Internship

HarithaTony/internship/docs/assitant for visually impaired

@@ -0,0 +1,99 @@
+
+Virtual assistant for the visually impaired 
+Vinayak Iyer  
+Department of Information Technology 
+Sardar Patel Institute of Technology 
+Mumbai, India 
+vinayak.iyer@spit.ac.in 
+Sahil Sheth  
+Department of Information Technology 
+Sardar Patel Institute of Technology 
+Mumbai, India 
+sahil.sheth@spit.ac.in 
+ Kshitij Shah  
+Department of Information Technology 
+Sardar Patel Institute of Technology 
+Mumbai, India 
+kshitij.shah@spit.ac.in 
+Kailas Devadkar  
+Department of Information Technology 
+Sardar Patel Institute of Technology 
+Mumbai, India 
+kailas_devadkar@spit.ac.in 
+
+Abstract—Research shows that people with visual 
+impairments are 31% less likely to access the internet than 
+individuals without disabilities. This paper illustrates the 
+implementation of software that provides assistance to the 
+visually impaired for accessing the internet. The software shall 
+prove instrumental in the way the internet has accessed and will 
+increase the ease of use drastically. Although technology has 
+grown leaps and bounds, the internet - especially websites are 
+still inaccessible by the visually impaired. The software provides 
+a way to interact with these websites with much ease. With the 
+use of voice commands instead of the traditional keyboard and 
+mouse, our software provides a new dimension to access and 
+provide commands to any website. The software will read out the 
+content of the website and then using speech to text and text to 
+speech modules along with selenium, our software can automate 
+any website. The user is free from remembering complex braille 
+keyboard commands or the hassle of typing, he/she can simply 
+voice out his/her command and the software will execute it. The 
+system also has the functionality of providing a summary of the 
+content on the website and answering questions asked by the user 
+with reference to the summary using a BERT model trained on 
+the Stanford Question Answer Dataset. This software will 
+revolutionize the internet and pave the way for Web3.0.  
+Keywords — Visually impaired; Voice control; automate 
+website; blind people  
+I.  INTRODUCTION  
+Today there are nearly 285 million people in the world that 
+are visually impaired [12]. Although technology has grown 
+leaps and bounds, the accessibility, especially that of the 
+internet for differently-abled people is still far-fetched. In this 
+modern world, more and more things can be performed online. 
+From shopping, ordering food, to booking train tickets 
+everything can be done online. For almost all of these online 
+facilities a person has to use a website. Using a website can be 
+a trivial task for most people but it is very difficult for visually 
+impaired people. The internet is a highly visual form of 
+communication, different "accessibility blockers" can hinder 
+different types of websites, unlike brick and mortar businesses where accessibility can be made by including a ramp for 
+wheelchairs or braille interfaces. For example, researchers 
+found that 80% of news sites “had significant accessibility 
+issues," while 70% of respondents said they were “unable to 
+access information and services through governmen t 
+websites.” Thus, wanted to come up with a unique way of 
+allowing visually impaired people to access the internet. 
+Although the W3C has a set of recommendations that stipulate 
+the rules to be followed when designing a website for the 
+visually impaired, not all websites necessarily stick to the high 
+standards in terms of accessibility.   
+The major challenge in developing a stable software is to 
+include as few keystrokes as possible and to provide an end to 
+end experience with the help of voice alone. The inclusion of 
+multiple languages and setting the right pace of the speech 
+when played back to the user are important factors to consider. 
+To support the widespread usage of the software, a crucial 
+parameter is the dependency of the software on the local 
+environment and operating systems. While the tech has 
+evolved greatly, the accessibility, especially the internet for 
+the differently abled is still stagnant. 
+Assistive technologies such as a screen reader or 
+magnifiers can enable visually impaired individuals to access 
+the internet. Unfortunately, these screen readers need to keep 
+the functionality of the website in mind otherwise it becomes 
+difficult to read data from the website. Some of the screen 
+readers work only with a particular kind of browser and some 
+require the user to remember complex commands thus screen 
+readers are not an effective solution to the problem at hand 
+and cannot be used to access the internet. 
+ There are the following two common themes visible in 
+most websites: 
+1. Web pages are partially accessible. Some parts are 
+usable for the visually impaired, while others are not. 
+2. The accessibility of some web pages regressed due to 
+updates on the web site. Proceedings of the Fifth International Conference on Communication and Electronics Systems (ICCES 2020)
+IEEE Conference Record # 48766; IEEE Xplore ISBN: 978-1-7281-5371-1
+978-1-7281-5371-1/20/$31.00 ©2020 IEEE 1057
+Authorized licensed use limited to: Western Sydney University. Downloaded on July 26,2020 at 05:06:57 UTC from IEEE Xplore.  Restrictions apply. 

HarithaTony/internship/interface.py

@@ -0,0 +1,80 @@
+import streamlit as st
+from pypdf import *
+import os
+from datetime import datetime
+st.title("PDF Reader")
+st.subheader("Upload your files to extract data!!!")
+
+def save_uploaded_file(uploaded_file, save_path):
+    with open(f"docs\\{uploaded_file.name}", "wb") as file:
+        file.write(uploaded_file.getbuffer())
+    return f"docs\\{uploaded_file.name}"
+
+def entry_db(title,year, journal, author,types, summary):
+    import openpyxl
+
+    # Load the existing workbook
+    wb = openpyxl.load_workbook('Blank Knowledge Database.xlsx')
+
+    # Select the active sheet
+    sheet = wb.active
+
+    # Determine the last row with data in column A
+    last_row = sheet.max_row + 1
+
+    # Data to be added
+
+    # Add data to the last row
+    sheet.cell(row=last_row, column=1).value = last_row-2
+    sheet.cell(row=last_row, column=2).value = title
+    sheet.cell(row=last_row, column=3).value = year
+    sheet.cell(row=last_row, column=4).value = journal
+    sheet.cell(row=last_row, column=5).value = author
+    sheet.cell(row=last_row, column=6).value = summary
+    sheet.cell(row=last_row, column=7).value = types
+
+    # Save the workbook
+    wb.save('Blank Knowledge Database.xlsx')
+
+uploaded_file = st.file_uploader("Upload your file as .pdf",type=["pdf"])
+
+
+
+
+
+
+if uploaded_file is not None:
+    save_path = os.path.join("docs\\", uploaded_file.name)
+    saved_path = save_uploaded_file(uploaded_file, save_path)
+    st.success(f"File saved")
+    reader = PdfReader(save_path)
+    text = ""
+    meta = reader.metadata
+    number_of_pages = len(reader.pages)
+    page = reader.pages[0]
+    text = page.extract_text()
+    title = pdf_title()
+    year = pdf_date()
+    journal = pdf_journal()
+    author = pdf_author()
+    summary = pdf_summary()
+    types =""
+    for i in author:
+        auth=i[:-1]
+    timestamp = datetime.fromisoformat(str(year))
+    year= timestamp.strftime("%Y-%m-%d")
+    author=str(auth)
+    if title != "":
+        st.text_input("Title",title)
+    if year != "":
+        st.text_input("Year",year)
+    if journal!= "":
+        st.text_input("Journal",journal)
+    if author != "":
+        st.text_input("Auther",author)
+    if summary != "":
+        st.text_area("Summary",summary)
+    if st.button("submit"):
+        entry_db(title,year, journal, author,types, summary)
+
+

HarithaTony/internship/pypdf.py

@@ -0,0 +1,61 @@
+from PyPDF2 import PdfReader
+from transformers import pipeline
+import dotenv
+import torch
+
+
+dotenv.load_dotenv()
+
+reader = PdfReader("assitant for visually impaired.pdf")
+text = ""
+meta = reader.metadata
+number_of_pages = len(reader.pages)
+page = reader.pages[0]
+text = page.extract_text()
+def pdf_title():
+    return meta.title #title
+def pdf_date():
+    return meta.creation_date #date
+def pdf_journal():
+    return meta.subject
+def pdf_author():
+    try:
+        with open("sample.txt", "w", encoding='utf-8') as file:
+            file.write(text)
+            print("File created successfully.")
+    except Exception as e:
+            print("An error occurred:", str(e)) 
+
+    ner = pipeline("ner", grouped_entities=True)
+    t = ner(text)
+    auth = []
+    for i in range(len(t)):
+        if t[i]['entity_group'] == 'PER':
+            if(len(t[i]['word'])<=1):
+                auth.append(t[i]['word'].replace(",","").replace("\'","").replace(" ","").replace("  ","") + ".")  #removing unwanted characters and white-spaces
+            else:
+                auth.append(t[i]['word'] + ",")
+    auth_name = " ".join(auth)
+    return auth
+# Initialize the summarization pipeline
+def pdf_summary():
+    summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
+
+    # Load text from a file
+    file_path = "sample.txt"
+    with open(file_path, "r", encoding="utf-8") as file:
+        text = file.read()
+
+    # Split the text into smaller chunks
+    max_chunk_length = 512  # Maximum sequence length supported by the model
+    chunks = [text[i:i + max_chunk_length] for i in range(0, len(text), max_chunk_length)]
+
+    # Summarize each chunk
+    summaries = []
+    for chunk in chunks:
+        summary = summarizer(chunk, min_length=10, max_length=30)[0]['summary_text']
+        summaries.append(summary)
+
+    # Combine the summaries into a single summary
+    final_summary = ' '.join(summaries)
+    return final_summary