temp.py

import csv
import re

# Define the names of your input and output files
INPUT_FILENAME = 'emails.txt'
OUTPUT_FILENAME = 'processed_emails.csv'

def process_email_line(email):
    """
    Parses a single email address to extract and capitalize the full name.
    """
    # 1. Extract the name part (before the first '@')
    name_part = email.split('@')[0]

    # 2. Remove the year (4 or more digits) which is typically at the end
    name_part_clean = re.sub(r'\d{4,}', '', name_part)

    # 3. Split the cleaned name part by the dot '.'
    name_components = name_part_clean.split('.')

    first_name = name_components[0]
    # Check if a last name component exists
    last_name = name_components[1] if len(name_components) > 1 else ''

    # 4. Capitalize each name part
    capitalized_first_name = first_name.title()
    capitalized_last_name = last_name.title()

    # 5. Combine the names
    full_name = f"{capitalized_first_name} {capitalized_last_name}".strip()

    return full_name, email

def process_emails_file(input_path, output_path):
    """
    Reads emails from an input file, deduplicates and sorts them,
    processes them, and writes the results to a new CSV file.
    """
    unique_emails = set()

    try:
        # First pass: Read all emails and store them in a set for deduplication
        with open(input_path, 'r', encoding='utf-8') as infile:
            for line in infile:
                email = line.strip()
                if email: # Only consider non-empty lines
                    unique_emails.add(email)

    except FileNotFoundError:
        print(f"❌ Error: The input file '{input_path}' was not found. Please make sure it exists.")
        return
    except Exception as e:
        print(f"❌ An unexpected error occurred during reading: {e}")
        return

    # Convert the set to a list and sort it alphabetically
    sorted_emails = sorted(list(unique_emails))

    try:
        # Second pass: Process the sorted, unique emails and write to CSV
        with open(output_path, 'w', newline='', encoding='utf-8') as outfile:
            csv_writer = csv.writer(outfile)
            # Write the header row
            csv_writer.writerow(['Name', 'Email'])

            for email in sorted_emails:
                # Process the email and get the name
                full_name, original_email = process_email_line(email)

                # Write the processed data to the CSV file
                csv_writer.writerow([full_name, original_email])

        print(f"✅ Success! {len(sorted_emails)} unique emails were processed and saved to '{output_path}'.")

    except Exception as e:
        print(f"❌ An unexpected error occurred during writing: {e}")


# Run the main function
process_emails_file(INPUT_FILENAME, OUTPUT_FILENAME)