Tutorial Pythin Script to Extract Email Data from Linkedin 2023 Database

Code:
import re

from tkinter import Tk, filedialog



def extract_info(row):

    # Define the regex pattern for email

    email_pattern = r'\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Z|a-z]{2,}\b'



    # Define the regex pattern for US or United States

    country_pattern = r'\b(?:US|United States)\b'



    # Find emails in the row

    emails = re.findall(email_pattern, row)



    # Check if any email matches the country pattern

    for email in emails:

        if re.search(country_pattern, row):

            # Extract name (assuming it's before the email) and country from the row

            name_match = re.search(r'([A-Za-z]+ [A-Za-z]+)', row)

            name = name_match.group(1) if name_match else None

            country = re.search(country_pattern, row).group() if re.search(country_pattern, row) else None

            return email, name, country

    return None, None, None



def process_file(input_file_path, output_file_path):

    try:

        with open(input_file_path, 'r') as input_file, open(output_file_path, 'w') as output_file:

            for line in input_file:

                email, name, country = extract_info(line)

                if email:

                    output_file.write(f"Email: {email}, Name: {name}, Country: {country}\n")

        print(f"Extraction completed. Results saved to {output_file_path}")

    except Exception as e:

        print(f"Error processing file: {e}")



def browse_files():

    Tk().withdraw()  # We don't want a full GUI, so keep the root window from appearing

    input_file_path = filedialog.askopenfilename(title="Select an input file")  # Show an "Open" dialog box and return the path to the selected file

    if input_file_path:

        output_file_path = filedialog.asksaveasfilename(title="Select an output file", defaultextension=".txt", filetypes=[("Text files", "*.txt")])

        if output_file_path:

            process_file(input_file_path, output_file_path)



# Create a simple GUI

if __name__ == "__main__":

    browse_files()