_message).strip()
    chat_history = [{"role": "system", "content": system_msg}]

    # reformulate query
    print("Reformulating...")
    reformulated_query_text = reformulate_query(message)
    print(f"[DEBUG] reformulated query: {reformulated_query_text}")

    print("Thinking...\n...")
    # rerank original query
    query_vec = encode_query(encoder, message)
    results = retrieve_docs(collection, query_vec, top_k=20)
    reranked = query_rerank(reranker, message, results, top_n=10)

    # rerank reformulated query
    reformulated_query_vec = encode_query(encoder, reformulated_query_text)
    reformulated_results = retrieve_docs(collection, reformulated_query_vec, top_k=20)
    reformulated_reranked = query_rerank(reranker, reformulated_query_text, reformulated_results, top_n=10)

    total_reranked = reranked + reformulated_reranked
    deduped = dedup_by_chapter_event(total_reranked, max_per_group=1)
    expanded_results = expand_with_neighbors(deduped[:5], collection)
    context = []
    for idx, text in enumerate(expanded_results):
        context.append(text[0]) if text else ""

    rag_prompt = build_rag_prompt(message, context, system_msg)
    messages = [
        {"role": "system", "content": system_msg},
        {"role": "user", "content": rag_prompt}
    ]

    response = ""
    stream = client.chat.completions.create(
        model=QWEN_MODEL,
        messages=messages,
        temperature=temperature,
        max_tokens=max_tokens,
        top_p=top_p,
        stream=True
    )
    for chunk in stream:
        delta = getattr(chunk.choices[0].delta, "content", None)
        if delta:
            response += delta
            yield response

    print("\n=== Answer ===")
    print(response)
    print("\n=== retrieved documents ===")
    for idx, (context, score, meta) in enumerate(expanded_results, 1):
        print(f"\n--- document {idx} (Score={score:.4f}) ---\n{context[:200]}...")
        print(meta)


# ========== Gradio ChatInterface with extra sidebar inputs ==========
chatbot = gr.ChatInterface(
    respond,
    type="messages",
    additional_inputs=[
        gr.Textbox(value="你是BangDream知识问答助手, 只能基于提供资料内容作答。", label="System message"),
        gr.Slider(minimum=64, maximum=8192, value=1536, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=2.0, value=0.2, step=0.05, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
    #examples=[
    #    ["在水族馆里爱音和灯发生了什么？"],
    #    ["RAS的目标是什么？"],
    #],
    description="输入你关于BangDream的问题，邦学家会基于资料库为你检索并作答\n\nGitHub项目地址: [GitHub repo](https://github.com/Cudd1es/dr-bang?tab=readme-ov-file)",
    title="Dr-Bang RAG QA Chatbot"
)

with gr.Blocks(title="Dr-Bang RAG QA") as demo:
    with gr.Sidebar():
        gr.Markdown(
            "## Dr-Bang QA\n\n"
            "[GitHub project](https://github.com/Cudd1es/dr-bang?tab=readme-ov-file)\n\n"
        )
        gr.LoginButton()
    chatbot.render()

if __name__ == "__main__":
    demo.launch()