generate_description.py

from dotenv import load_dotenv
from openai_client import openai_client
load_dotenv()

client = openai_client()

def read_srt_file(file_path):
    subtitles = []
    with open(file_path, 'r') as file:
        lines = file.readlines()
        subtitle = {'index': None, 'start': None, 'end': None, 'text': ''}
        for line in lines:
            line = line.strip()
            if line.isdigit():
                if subtitle['index'] is not None:
                    subtitles.append(subtitle)
                subtitle = {'index': int(line), 'start': None, 'end': None, 'text': ''}
            elif ' --> ' in line:
                start, end = line.split(' --> ')
                subtitle['start'] = start.strip()
                subtitle['end'] = end.strip()
            elif line == '':
                continue
            else:
                subtitle['text'] += line + ' '
        if subtitle['index'] is not None:
            subtitles.append(subtitle)
    return subtitles

def split_subtitles(subtitles, chunk_size=28):
    chunks = []
    for i in range(0, len(subtitles), chunk_size):
        chunks.append(subtitles[i:i+chunk_size])
    return chunks

def summ_desc(section):
    prompt = "The format you will use is the following: 00:00 - name\n. Return only 2-3 words in the format in the place of the word 'name'. Only return 1 sentence, no more. so you just summarize the list into 1 row using that template i gave you. So the time is in minutes and seconds, the data: "
    try:
        response = client.completions.create(
            model="gpt-3.5-turbo-instruct",
            prompt=prompt + str(section),
            max_tokens=50,
            temperature=0.8,
            )
        return response.choices[0].text.strip()

    except Exception as e:
        print("An error occurred:", str(e))
    

def generate_description(title, ythandle, channel):
    
    srt_path = 'output/videos/' + channel + "/" + title + '/subtitles.srt'
    subtitles = read_srt_file(srt_path)
    
    subtitle_chunks = split_subtitles(subtitles)
    items = []
    for i, chunk in enumerate(subtitle_chunks, start=1):
        items.append(summ_desc(chunk))
    intro_template = "Subscribe " + ythandle + "\n\n-timestamps-\n"
    total = intro_template + "\n".join(items)  
    path = "output/videos/" + channel + "/" + title + "/"
    output_file = path + "description.txt"
    with open(output_file, 'w') as f:
            f.write(total)
            print("Description saved to", output_file)  
    
    
    

# generate_description("TEST_TITLE", "@TheUniverseDecoded", "the_universe_decoded")