internship

main.py

@@ -0,0 +1,185 @@
+import streamlit as st
+import psycopg2
+from PyPDF2 import PdfReader
+from datetime import datetime
+import os
+import pytesseract
+from PIL import Image
+import google.generativeai as genai
+
+genai.configure(api_key='AIzaSyD6HL9315oCQIYgYdvA1z8xU4z5ysuuPbw')
+generation_config={"temperature": 0.9,"top_p": 1,"top_k": 1,"max_output_tokens": 2048}
+# Set path to Tesseract OCR executable
+pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'
+safety_settings = [
+    {
+        "category": "HARM_CATEGORY_HARASSMENT",
+        "threshold": "BLOCK_NONE"
+    },
+    {
+        "category": "HARM_CATEGORY_HATE_SPEECH",
+        "threshold": "BLOCK_NONE"
+    },
+    {
+        "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
+        "threshold": "BLOCK_NONE"
+    },
+    {
+        "category": "HARM_CATEGORY_DANGEROUS_CONTENT",
+        "threshold": "BLOCK_NONE"
+    }
+]
+model = genai.GenerativeModel('gemini-1.0-pro-latest')
+convo = model.start_chat(history=[])
+
+# Database connection
+conn = psycopg2.connect(
+    dbname="internship",
+    user="postgres",
+    password="soloman",
+    host="localhost"
+)
+cur = conn.cursor()
+
+# Function to extract text from an image using Tesseract OCR
+def extract_text_from_image(image):
+    return pytesseract.image_to_string(image)
+
+
+# Function to extract metadata and text from PDF and insert it into the database
+def extract_metadata_and_insert(pdf_file):
+    pdf = PdfReader(pdf_file)
+    metadata = pdf.metadata
+
+    for page_number in range(len(pdf.pages)):
+        image = pdf.pages[page_number].to_image()
+        text = extract_text_from_image(image)
+
+        # Send prompts to Gemini for specific information extraction
+        try:
+            convo.send_message(
+                f'''
+                    Get the title of the text provided in {text}.
+                    Return the output as a string of the exact name as the title of the content.
+                    Do not add any other content of your own in the output.
+                '''
+            )
+            title=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the year of the text provided in {text}.
+                    Return the output as a string of the exact year as the year of the content.
+                    Do not add any other content of your own in the output.
+                '''
+            )
+            year=int(convo.last.text)
+            convo.send_message(
+                f'''
+                    Get the journal of publication of the text provided in {text}.
+                    Return the output as a string of the exact journal as the journal of the content.
+                    Do not add any other content of your own in the output.
+                '''
+            )
+            journal=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the names of the authors in {text}.
+                    Return the output as a string of all the names of authors separated by commas.
+                    Do not add any other content of your own in the output.
+                '''
+            )
+            authors=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the type of the text provided in {text}.
+                    Return the output as a string of the type of the content.
+                '''
+            )
+            genre=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the short description of the text provided in {text}.
+                    Return the output as a string of the description of the content.
+                '''
+            )
+            description=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the focal constructs of the text provided in {text}.
+                    Return the output as a string of the focal constructs of the content.
+                '''
+            )
+            constructs=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the theoretical perspectives of the text provided in {text}.
+                    Return the output as a string of the theoretical perspectives of the content.
+                '''
+            )
+            perspectives=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the context of the text provided in {text}.
+                    Return the output as a string of the context of the content.
+                '''
+            )
+            context=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the study design of the text provided in {text}.
+                    Return the output as a string of the study design of the content.
+                '''
+            )
+            study=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the levels of the text provided in {text}.
+                    Return the output as a string of the levels of the content.
+                '''
+            )
+            levels=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the findings of the text provided in {text}.
+                    Return the output as a string of the findings of the content.
+                '''
+            )
+            findings=convo.last.text
+            convo.send_message(
+                f'''
+                    Get the summary of the text provided in {text}.
+                    Return the output as a string of the summary of the content.
+                '''
+            )
+            summary=convo.last.text
+        except Exception as e:
+                i=i-1
+                continue 
+
+        # Use extracted creation date from metadata or parse creation_date_str to datetime object
+        creation_date = metadata.get('/CreationDate', '')
+
+        cur.execute("INSERT INTO pdf (author, title, creation_date, description, year, journal, genre) VALUES (%s, %s, %s, %s, %s, %s, %s)",
+                    (authors, title, creation_date, description, year, journal, genre))
+        conn.commit()
+
+# Streamlit app
+st.title("PDF Upload and View")
+# Authenticate with Gemini API
+
+# File upload
+uploaded_file = st.file_uploader("Upload a PDF file", type="pdf")
+if uploaded_file is not None:
+    st.write("File Uploaded Successfully")
+    extract_metadata_and_insert(uploaded_file)
+
+# View table contents
+st.subheader("Table Contents")
+cur.execute("SELECT * FROM pdf")
+rows = cur.fetchall()
+for row in rows:
+    st.write(row)
+
+# Close database connection
+cur.close()
+conn.close()

soloman.txt

@@ -0,0 +1,3 @@
+I wanted to learn about LLM . I need a platform to get more familiar with LLM. I can explore new applications.These provides me the opportunity to connect with the proffessionals.I can enhance my skill .
+
+In one year I want to work in a best company. I have to sharpening my skills. So I should be in a good reputed company. In third year I want to get a good position in company. The good position means team lead for a project. In fifth year I want to be a project manager in my company.