import gradio as gr
import pandas as pd
from googleapiclient.discovery import build
from googleapiclient.errors import HttpError

# Function to fetch comments from a video
def get_filtered_video_comments(api_key, video_id, username_filter, text_filter):
    youtube = build('youtube', 'v3', developerKey=api_key)
    comments = []
    next_page_token = None

    while True:
        try:
            response = youtube.commentThreads().list(
                part='snippet,replies',
                videoId=video_id,
                maxResults=100,
                pageToken=next_page_token
            ).execute()

            for item in response['items']:
                comment = item['snippet']['topLevelComment']['snippet']
                author_name = comment.get('authorDisplayName', '')
                comment_text = comment['textDisplay']

                # Check if the comment matches the username or contains the specified text
                if (username_filter.lower() in author_name.lower() if username_filter else True) and \
                   (text_filter.lower() in comment_text.lower() if text_filter else True):
                    comments.append({
                        'VideoID': video_id,
                        'AuthorName': author_name,
                        'CommentedDateTime': comment['publishedAt'],
                        'CommentText': comment_text,
                        'NumOfCommentLikes': comment['likeCount'],
                        'CommentedUserID': comment['authorChannelId']['value']
                    })

            next_page_token = response.get('nextPageToken')
            if not next_page_token:
                break

        except HttpError as e:
            return None, f"Error fetching comments: {e}"

    return comments, None

# Main function
def scrape_filtered_comments(api_key, video_id, username_filter, text_filter):
    if not api_key or not video_id:
        return None, "API key and video ID are required."

    comments, error = get_filtered_video_comments(api_key, video_id, username_filter, text_filter)
    if error:
        return None, error

    if comments:
        df_comments = pd.DataFrame(comments)
        csv_data = df_comments.to_csv(index=False)
        return csv_data, None
    else:
        return None, "No comments matching the criteria were found."

# Gradio Interface
with gr.Blocks() as demo:
    gr.Markdown("### YouTube Filtered Comments Extractor")
    api_key = gr.Textbox(label="YouTube API Key", type="password")
    video_id = gr.Textbox(label="Video ID")
    username_filter = gr.Textbox(label="Filter by Username (Optional)")
    text_filter = gr.Textbox(label="Filter by Text in Comments (Optional)")

    output_file = gr.File(label="Download Filtered Comments as CSV")
    error_msg = gr.Textbox(label="Error Message", interactive=False)

    def main(api_key, video_id, username_filter, text_filter):
        csv_data, error = scrape_filtered_comments(api_key, video_id, username_filter, text_filter)
        if error:
            return None, error
        else:
            return csv_data, None

    extract_button = gr.Button("Extract Filtered Comments")
    extract_button.click(main, [api_key, video_id, username_filter, text_filter], [output_file, error_msg])

demo.launch()