mirror of
https://github.com/marcogll/talia_bot.git
synced 2026-01-13 21:35:19 +00:00
81efa4babd
This commit addresses critical issues in the Talia Bot system, including fixing missing admin flows, correcting agenda privacy logic, implementing voice message transcription, and enforcing RAG guardrails. Key changes include: - Modified the onboarding module to dynamically generate admin menus from available JSON flows, making all admin functions accessible. - Updated the agenda module to correctly use separate work and personal Google Calendar IDs, ensuring privacy and accurate availability. - Implemented audio transcription using the OpenAI Whisper API, replacing placeholder logic and enabling multimodal interaction. - Reworked the sales RAG module to prevent it from generating generic responses when it lacks sufficient context. Additionally, this commit introduces comprehensive documentation as requested: - `AGENTS.md`: Defines the roles and responsibilities of each system agent. - `Agent_skills.md`: Details the technical capabilities and business rules for each agent. - `plan_de_pruebas.md`: Provides a step-by-step test plan to verify the fixes. - `reparacion_vs_refactor.md`: Outlines the immediate repairs performed and proposes a strategic, incremental plan for long-term architectural improvements.
57 lines
1.9 KiB
Python
57 lines
1.9 KiB
Python
# talia_bot/modules/llm_engine.py
|
|
# Este script se encarga de la comunicación con la inteligencia artificial de OpenAI.
|
|
|
|
import openai
|
|
from talia_bot.config import OPENAI_API_KEY, OPENAI_MODEL
|
|
|
|
def get_smart_response(prompt):
|
|
"""
|
|
Genera una respuesta inteligente usando la API de OpenAI.
|
|
|
|
Parámetros:
|
|
- prompt: El texto o pregunta que le enviamos a la IA.
|
|
"""
|
|
# Verificamos que tengamos la llave de la API configurada
|
|
if not OPENAI_API_KEY:
|
|
return "Error: La llave de la API de OpenAI no está configurada."
|
|
|
|
try:
|
|
# Creamos el cliente de OpenAI
|
|
client = openai.OpenAI(api_key=OPENAI_API_KEY)
|
|
|
|
# Solicitamos una respuesta al modelo configurado
|
|
response = client.chat.completions.create(
|
|
model=OPENAI_MODEL,
|
|
messages=[
|
|
{"role": "system", "content": "Eres un asistente útil."},
|
|
{"role": "user", "content": prompt},
|
|
],
|
|
)
|
|
# Devolvemos el contenido de la respuesta limpia (sin espacios extras)
|
|
return response.choices[0].message.content.strip()
|
|
except Exception as e:
|
|
# Si algo sale mal, devolvemos el error
|
|
return f"Ocurrió un error al comunicarse con OpenAI: {e}"
|
|
|
|
def transcribe_audio(audio_file_path):
|
|
"""
|
|
Transcribes an audio file using OpenAI's Whisper model.
|
|
|
|
Parameters:
|
|
- audio_file_path: The path to the audio file.
|
|
"""
|
|
if not OPENAI_API_KEY:
|
|
return "Error: OPENAI_API_KEY is not configured."
|
|
|
|
try:
|
|
client = openai.OpenAI(api_key=OPENAI_API_KEY)
|
|
|
|
with open(audio_file_path, "rb") as audio_file:
|
|
transcript = client.audio.transcriptions.create(
|
|
model="whisper-1",
|
|
file=audio_file
|
|
)
|
|
return transcript.text
|
|
except Exception as e:
|
|
return f"Error during audio transcription: {e}"
|