app.py

import gradio as gr

from huggingface_hub import snapshot_download
from langchain.document_loaders import (
    CSVLoader,
    EverNoteLoader,
    PDFMinerLoader,
    TextLoader,
    UnstructuredEmailLoader,
    UnstructuredEPubLoader,
    UnstructuredHTMLLoader,
    UnstructuredMarkdownLoader,
    UnstructuredODTLoader,
    UnstructuredPowerPointLoader,
    UnstructuredWordDocumentLoader,
)
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.vectorstores import Chroma
from langchain.embeddings import HuggingFaceEmbeddings
from langchain.docstore.document import Document
from chromadb.config import Settings
from llama_cpp import Llama


SYSTEM_PROMPT = "Ты — Сайга, русскоязычный автоматический ассистент. Ты разговариваешь с людьми и помогаешь им."
SYSTEM_TOKEN = 1788
USER_TOKEN = 1404
BOT_TOKEN = 9225
LINEBREAK_TOKEN = 13

ROLE_TOKENS = {
    "user": USER_TOKEN,
    "bot": BOT_TOKEN,
    "system": SYSTEM_TOKEN
}

LOADER_MAPPING = {
    ".csv": (CSVLoader, {}),
    ".doc": (UnstructuredWordDocumentLoader, {}),
    ".docx": (UnstructuredWordDocumentLoader, {}),
    ".enex": (EverNoteLoader, {}),
    ".epub": (UnstructuredEPubLoader, {}),
    ".html": (UnstructuredHTMLLoader, {}),
    ".md": (UnstructuredMarkdownLoader, {}),
    ".odt": (UnstructuredODTLoader, {}),
    ".pdf": (PDFMinerLoader, {}),
    ".ppt": (UnstructuredPowerPointLoader, {}),
    ".pptx": (UnstructuredPowerPointLoader, {}),
    ".txt": (TextLoader, {"encoding": "utf8"}),
}


MODEL_NAME = "ggml-model-q4_1.bin"
snapshot_download(
    repo_id="IlyaGusev/saiga_7b_lora_llamacpp",
    local_dir=".",
    allow_patterns=MODEL_NAME
)


model = Llama(
    model_path=MODEL_NAME,
    n_ctx=2000,
    n_parts=1,
)

max_new_tokens = 1500
top_k = 30
top_p = 0.9
temp = 0.1
repeat_penalty = 1.15
chunk_size = 300

embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/paraphrase-multilingual-mpnet-base-v2")


def load_single_document(file_path: str) -> Document:
    ext = "." + file_path.rsplit(".", 1)[-1]
    assert ext in LOADER_MAPPING
    loader_class, loader_args = LOADER_MAPPING[ext]
    loader = loader_class(file_path, **loader_args)
    return loader.load()[0]


def get_message_tokens(model, role, content):
    message_tokens = model.tokenize(content.encode("utf-8"))
    message_tokens.insert(1, ROLE_TOKENS[role])
    message_tokens.insert(2, LINEBREAK_TOKEN)
    message_tokens.append(model.token_eos())
    return message_tokens


def get_system_tokens(model):
    system_message = {"role": "system", "content": SYSTEM_PROMPT}
    return get_message_tokens(model, **system_message)


def upload_files(files, file_paths):
    file_paths = [f.name for f in files]
    return file_paths


def build_index(file_paths, db):
    documents = [load_single_document(path) for path in file_paths]
    text_splitter = RecursiveCharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=20)
    texts = text_splitter.split_documents(documents)
    def fix_lines(text):
        lines = text.split("\n")
        lines = [line for line in lines if len(line.strip()) > 2]
        return "\n".join(lines)
    fixed_texts = []
    for text in texts:
        text.page_content = fix_lines(text.page_content)
        if len(text.page_content) < 10:
            continue
        fixed_texts.append(text)

    db = Chroma.from_documents(
        fixed_texts,
        embeddings,
        client_settings=Settings(
            anonymized_telemetry=False
        )
    )
    return db


def user(message, history, system_prompt):
    new_history = history + [[message, None]]
    return "", new_history


def bot(history, system_prompt, conversation_id, db):
    if not history:
        return

    tokens = get_system_tokens(model)[:]
    tokens.append(LINEBREAK_TOKEN)

    for user_message, bot_message in history[:-1]:
        message_tokens = get_message_tokens(model=model, role="user", content=user_message)
        tokens.extend(message_tokens)
        if bot_message:
            message_tokens = get_message_tokens(model=model, role="bot", content=bot_message)
            tokens.extend(message_tokens)

    last_user_message = history[-1][0]
    if db:
        retriever = db.as_retriever(search_kwargs={"k": 2})
        docs = retriever.get_relevant_documents(last_user_message)
        context = "\n\n".join([doc.page_content for doc in docs])
        last_user_message = f"Контекст: {context}\n\nИспользуя контекст, ответь на вопрос: {last_user_message}"
    message_tokens = get_message_tokens(model=model, role="user", content=last_user_message)
    tokens.extend(message_tokens)

    role_tokens = [model.token_bos(), BOT_TOKEN, LINEBREAK_TOKEN]
    tokens.extend(role_tokens)
    generator = model.generate(
        tokens,
        top_k=top_k,
        top_p=top_p,
        temp=temp,
        repeat_penalty=repeat_penalty
    )

    completion_tokens = []
    partial_text = ""
    for i, token in enumerate(generator):
        completion_tokens.append(token)
        if token == model.token_eos() or (max_new_tokens is not None and i >= max_new_tokens):
            break
        partial_text = model.detokenize(completion_tokens).decode("utf-8", "ignore")
        history[-1][1] = partial_text
        yield history


with gr.Blocks(
    theme=gr.themes.Soft()
) as demo:
    db = gr.State(None)
    conversation_id = gr.State(get_uuid)
    favicon = '<img src="https://cdn.midjourney.com/b88e5beb-6324-4820-8504-a1a37a9ba36d/0_1.png" width="48px" style="display: inline">'
    gr.Markdown(
        f"""<h1><center>{favicon}Saiga 7B Retrieval QA Llama.cpp</center></h1>
        """
    )

    system_prompt = gr.Textbox(label="Системный промпт", placeholder="", value=SYSTEM_PROMPT)

    file_output = gr.File(file_count="multiple")
    file_paths = gr.State([])

    chatbot = gr.Chatbot().style(height=400)
    with gr.Row():
        with gr.Column():
            msg = gr.Textbox(
                label="Отправить сообщение",
                placeholder="Отправить сообщение",
                show_label=False,
            ).style(container=False)
        with gr.Column():
            with gr.Row():
                submit = gr.Button("Отправить")
                stop = gr.Button("Остановить")
                clear = gr.Button("Очистить")

    upload_event = file_output.change(
        fn=upload_files,
        inputs=[file_output, file_paths],
        outputs=[file_paths],
        queue=False,
    ).then(
        fn=build_index,
        inputs=[file_paths, db],
        outputs=[db],
        queue=True
    )

    submit_event = msg.submit(
        fn=user,
        inputs=[msg, chatbot, system_prompt],
        outputs=[msg, chatbot],
        queue=False,
    ).then(
        fn=bot,
        inputs=[chatbot, system_prompt, conversation_id, db],
        outputs=chatbot,
        queue=True,
    )

    submit_click_event = submit.click(
        fn=user,
        inputs=[msg, chatbot, system_prompt],
        outputs=[msg, chatbot],
        queue=False,
    ).then(
        fn=bot,
        inputs=[chatbot, system_prompt, conversation_id, db],
        outputs=chatbot,
        queue=True,
    )
    stop.click(
        fn=None,
        inputs=None,
        outputs=None,
        cancels=[submit_event, submit_click_event],
        queue=False,
    )
    clear.click(lambda: None, None, chatbot, queue=False)

demo.queue(max_size=128, concurrency_count=1)
demo.launch()