app-llamaindex-v0.10.py

from llama_index.core import Settings
from llama_index.legacy.embeddings import AzureOpenAIEmbedding
from llama_index.legacy.llms.openai import AzureOpenAI
from llama_index.core import (
    VectorStoreIndex,
    SimpleDirectoryReader,
    StorageContext,
    load_index_from_storage,
)
from llama_index.core.node_parser import SimpleNodeParser

import tiktoken
import warnings
import os
import logging
import sys

warnings.filterwarnings("ignore")
from dotenv import load_dotenv

from colored_log_handler import ColoredConsoleHandler

logging.basicConfig(stream=sys.stdout, level=logging.INFO)
logging.getLogger().addHandler(logging.StreamHandler(stream=sys.stdout))

# Load environment variables from a .env file
load_dotenv()


azure_endpoint = os.getenv("OPENAI_API_BASE")
api_key = os.getenv("OPENAI_API_KEY")
api_version = os.getenv("OPENAI_API_VERSION")

# Configuring Embedding ada 002 model API deployed on Azure
embedding_model = AzureOpenAIEmbedding(
    model="text-embedding-ada-002",
    deployment_name="text-embedding-ada-002",
    azure_endpoint=azure_endpoint,
    api_version=api_version,
    api_key=api_key,
    embed_batch_size=10,
)
Settings.embed_model = embedding_model
Settings.chunk_size = 256
Settings.chunk_overlap = 40

documents = SimpleDirectoryReader(
    input_files=["/Users/sumved/Downloads/Demo.pdf"]
).load_data()

# Chunking the documents
parser = SimpleNodeParser.from_defaults()
nodes = parser.get_nodes_from_documents(documents)
# Indexing the documents chunks
index = VectorStoreIndex(nodes=nodes)

# Configure Azure OpenAI Service API
llm = AzureOpenAI(
    azure_deployment="gpt-35-turbo",
    azure_endpoint=azure_endpoint,
    api_version=api_version,
    api_key=api_key,
)
Settings.llm = llm

Settings.tokenizer = tiktoken.encoding_for_model("gpt-3.5-turbo").encode

# Initialize query engine
query_engine = index.as_query_engine()

print("\nReady to answer your queries. Press Ctrl+Q to exit.")

while True:
    try:
        query = input("\nEnter your query: ")

        # Check for exit command
        if query.lower() == "ctrl+q":
            print("Exiting program.")
            break

        answer = query_engine.query(query).response

        print("\nQuery:", query)
        print("\nAnswer:", answer)

    except KeyboardInterrupt:
        print("\nInterrupt received, exiting.")
        break