train.py

import openai
import pinecone
from llama_index import StorageContext, VectorStoreIndex, download_loader
from llama_index.vector_stores import PineconeVectorStore

from environments import PINECONE_API_KEY, PINECONE_INDEX, OPENAI_API_KEY

openai.api_key = OPENAI_API_KEY

print('Start Loading Data ...')

UnstructuredURLLoader = download_loader("UnstructuredURLLoader")
urls = [
    'https://tmdhc.org.hk/tc/intro',
    'https://tmdhc.org.hk/tc/service-intro',
    'https://tmdhc.org.hk/tc/service-flow',
    'https://tmdhc.org.hk/tc/contact-us',
    'https://tmdhc.org.hk/tc/service-charge',
]

loader = UnstructuredURLLoader(urls=urls, continue_on_failure=True, headers={
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
                  "AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537",
})
documents = loader.load()

pinecone.init(
    api_key=PINECONE_API_KEY,
    environment='gcp-starter'
)
pinecone_index = pinecone.Index(PINECONE_INDEX)
vector_store = PineconeVectorStore(pinecone_index=pinecone_index)
storage_context = StorageContext.from_defaults(vector_store=vector_store)
index = VectorStoreIndex.from_documents(documents, storage_context=storage_context)

print('Done!')