from langchain_community.document_loaders import AssemblyAIAudioTranscriptLoader
from langchain_community.document_loaders.assemblyai import TranscriptFormat
import assemblyai as aai
from dotenv import load_dotenv, find_dotenv
import os


class Audio:
    audio_file_path = ""
    api_key = ""

    def __init__(self):
        load_dotenv(find_dotenv())
        self.api_key = os.getenv("AI_AUDIO_TRANSCRIPT_KEY")

    def transcribe(self, audio_file_path):
        # Validar que el archivo existe
        self.audio_file_path = audio_file_path

        config = aai.TranscriptionConfig(
            speaker_labels=True,
            auto_chapters=False,
            entity_detection=True,
            language_code=aai.LanguageCode.es,
        )

        loader = AssemblyAIAudioTranscriptLoader(
            file_path=self.audio_file_path,
            api_key=self.api_key,
            config=config,
            transcript_format=TranscriptFormat.TEXT,
        )
        docs = loader.load()

        return docs[0].page_content