app.py

import gradio as gr

from datasets import load_dataset
from pathlib import Path
from llama_index.core import SimpleDirectoryReader
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.core.node_parser import SentenceSplitter
from llama_index.core.ingestion import IngestionPipeline
from llama_index.core import VectorStoreIndex
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.llms.huggingface_api import HuggingFaceInferenceAPI
import nest_asyncio
from llama_index.core.evaluation import FaithfulnessEvaluator
import chromadb
from llama_index.vector_stores.chroma import ChromaVectorStore

def query(input):
    dataset = load_dataset(path="dvilasuero/finepersonas-v0.1-tiny", split="train")
    # dataset
    # for i, persona in enumerate(dataset):
    #   print(f"{persona["persona"]}")
    #   # persona
    #   break
    
    # 读取数据集，写入文档。
    Path("data").mkdir(parents=True, exist_ok=True)
    # enumerate(dataset)对dataset加上序号标签后进行遍历
    for i, persona in enumerate(dataset):
        with open(Path("data") / f"persona_{i}.txt", "w") as f:
            f.write(persona["persona"])
    
    
    # 加载文档，分解为node对象，输入到HuggingFaceEmbedding。
    reader = SimpleDirectoryReader(input_dir="data")
    documents = reader.load_data()
    # len(documents)
    # create the pipeline with transformations
    embed_model = HuggingFaceEmbedding(model_name="BAAI/bge-small-en-v1.5")
    pipeline = IngestionPipeline(
        transformations=[
            SentenceSplitter(),
            embed_model,
        ]
    )
    # run the pipeline sync or async
    nodes = pipeline.arun(documents=documents[:10])
    # nodes
    
    # 建立向量存储
    db = chromadb.PersistentClient(path="./wgf_chroma_db")
    chroma_collection = db.get_or_create_collection(name="wgf")
    vector_store = ChromaVectorStore(chroma_collection=chroma_collection)
    index = VectorStoreIndex.from_vector_store(
        vector_store=vector_store, embed_model=embed_model
    )
    
    # 应用补丁，允许嵌套异步
    nest_asyncio.apply()  # This is needed to run the query engine
    llm = HuggingFaceInferenceAPI(model_name="Qwen/Qwen2.5-Coder-32B-Instruct")
    query_engine = index.as_query_engine(
        llm=llm,
        response_mode="tree_summarize",
    )
    response = query_engine.query(input)
    # response
    
    # 对响应进行忠实度测评。
    # evaluator = FaithfulnessEvaluator(llm=llm)
    # eval_result = evaluator.evaluate_response(response=response)
    # eval_result.passing
    return response

demo = gr.Interface(fn=query, inputs="text", outputs="text")
demo.launch()