Smart-Retrieval / utlis /helper.py
Nechba's picture
fisrt commit1
a4200f5
raw
history blame
7.55 kB
import pdfplumber
import streamlit as st
import requests
import json
import redis
import redis.commands.search
from redis.commands.search.field import TagField, VectorField, TextField
from redis.commands.search.indexDefinition import IndexDefinition, IndexType
import logging
from redis.commands.search.query import Query
import numpy as np
from typing import List, Dict, Any
from semantic_text_splitter import TextSplitter
from tokenizers import Tokenizer
from sentence_transformers import SentenceTransformer
from utlis.constant import *
from PIL import Image
import google.generativeai as genai
genai.configure(api_key="AIzaSyAhz9UBzkEIYI886zZRm40qqB1Kd_9Y4-0")
def initialize_session_state():
if "token" not in st.session_state:
st.session_state["token"] ="abcd"
if "service" not in st.session_state:
st.session_state["service"] = None
if "use_document" not in st.session_state:
st.session_state.use_document = False
if "flag" not in st.session_state:
st.session_state.flag = False
if "embdding_model" not in st.session_state:
st.session_state["embdding_model"] = None
if "indexing_method" not in st.session_state:
st.session_state["indexing_method"] = None
if "uploaded_files" not in st.session_state:
st.session_state["uploaded_files"] = None
if "messages" not in st.session_state:
st.session_state["messages"] = [{"role": "assistant", "content": "How can I help you?"}]
def extract_text_from_pdf(pdf_path):
text=""
with pdfplumber.open(pdf_path) as pdf:
for page_number, page in enumerate(pdf.pages, start=1):
# Try to extract the text
text+= page.extract_text(x_tolerance=2, y_tolerance=4, layout=True, x_density=5, y_density=10)
return text
def delete_service(token,service_slected_to_delete):
for srevice_name in service_slected_to_delete:
url = REMOVE_SERVICE_API
# JSON payload to be sent
data = {
"token": token,
"servicename": srevice_name
}
json_data = json.dumps(data)
# Set the headers to specify that the content type is JSON
headers = {'Content-Type': 'application/json'}
# Send the POST request
response = requests.post(url, data=json_data, headers=headers)
if json.loads( response.text).get("success")==True:
st.success(f"{srevice_name} deleted successfully")
else:
st.error(f"{srevice_name} not deleted successfully")
def delete_document(token, service,document_slected_to_delete):
for document_name in document_slected_to_delete:
url = REMOVE_DOCUMENT_API
# JSON payload to be sent
data = {
"token": token,
"servicename": service,
"documentname":document_name}
# Convert the dictionary to a JSON formatted string
json_data = json.dumps(data)
# Set the headers to specify that the content type is JSON
headers = {'Content-Type': 'application/json'}
# Send the POST request
response = requests.post(url, data=json_data, headers=headers)
if json.loads( response.text).get("status")=="success":
st.success(f"{document_name} deleted successfully")
else:
st.error(f"{document_name} not deleted successfully")
def gemini_vision(file):
load_image = Image.open(file)
prompt= "please extract all text fromt this image"
model = genai.GenerativeModel('gemini-pro-vision')
response = model.generate_content([prompt, load_image])
return response.text
def add_service(token,servicename,embdding_model):
url = ADD_SERVICES_API
# JSON payload to be sent
data = {
"token": token,
"services": [
{
"servicename": servicename,
"modelname": embdding_model
}
]
}
# Convert the dictionary to a JSON formatted string
json_data = json.dumps(data)
# Set the headers to specify that the content type is JSON
headers = {'Content-Type': 'application/json'}
# Send the POST request
response = requests.post(url, data=json_data, headers=headers)
if json.loads( response.text).get("added_services"):
st.success(f"{servicename} added successfully")
else:
st.error(response.text)
def add_document(token,servicename):
for file in st.session_state.uploaded_files:
if file.type.split('/')[-1]=='pdf':
text= extract_text_from_pdf(file)
else:
text = gemini_vision(file)
print(text)
if text:
url = CHUNK_STORE_API
# JSON payload to be sent
document_name = file.name.replace(" ","")
#document_name = document_name.replace(".pdf","")
document_name = document_name.replace("(","_")
document_name = document_name.replace(")","_")
document_name = document_name.replace("-","_")
data = {
"text": text,
"document_name":document_name,
"user_id": token,
"service_name": servicename
}
# Convert the dictionary to a JSON formatted string
json_data = json.dumps(data)
# Set the headers to specify that the content type is JSON
headers = {'Content-Type': 'application/json'}
# Send the POST request
response = requests.post(url, data=json_data, headers=headers)
document_name = file.name.replace(" ","_")
if json.loads( response.text).get("success")==True:
st.success(f"{document_name} uploaded successfully")
else:
st.error(f"{document_name} not uploaded successfully")
else:
st.error("we can't extract text from {}".format(file.name))
def get_context(prompt,token,service_name,top_k):
url = SEARCH_API
# JSON payload to be sent
data = {
"userid": token,
"service_name": service_name,
"query_str": prompt,
"document_names":st.session_state.doument_slected_to_chat ,
"top_k": top_k
}
# Convert the dictionary to a JSON formatted string
json_data = json.dumps(data)
# Set the headers to specify that the content type is JSON
headers = {'Content-Type': 'application/json'}
# Send the POST request
response = requests.post(url, data=json_data, headers=headers)
if json.loads( response.text).get("results"):
context = []
for chunk in json.loads( response.text).get("results"):
context.append(chunk['chunk'])
return context
else:
return []
def query(payload):
response = requests.post(API_URL, headers=HEADERS, json=payload)
return response.json()
def generate_response(llm_name, question, context = None):
url = CHAT_API
#st.chat_message("assistant", avatar="πŸ€–").write(context)
# JSON payload to be sent
data = {
"context": context,
"question": question,
"model_name": llm_name,
}
# Convert the dictionary to a JSON formatted string
json_data = json.dumps(data)
# Set the headers to specify that the content type is JSON
headers = {'Content-Type': 'application/json'}
# Send the POST request
response = requests.post(url, data=json_data, headers=headers)
return json.loads( response.text).get("response", "429 Quota exceeded for quota metric.")