Integrate / app.py
shahed125238's picture
Update app.py
2852f8f verified
Raw History Blame Contribute Delete
9.24 kB
import os
import streamlit as st
import pandas as pd
import pickle
from PyPDF2 import PdfReader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_google_genai import GoogleGenerativeAIEmbeddings
import google.generativeai as genai
from langchain.vectorstores import FAISS
from langchain_google_genai import ChatGoogleGenerativeAI
from langchain.chains.question_answering import load_qa_chain
from langchain.prompts import PromptTemplate
from dotenv import load_dotenv
# Load environment variables
load_dotenv()
# Configure Google API
genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
# Function to extract text from PDFs
def get_pdf_text(pdf_docs):
text = ""
for pdf in pdf_docs:
pdf_reader = PdfReader(pdf)
for page in pdf_reader.pages:
text += page.extract_text()
return text
# Function to split text into chunks
def get_text_chunks(text):
text_splitter = RecursiveCharacterTextSplitter(chunk_size=10000, chunk_overlap=1000)
chunks = text_splitter.split_text(text)
return chunks
# Function to store text chunks in FAISS vector store
def get_vector_store(text_chunks):
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
vector_store = FAISS.from_texts(text_chunks, embedding=embeddings)
vector_store.save_local("faiss_index")
# Function to load the conversational chain for Q&A
def get_conversational_chain():
prompt_template = """
Answer the question as detailed as possible from the provided context. Provide all details in English.
If the answer is not in the provided context, just say, 'The answer is not available in the context.'
Do not provide incorrect information.
Context:\n{context}\n
Question:\n{question}\n
Answer:
"""
model = ChatGoogleGenerativeAI(model="gemini-pro", temperature=0.3)
prompt = PromptTemplate(template=prompt_template, input_variables=["context", "question"])
chain = load_qa_chain(model, chain_type="stuff", prompt=prompt)
return chain
# Function to handle user input and provide answers
def user_input(user_question):
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
# Load the vector store with dangerous deserialization allowed
new_db = FAISS.load_local("faiss_index", embeddings, allow_dangerous_deserialization=True)
# Perform similarity search
docs = new_db.similarity_search(user_question)
# Load the conversational chain
chain = get_conversational_chain()
# Get the response based on the search
response = chain({"input_documents": docs, "question": user_question}, return_only_outputs=True)
# Print and display the response
print(response)
return response["output_text"]
# Load the saved model and feature columns for the machine health predictor
with open('logistic_regression_model.pkl', 'rb') as model_file:
model = pickle.load(model_file)
with open('feature_columns.pkl', 'rb') as col_file:
feature_columns = pickle.load(col_file)
def predict_status(air_temp, proc_temp, speed, torque, tool_wear):
# Prepare the input data
sample_input = {
"Air temperature [K]": air_temp,
"Process temperature [K]": proc_temp,
"Rotational speed [rpm]": speed,
"Torque [Nm]": torque,
"Tool wear [min]": tool_wear
}
# Convert the sample input to a DataFrame
sample_df = pd.DataFrame([sample_input])
# Ensure all columns from training data are present in the sample DataFrame
for col in feature_columns:
if col not in sample_df.columns:
sample_df[col] = 0 # Add missing columns and fill with 0
# Reorder columns to match the order in the training data
sample_df = sample_df[feature_columns]
# Make a prediction
prediction = model.predict(sample_df)[0]
# Initialize the reason list
reasons = []
# Define acceptable ranges for each feature with user-friendly feedback
acceptable_ranges = {
"Air temperature [K]": (150, 500),
"Process temperature [K]": (150, 500),
"Rotational speed [rpm]": (500, 1500),
"Torque [Nm]": (100, 800),
"Tool wear [min]": (50, 300)
}
# User-friendly reasons for each feature
if air_temp < 150:
reasons.append("• The air temperature is too low, which might be causing the machine to underperform.")
elif air_temp > 500:
reasons.append("• The air temperature is too high, which could be causing the machine to overheat.")
if proc_temp < 150:
reasons.append("• The process temperature is below the normal range, possibly affecting machine efficiency.")
elif proc_temp > 500:
reasons.append("• The process temperature is too high, which might lead to overheating.")
if speed < 500:
reasons.append("• The rotational speed is too low, causing the machine to operate inefficiently.")
elif speed > 1500:
reasons.append("• The rotational speed is too high, which might be causing excessive strain on the machine.")
if torque < 100:
reasons.append("• The torque is too low, which may indicate that the machine is not operating at its full capacity.")
elif torque > 800:
reasons.append("• The torque is too high, possibly putting excessive stress on the machine's components.")
if tool_wear < 50:
reasons.append("• The tool wear is minimal, but routine checks are still recommended to prevent future issues.")
elif tool_wear > 300:
reasons.append("• The tool wear is high, and the tool may need to be replaced soon to maintain optimal performance.")
# Generate the final result message
if prediction == 0:
result_message = "🔴 Machine is NOT working.\n" # First line for machine status
if reasons:
result_message += "\n".join(reasons) # Second row onward for reasons with bullet points
else:
result_message += "No specific reason identified." # In case no issues found
else:
result_message = "🟢 Machine is working fine."
return result_message
# Main function for the Streamlit app
def main():
st.set_page_config("Multi-PDF Chatbot & Machine Health Predictor", page_icon=":scroll:")
# Sidebar for navigation
st.sidebar.title("Navigation")
app_mode = st.sidebar.radio("Choose an option", ["Combined Screen"])
# Define the PDF directory path
pdf_directory = r"data"
# Get all PDF files from the directory
pdf_docs = [os.path.join(pdf_directory, file) for file in os.listdir(pdf_directory) if file.endswith('.pdf')]
# Initialize session state for prediction message and user question
if "prediction_message" not in st.session_state:
st.session_state.prediction_message = ""
if "user_question" not in st.session_state:
st.session_state.user_question = ""
# Slider defaults for machine health predictor
if "air_temp" not in st.session_state:
st.session_state.air_temp = 300
if "proc_temp" not in st.session_state:
st.session_state.proc_temp = 300
if "speed" not in st.session_state:
st.session_state.speed = 1000
if "torque" not in st.session_state:
st.session_state.torque = 500
if "tool_wear" not in st.session_state:
st.session_state.tool_wear = 100
# Combined Screen
if app_mode == "Combined Screen":
st.header("Smart Machine Health Predictor & PDF Chatbot")
# Machine Health Predictor
st.subheader("Machine Health Predictor")
air_temp = st.slider("Air Temperature (K)", 0, 1000, st.session_state.air_temp, key='air_temp')
proc_temp = st.slider("Process Temperature (K)", 0, 1000, st.session_state.proc_temp, key='proc_temp')
speed = st.slider("Rotational Speed (rpm)", 0, 2000, st.session_state.speed, key='speed')
torque = st.slider("Torque (Nm)", 0, 1000, st.session_state.torque, key='torque')
tool_wear = st.slider("Tool Wear (min)", 0, 500, st.session_state.tool_wear, key='tool_wear')
if st.button("Predict Machine Status"):
result_message = predict_status(air_temp, proc_temp, speed, torque, tool_wear)
st.session_state.prediction_message = result_message
# Automatically populate the PDF Reader Chatbot input with the result message
st.session_state.user_question = result_message # Set the question to the result
st.markdown(st.session_state.prediction_message)
# PDF Reader Chatbot
st.subheader("Ask a question about the PDF documents:")
user_question = st.text_input("Question:", value=st.session_state.user_question) # Set the input box to the prediction message
if user_question:
pdf_text = get_pdf_text(pdf_docs)
text_chunks = get_text_chunks(pdf_text)
get_vector_store(text_chunks)
answer = user_input(user_question)
st.success("Answer retrieved successfully!")
st.write(answer)
else:
st.warning("Please enter a question.")
# Run the app
if __name__ == "__main__":
main()