salt/summarizeyt
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Wrap in a main function and use args

Browse Source
This commit is contained in:
Salt
2025-01-08 13:46:17 -06:00
parent de310ca85e
commit fca8c6e232
1 changed files with 46 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

33
summarize.py
View File

@@ -1,13 +1,11 @@
#! /usr/bin/env python3 #! /usr/bin/env python3
import argparse
import io import io
import json import json
import yt_dlp import yt_dlp
from ollama import chat, ChatResponse, Client from ollama import chat, ChatResponse, Client
#video_url = 'https://youtu.be/jl4HOY8ZaEA'
video_url = 'https://www.youtube.com/watch?v=kTctVqjhDEw'
ydl_opts = { ydl_opts = {
'writesubtitles': True, # Enable downloading subtitles 'writesubtitles': True, # Enable downloading subtitles
'subtitleslangs': ['en'], # Specify subtitle language(s) 'subtitleslangs': ['en'], # Specify subtitle language(s)
@@ -69,9 +67,25 @@ def concatenate_subtitles(subtitle_json):
# Join all collected text with spaces and return # Join all collected text with spaces and return
return ' '.join(result) return ' '.join(result)
with yt_dlp.YoutubeDL(ydl_opts) as ydl: def main():
parser = argparse.ArgumentParser(
description="Download subtitles from a video and summarize it using a local Ollama instance."
)
parser.add_argument(
'url',
metavar='URL',
type=str,
help="The URL of the video to process."
)
# Parse out arguments
args = parser.parse_args()
video_url = args.url
with yt_dlp.YoutubeDL(ydl_opts) as ydl:
# Extract metadata without downloading the video # Extract metadata without downloading the video
info = ydl.extract_info(video_url, download=False) info = ydl.extract_info(video_url, download=False)
print(f"Summarizing video: {info.get('title', 'Unknown Title')}...")
# Check if subtitles are available # Check if subtitles are available
subtitle_lang = 'en' # Change this to your desired language subtitle_lang = 'en' # Change this to your desired language
@@ -79,23 +93,26 @@ with yt_dlp.YoutubeDL(ydl_opts) as ydl:
automatic_subtitles_available = info.get('automatic_captions', {}) automatic_subtitles_available = info.get('automatic_captions', {})
if subtitle_lang in subtitles_available: if subtitle_lang in subtitles_available:
print(f"Downloading manual subtitles for language: {subtitle_lang}") print(f"Downloading manual subtitles for language: {subtitle_lang}...")
subtitle_url = subtitles_available[subtitle_lang][0]['url'] subtitle_url = subtitles_available[subtitle_lang][0]['url']
elif subtitle_lang in automatic_subtitles_available: elif subtitle_lang in automatic_subtitles_available:
print(f"No manual subtitles available. Falling back to auto-generated subtitles for language: {subtitle_lang}") print(f"No manual subtitles available. Falling back to auto-generated subtitles for language: {subtitle_lang}...")
subtitle_url = automatic_subtitles_available[subtitle_lang][0]['url'] subtitle_url = automatic_subtitles_available[subtitle_lang][0]['url']
else: else:
print(f"No subtitles (manual or auto-generated) available for language: {subtitle_lang}") print(f"No subtitles (manual or auto-generated) available for language: {subtitle_lang}!")
subtitle_url = None subtitle_url = None
exit(51)
# If a subtitle URL was found, download the subtitles # If a subtitle URL was found, download the subtitles
if subtitle_url: if subtitle_url:
subtitle_data = ydl.urlopen(subtitle_url).read().decode('utf-8') subtitle_data = ydl.urlopen(subtitle_url).read().decode('utf-8')
else: else:
print("Failed to download subtitles.") print("Failed to download subtitles!")
exit(50) exit(50)
subs = concatenate_subtitles(json.loads(subtitle_data)) subs = concatenate_subtitles(json.loads(subtitle_data))
print("Getting summary...") print("Getting summary...")
summary = get_summary(subs) summary = get_summary(subs)
print(summary) print(summary)
main()