WebPage QA

This is an example of a QA example， reference llama_index example.

It works by first obtaining a list of URLs from the user, then extracting relevant information from the pages associated with those URLs.Next, a vector index is created based on this information, and finally, the program is able to answer questions using the indexed information.

Before running this example, please set OPENAI_API_KEY environment param.

Init GPTCache

import hashlib
from gptcache import Cache
from gptcache.adapter.api import init_similar_cache


def get_hashed_name(name):
    return hashlib.sha256(name.encode()).hexdigest()


def init_gptcache(cache_obj: Cache, llm: str):
    hashed_llm = get_hashed_name(llm)
    init_similar_cache(cache_obj=cache_obj, data_dir=f"similar_cache_{hashed_llm}")

gptcache_obj = GPTCache(init_gptcache)

Load WebPage Data

from llama_index import (
    GPTVectorStoreIndex,
    ServiceContext,
    LLMPredictor,
    SimpleWebPageReader,
)

loader = SimpleWebPageReader(html_to_text=True)
documents = loader.load_data(urls=["https://milvus.io/docs/overview.md"])

Build Index and Get Query Engine

index = GPTVectorStoreIndex.from_documents(
    documents,
    service_context=ServiceContext.from_defaults(
        llm_predictor=LLMPredictor(cache=gptcache_obj)
    ),
)
query_engine = index.as_query_engine()

Query

%%time
print(query_engine.query("What is milvus?"))

Milvus is an open source vector database for building and managing large-scale AI applications. It provides fast and accurate vector search capabilities, enabling users to quickly search and retrieve vectors from large datasets.
CPU times: user 1.21 s, sys: 206 ms, total: 1.42 s
Wall time: 9.69 s

%%time
print(query_engine.query("What's milvus?"))

Milvus is an open source vector database for building and managing large-scale AI applications. It provides fast and accurate vector search capabilities, enabling users to quickly search and retrieve vectors from large datasets.
CPU times: user 784 ms, sys: 17.8 ms, total: 801 ms
Wall time: 940 ms