pdf-search-api / main.py
Ronr's picture
Upload 2 files
3a0e6a8
Raw History Blame Contribute Delete
6.45 kB
from fastapi import FastAPI, HTTPException, Body, File, Form, UploadFile
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import FileResponse
import os
import openai
from langchain.chains import RetrievalQA
from langchain.llms import OpenAI
from langchain.document_loaders import TextLoader
from langchain.document_loaders import PyPDFLoader
from langchain.document_loaders import DirectoryLoader
from langchain.text_splitter import CharacterTextSplitter
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.embeddings import OpenAIEmbeddings
from langchain.vectorstores import Chroma
from langchain.callbacks import get_openai_callback
from langchain.chat_models import AzureChatOpenAI
from langchain.llms import AzureOpenAI
# Import Azure OpenAI
from langchain.llms import AzureOpenAI
import uvicorn
import datetime
import shutil
app = FastAPI()
origins = ['*']
app.add_middleware(
CORSMiddleware,
allow_origins=origins,
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"]
)
os.environ["OPENAI_API_TYPE"] = "Azure"
os.environ["OPENAI_API_VERSION"] ="2023-05-15"
os.environ["OPENAI_API_BASE"] = "https://rgenopenaitest.openai.azure.com/"
os.environ["DEPLOYMENT_NAME"]= "rgen_pdf_analyzer"
@app.get("/")
async def read_item():
return {"message": "App Working"}
@app.post("/search")
async def upload_pdf(file: UploadFile = None, api_key: str = Form(...), text: str = Form(...)):
timestamp = datetime.datetime.now()
print(f'---------------------------------------------------{timestamp}------------------------------------------------------')
print(f'query: {text}')
try:
if file is not None:
folder_path = "./pdf_temp_storage"
# Create the folder if it doesn't exist
if not os.path.exists(folder_path):
os.makedirs(folder_path)
file_name = os.path.join('./pdf_temp_storage', file.filename)
await save_pdf(file_name, file)
openAi_api = api_key
query = text+"\n" +"Response should not be longer than 300 words"
response = qa_result(openAi_api, query, file_name, "map_reduce")
print(response)
print('-----------------------------------------------------------------------------------------------------------------------------------')
return response
else:
return chat_openai(api_key, text)
except Exception as e:
print("Error: ", e)
return e
finally:
# Cleanup operations regardless of success or failure
# shutil.rmtree(folder_path)
if file is not None:
rem_documents(file_name)
print('-----------------------------------------------------------------------------------------------------------------------------------')
def chat_openai(api_key,prompt_text):
# os.environ["OPENAI_API_KEY"] = api_key
print('Inside chat')
openai.api_key =api_key
openai.api_type = "azure"
openai.api_base = "https://rgenopenaitest.openai.azure.com/"
openai.api_version = "2023-03-15-preview"
# openai.api_key = os.getenv("OPENAI_API_KEY")
if prompt_text:
response = openai.ChatCompletion.create(
engine="rgen_pdf_analyzer",
messages = [{"role":"system","content":"You are an AI assistant that helps people find information."},
{"role": "user", "content": prompt_text }],
temperature=0.7,
max_tokens=800,
top_p=0.95,
frequency_penalty=0,
presence_penalty=0,
stop=None)
print("Answers to the prommpt :------->", response["choices"][0]["message"]["content"])
return {
"response": response["choices"][0]["message"]["content"],
}
def qa(file , query, chain_type, k):
if os.path.splitext(file)[1].lower() == ".pdf":
loader = PyPDFLoader(file)
documents = loader.load()
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
texts = text_splitter.split_documents(documents)
else:
# loader = DirectoryLoader('./pdf_temp_storage', glob="*", loader_cls=TextLoader)
# documents = loader.load()
with open(file) as f:
documents = f.read()
text_splitter = RecursiveCharacterTextSplitter(
# Set a really small chunk size, just to show.
chunk_size = 1000,
chunk_overlap = 100,
length_function = len,
)
texts = text_splitter.create_documents([documents])
azure_llm = AzureChatOpenAI(
openai_api_base=os.environ["OPENAI_API_BASE"],
openai_api_version="2023-05-15",
deployment_name=os.environ["DEPLOYMENT_NAME"],
openai_api_type="azure",
)
embeddings = OpenAIEmbeddings(model='text-embedding-ada-002',
deployment='text_embeddings',
openai_api_base='https://rgenopenaitest.openai.azure.com/',
openai_api_type='azure',
chunk_size=1)
db = Chroma.from_documents(texts, embeddings)
retriever = db.as_retriever(search_type="similarity", search_kwargs={"k": k})
qa = RetrievalQA.from_chain_type(llm=azure_llm, chain_type=chain_type, retriever=retriever, return_source_documents=True)
result = qa({"query": query})
print(result['result'])
return result
def qa_result(api_key, prompt_text, file, Chain_type):
os.environ["OPENAI_API_KEY"] = api_key
if prompt_text:
with get_openai_callback() as cb:
result = qa(file=file, query=prompt_text, chain_type=Chain_type, k=2)
return {
"response": result["result"],
"reference": result["source_documents"],
"Total Tokens": {cb.total_tokens},
"Prompt Tokens": {cb.prompt_tokens},
"Completion Tokens": {cb.completion_tokens}
}
async def save_pdf(file_name, file):
with open(file_name, "wb") as f:
f.write(await file.read())
def rem_documents(file_path):
try:
os.remove(file_path)
print("File deleted successfully:", file_path)
except FileNotFoundError:
print("File not found:", file_path)
except Exception as e:
print("An error occurred while deleting the file:", str(e))