Spaces:
Sleeping
Sleeping
| from langchain_core.tools import tool | |
| import datetime | |
| import requests | |
| import openai | |
| import os | |
| import tempfile | |
| from urllib.parse import urlparse | |
| from openai import OpenAI | |
| from youtube_transcript_api import YouTubeTranscriptApi | |
| from pytube import extract | |
| def current_date(_) -> str : | |
| """ Returns the current date in YYYY-MM-DD format """ | |
| return datetime.datetime.now().strftime("%Y-%m-%d") | |
| def day_of_week(_) -> str : | |
| """ Returns the current day of the week (e.g., Monday, Tuesday) """ | |
| return datetime.datetime.now().strftime("%A") | |
| def days_until(date_str: str) -> str : | |
| """ Returns the number of days from today until a given date (input format: YYYY-MM-DD) """ | |
| try: | |
| future_date = datetime.datetime.strptime(date_str, "%Y-%m-%d").date() | |
| today = datetime.date.today() | |
| delta_days = (future_date - today).days | |
| return f"{delta_days} days until {date_str}" | |
| except Exception as e: | |
| return f"Error parsing date: {str(e)}" | |
| datetime_tools = [current_date, day_of_week, days_until] | |
| def transcribe_audio(audio_file: str, file_extension: str) -> str: | |
| """ Transcribes an audio file to text | |
| Args: | |
| audio_file (str): local file path to the audio file (.mp3, .m4a, etc.) | |
| file_extension (str): file extension of the audio, e.g. mp3 | |
| Returns: | |
| str: The transcribed text from the audio. | |
| """ | |
| try: | |
| response = requests.get(audio_file) # download the audio_file | |
| response.raise_for_status() # check if the http request was successful | |
| # clean file extension and save to disk | |
| file_extension = file_extension.replace('.','') | |
| filename = f'tmp.{file_extension}' | |
| with open(filename, 'wb') as file: # opens a new file for writing with a name like, e.g. tmp.mp3 | |
| file.write(response.content) # write(w) the binary(b) contents (audio file) to disk | |
| # transcribe audio with OpenAI Whisper | |
| client = OpenAI() | |
| # read(r) the audio file from disk in binary(b) mode "rb"; the "with" block ensures the file is automatically closed afterward | |
| with open(filename, "rb") as audio_content: | |
| transcription = client.audio.transcriptions.create( | |
| model="whisper-1", | |
| file=audio_content | |
| ) | |
| return transcription.text | |
| except Exception as e: | |
| return f"transcribe_audio failed: {e}" | |
| # TESTTTTT | |
| def get_youtube_transcript(page_url: str) -> str: | |
| """Get the transcript of a YouTube video | |
| Args: | |
| page_url (str): YouTube URL of the video | |
| """ | |
| try: | |
| # get video ID from URL | |
| video_id = extract.video_id(page_url) | |
| # get transcript | |
| ytt_api = YouTubeTranscriptApi() | |
| transcript = ytt_api.fetch(video_id) | |
| # keep only text | |
| txt = '\n'.join([s.text for s in transcript.snippets]) | |
| return txt | |
| except Exception as e: | |
| return f"get_youtube_transcript failed: {e}" |