Update app.py

app.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-"""saivv_protoype
+"""saivv_prototype
 
 Automatically generated by Colab.
 
@@ -7,25 +7,10 @@ Original file is located at
     https://colab.research.google.com/drive/10LVNp23wimf3FbDIurOgtyckwjSIo3lK
 """
 
-pip install speechrecognition
-
-pip install gradio speechrecognition
-
-pip install pytesseract
-
-pip install gradio
-
+# Import necessary libraries
 import cv2  # For image processing with OpenCV
 import pytesseract  # For Optical Character Recognition (OCR) on receipts
 import gradio as gr  # For creating the Gradio interface
-
-apt update
-apt install -y tesseract-ocr
-
-pip install langchain langchain-community langchain-core transformers
-
-pip install bitsandbytes accelerate
-
 from torch import cuda, bfloat16
 import transformers
 from transformers import AutoTokenizer
@@ -33,10 +18,12 @@ import torch
 from langchain.llms import HuggingFacePipeline
 from langchain.chains import RetrievalQA
 from langchain.vectorstores import chroma
+import os
+import speech_recognition as sr  # For voice recording
 
-model_id='HuggingFaceH4/zephyr-7b-beta'
-
-device= f'cuda:{cuda.current_device()}' if cuda.is_available() else 'cpu'
+# Model configuration
+model_id = 'HuggingFaceH4/zephyr-7b-beta'
+device = f'cuda:{cuda.current_device()}' if cuda.is_available() else 'cpu'
 bnb_config = transformers.BitsAndBytesConfig(
     load_in_4bit=True,
     bnb_4bit_quant_type='nf4',
@@ -45,13 +32,13 @@ bnb_config = transformers.BitsAndBytesConfig(
 )
 print(device)
 
-import os
+# Set CUDA environment variable
 os.environ['CUDA_LAUNCH_BLOCKING'] = '1'
-model_config =  transformers.AutoConfig.from_pretrained(
+model_config = transformers.AutoConfig.from_pretrained(
     model_id,
     trust_remote_code=True,
     max_new_tokens=1024
-    )
+)
 model = transformers.AutoModelForCausalLM.from_pretrained(
     model_id,
     trust_remote_code=True,
@@ -61,6 +48,7 @@ model = transformers.AutoModelForCausalLM.from_pretrained(
 )
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 
+# Setup pipeline
 query_pipeline = transformers.pipeline(
     'text-generation',
     model=model,
@@ -71,11 +59,12 @@ query_pipeline = transformers.pipeline(
     device_map="auto"
 )
 
+# Function to format text with colors
 from IPython.display import display, Markdown
 def colorize_text(text):
-  for word, color in zip(["Reasoning", "Question", "Answer","Total time"], ["blue", "red", "green","magenta"]):
-    text = text.replace(f"{word}:", f"\n\n**<font color='{color}'>{word}:</font>**")
-  return text
+    for word, color in zip(["Reasoning", "Question", "Answer", "Total time"], ["blue", "red", "green", "magenta"]):
+        text = text.replace(f"{word}:", f"\n\n**<font color='{color}'>{word}:</font>**")
+    return text
 
 llm = HuggingFacePipeline(pipeline=query_pipeline)
 
@@ -91,7 +80,7 @@ Groceries: $496.0, Supplies: $454.42, Food: $341.69, Electronics: $351.92,
 Home Shopping: $235.68, Others: $253.45
 """
 
-question = "Based on this data, can i buy a lamborgini?"
+question = "Based on this data, can I buy a Lamborghini?"
 
 # Combine structured data into prompt
 prompt = f"{user_profile}\n\nQuestion: {question}"
@@ -103,8 +92,7 @@ response = llm(prompt=prompt)
 full_response = f"**Question:** {question}\n\n**Answer:** {response}"
 display(Markdown(colorize_text(full_response)))
 
-# Placeholder function for receipt scanning with OCR
-import speech_recognition as sr
+# Function for receipt scanning with OCR
 def scan_receipt(image):
     try:
         img_orig = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
@@ -114,7 +102,7 @@ def scan_receipt(image):
     except Exception as e:
         return f"An error occurred: {str(e)}"
 
-# Placeholder function for voice recording
+# Function for voice recording
 def record_expense(audio_path):
     recognizer = sr.Recognizer()
     with sr.AudioFile(audio_path) as source:
@@ -213,4 +201,4 @@ with gr.Blocks() as iface:
     )
 
 # Launch the app
-iface.launch()
+iface.launch()