Python Samples

Extract Folder and File Names to Multi-Sheet Excel

import os
import pandas as pd

def extract_folder_file_names(root_dir, output_excel):
    """
    Extracts folder and file names from a directory and saves them to an Excel file,
    splitting data into multiple sheets if the total number of rows exceeds Excel's limit.

    Args:
        root_dir (str): The path to the root directory.
        output_excel (str): The path to the output Excel file (e.g., "output.xlsx").
    """
    all_data = []
    print(f"Scanning directory: {root_dir}...")
    for root, dirs, files in os.walk(root_dir):
        for dir_name in dirs:
            all_data.append({'Type': 'Folder', 'Name': dir_name, 'Path': os.path.join(root, dir_name)})
        for file_name in files:
            all_data.append({'Type': 'File', 'Name': file_name, 'Path': os.path.join(root, file_name)})
    print(f"Found {len(all_data)} items (folders/files).")

    # Define the maximum rows per Excel sheet (1,048,576 total rows, including header)
    # So, we can write 1,048,575 data rows per sheet.
    MAX_ROWS_PER_SHEET = 1048575 

    if not all_data:
        print("No folders or files found to write to Excel.")
        return

    # Create an Excel writer object
    # The 'with' statement ensures the Excel file is properly closed
    with pd.ExcelWriter(output_excel, engine='openpyxl') as writer:
        # Calculate the number of sheets needed
        num_sheets = (len(all_data) + MAX_ROWS_PER_SHEET - 1) // MAX_ROWS_PER_SHEET
        print(f"Writing data to {num_sheets} sheet(s) in {output_excel}...")

        for i in range(num_sheets):
            start_row = i * MAX_ROWS_PER_SHEET
            end_row = min((i + 1) * MAX_ROWS_PER_SHEET, len(all_data))
            
            # Get the chunk of data for the current sheet
            sheet_data = all_data[start_row:end_row]
            
            # Create a DataFrame for the current sheet's data
            df_sheet = pd.DataFrame(sheet_data)
            
            # Define the sheet name
            sheet_name = f"Sheet{i + 1}"
            
            # Write the DataFrame to the current sheet
            df_sheet.to_excel(writer, sheet_name=sheet_name, index=False)
            print(f"  - Wrote {len(sheet_data)} rows to '{sheet_name}'.")

    print(f"Successfully extracted folder and file names to '{output_excel}'.")

if __name__ == "__main__":
    # Example usage:
    # root_directory = "C:\\Users\\Kishore\\Desktop\\New"
    # output_file = "C:\\Users\\Kishore\\Desktop\\New\\output.xlsx"

    root_directory = input("Enter the path to the root directory: ")
    output_file = input("Enter the path to the output Excel file (e.g., C:\\path\\to\\file.xlsx): ")

    # Basic validation for output file extension
    if not output_file.lower().endswith(('.xlsx', '.xls')):
        print("Error: Output file must have a .xlsx or .xls extension.")
    else:
        extract_folder_file_names(root_directory, output_file)

Input Commands & Output  

PS D:\Python\Extract> python D:\Python\Extract\main.py
Enter the path to the root directory: G:\
Enter the path to the output Excel file (e.g., C:\path\to\file.xlsx): C:\Users\Kishore\Desktop\New\output.xlsx
Scanning directory: G:\...
Found 1301168 items (folders/files).
Writing data to 2 sheet(s) in C:\Users\Kishore\Desktop\New\output.xlsx...
  - Wrote 1048575 rows to 'Sheet1'.
  - Wrote 252593 rows to 'Sheet2'.
Successfully extracted folder and file names to 'C:\Users\Kishore\Desktop\New\output.xlsx'.
PS D:\Python\Extract>

Youtube Video Summary and Chapters using Bedrock LLM

# install dependencies first:
# pip install youtube-transcript-api openai pytube

import boto3
import json
from youtube_transcript_api import YouTubeTranscriptApi
from urllib.parse import urlparse, parse_qs

# Configure AWS Bedrock client
bedrock_runtime = boto3.client(
    "bedrock-runtime",
    region_name="us-east-1"  # Use the region where Llama 3.3 70B is available
)

MODEL_ID = "anthropic.claude-v2"  # Llama 3.3 70B Instruct


def extract_video_id(url):
    query = urlparse(url)
    if query.hostname == 'youtu.be':
        return query.path[1:]
    elif query.hostname in ('www.youtube.com', 'youtube.com'):
        if query.path == '/watch':
            return parse_qs(query.query)['v'][0]
        elif query.path[:7] == '/embed/':
            return query.path.split('/')[2]
        elif query.path[:3] == '/v/':
            return query.path.split('/')[2]
    return None

def get_transcript(video_id):
    transcript = YouTubeTranscriptApi.get_transcript(video_id)
    #print("full_text_transcript", transcript)
    full_text = ' '.join([entry['text'] for entry in transcript])
    #print("full_text", full_text)
    return full_text

def summarize_and_chapterize(transcript):
    system_prompt = (
        "You are an expert summarizer.\n\n"
        "TASK:\n"
        "1. Summarize the following video transcript.\n"
        "2. Break the content into well-titled chapters with timestamps (assume even spacing).\n"
        "3. Format output like this:\n\n"
        "### Summary\n"
        "<"Summary here">\n\n"
        "### Chapters\n"
        "- 00:00 - Introduction\n"
        "- 02:13 - Topic 1: ...\n"
        "- 05:47 - Topic 2: ...\n\n"
        "TRANSCRIPT:\n"
    )

    prompt = f"{system_prompt}{transcript[:7000]}"  # Claude v2 can handle more, but keep conservative

    body = {
        "prompt": f"\n\nHuman: {prompt}\n\nAssistant:",
        "max_tokens_to_sample": 2048,
        "temperature": 0.5,
        "top_k": 250,
        "top_p": 0.9,
        "stop_sequences": ["\n\nHuman:"]
    }

    response = bedrock_runtime.invoke_model(
        modelId=MODEL_ID,
        contentType="application/json",
        accept="application/json",
        body=json.dumps(body)
    )

    result = json.loads(response['body'].read())
    return result['completion']

def summarize_youtube_video(url):
    video_id = extract_video_id(url)
    if not video_id:
        return "Invalid YouTube URL"
    
    print(f"[+] Fetching transcript for video ID: {video_id}")
    transcript = get_transcript(video_id)
    print("transcript", transcript)

    print("[+] Sending to LLM for summary and chapter creation...")
    result = summarize_and_chapterize(transcript)
    print("result", result)

    return result
    #return transcript

# Example usage
if __name__ == "__main__":
    url = input("Enter YouTube video URL: ")
    output = summarize_youtube_video(url)
    print("\n--- Generated Summary and Chapters ---\n")
    print(output)