Se agrega el proyecto al repositorio

api.py

@@ -0,0 +1,106 @@
+from flask import Flask, request, jsonify
+from langchain_community.tools.tavily_search import TavilySearchResults
+from typing import Annotated
+from typing_extensions import TypedDict
+from langgraph.graph.message import add_messages
+from langchain_openai import ChatOpenAI
+from dotenv import load_dotenv
+from langgraph.prebuilt import create_react_agent
+from langchain_core.prompts import ChatPromptTemplate
+from langgraph.checkpoint.memory import MemorySaver
+from langchain_tools.agent_tools import (
+    redact_email, list_calendar_events,
+    create_calendar_event, get_company_info,
+    get_current_date_and_time
+)
+from langchain_community.tools.gmail.utils import (
+    build_resource_service, get_gmail_credentials
+)
+from langchain_community.agent_toolkits import GmailToolkit
+
+# Cargar las variables de entorno
+load_dotenv()
+
+# Inicializar la app Flask
+app = Flask(__name__)
+
+# Inicializar el modelo LLM de OpenAI
+llm = ChatOpenAI(
+    model="gpt-4o-mini",
+    temperature=0
+)
+
+# Configuración de Gmail
+toolkit = GmailToolkit()
+credentials = get_gmail_credentials(
+    token_file="token.json",
+    scopes=["https://mail.google.com/"],
+    client_secrets_file="credentials.json",
+)
+api_resource = build_resource_service(credentials=credentials)
+toolkit = GmailToolkit(api_resource=api_resource)
+
+# Crear herramientas
+tools = toolkit.get_tools()
+search = TavilySearchResults(max_results=2)
+tools.extend([search, redact_email, list_calendar_events,
+              create_calendar_event, get_company_info, get_current_date_and_time])
+
+# Definir el sistema prompt
+system_prompt = ChatPromptTemplate.from_messages(
+    [
+        ("system", "Eres Mariana, el asistente virtual de OneCluster, una empresa de software que ofrece soluciones personalizadas. Asume el tono de J.A.R.V.I.S.: cordial, atento y con tacto en todo momento."),
+        ("system", "Preséntate como Mariana en el primer mensaje y pregunta el nombre del usuario si no lo tienes registrado."),
+        ("system", "Si el usuario ya ha interactuado antes, usa su nombre sin necesidad de volver a preguntar."),
+        ("system", "OneCluster es una empresa de software especializada en desarrollo a medida. Solo responde a preguntas y solicitudes relacionadas con la empresa y sus servicios."),
+        ("system", "Si necesitas información adicional sobre la empresa, usa la función get_company_info."),
+        ("system", "Antes de enviar correos o crear eventos, muestra los detalles al usuario para que los confirme antes de ejecutar la tarea."),
+        ("system", "Si te preguntan algo no relacionado con los servicios de OneCluster, responde que solo puedes ayudar con temas relacionados con la empresa y sus soluciones."),
+        ("system", "Evita mencionar o hacer alusión a las herramientas que utilizas internamente. Esa información es confidencial."),
+        ("placeholder", "{messages}"),
+    ]
+)
+
+# Definir el estado del asistente
+
+
+class State(TypedDict):
+    messages: Annotated[list, add_messages]
+    is_last_step: bool
+
+
+# Crear el graph con el estado definido
+graph = create_react_agent(
+    model=llm,
+    tools=tools,
+    state_schema=State,
+    state_modifier=system_prompt,
+    checkpointer=MemorySaver()
+)
+
+# Ruta de la API para procesar texto
+
+
+@app.route('/process_text', methods=['POST'])
+def process_text():
+    user_input = request.json.get('text')
+
+    # Procesar el texto con LangChain
+    events = graph.stream(
+        {"messages": [("user", user_input)], "is_last_step": False},
+        config={"configurable": {"thread_id": "thread-1", "recursion_limit": 50}},
+        stream_mode="updates"
+    )
+
+    # Preparar la respuesta
+    response = []
+    for event in events:
+        if "agent" in event:
+            response.append(event["agent"]["messages"][-1].content)
+
+    return jsonify({'response': response})
+
+
+# Ejecutar la app Flask
+if __name__ == '__main__':
+    app.run(port=5000)

api_openai/whisper.py

@@ -0,0 +1,45 @@
+from dotenv import load_dotenv
+import os
+from openai import OpenAI
+
+
+def whisper_api(file_path: str) -> str:
+    # cargar la variable del entorno desde el archivo .env
+    load_dotenv()
+
+    # Usar la variable de entorno API_KEY
+    api_key: str = os.getenv("OPENAI_API_KEY")
+
+    # Cargar el modelo whisper
+    client = OpenAI(api_key=api_key)
+
+    audio_file = open(file_path, "rb")
+    transcription = client.audio.transcriptions.create(
+        model="whisper-1",
+        file=audio_file,
+        response_format="text"
+    )
+
+    return transcription
+
+
+def tts_api(text: str):
+    # cargar la variable del entorno desde el archivo .env
+    load_dotenv()
+
+    # Usar la variable de entorno API_KEY
+    api_key: str = os.getenv("OPENAI_API_KEY")
+
+    # Cargar el modelo whisper
+    client = OpenAI(api_key=api_key)
+
+    speech_file_path = os.path.join('audios', 'voice.ogg')
+    response = client.audio.speech.create(
+        model="tts-1",
+        voice="alloy",
+        input=text
+    )
+
+    response.stream_to_file(speech_file_path)
+
+    return speech_file_path

credentials.json

@@ -0,0 +1 @@
+{"installed":{"client_id":"19011937557-bi5nh4afvg4tuqr87v6dp55qj9a9o1h2.apps.googleusercontent.com","project_id":"oc-aassistent","auth_uri":"https://accounts.google.com/o/oauth2/auth","token_uri":"https://oauth2.googleapis.com/token","auth_provider_x509_cert_url":"https://www.googleapis.com/oauth2/v1/certs","client_secret":"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","redirect_uris":["http://localhost"]}}

credentials_2.json

@@ -0,0 +1 @@
+{"installed":{"client_id":"629922809906-pl9l1ipout6d5hh19ku50sfvnqgu8ir2.apps.googleusercontent.com","project_id":"calendar-424503","auth_uri":"https://accounts.google.com/o/oauth2/auth","token_uri":"https://oauth2.googleapis.com/token","auth_provider_x509_cert_url":"https://www.googleapis.com/oauth2/v1/certs","client_secret":"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","redirect_uris":["http://localhost"]}}

langchain_tools/agent_tools.py

@@ -0,0 +1,323 @@
+from dotenv import load_dotenv
+from langchain_openai import ChatOpenAI
+from langchain_core.tools import tool
+from datetime import datetime, timezone
+from google.oauth2.credentials import Credentials
+from google_auth_oauthlib.flow import InstalledAppFlow
+from google.auth.transport.requests import Request
+from googleapiclient.discovery import build
+from rag.split_docs import load_split_docs
+from rag.llm import load_llm_openai
+from rag.embeddings import load_embeddins
+from rag.retriever import create_retriever
+from rag.vectorstore import create_verctorstore
+from rag.rag_chain import create_rag_chain
+from datetime import datetime
+import pytz
+import telebot
+import os
+
+
+class LangChainTools:
+    def load_llm_openai(self):
+        load_dotenv()
+        # model = "gpt-3.5-turbo-0125"
+        # model = "gpt-4o"
+        model = "gpt-4o-mini"
+
+        llm = ChatOpenAI(
+            model=model,
+            temperature=0.1,
+            max_tokens=2000,
+        )
+        return llm
+
+
+@tool
+def multiply(first_int: int, second_int: int) -> int:
+    """Multiply two integers together."""
+    return first_int * second_int
+
+
+@tool
+def redact_email(topic: str) -> str:
+    """Use this tool to draft the content of an email based on a topic."""
+
+    # Load LLM model
+    langChainTools = LangChainTools()
+
+    llm = langChainTools.load_llm_openai()
+    # Create prompt for the LLM
+    prompt = (
+        "Please redact a email based on the topic:\n\n"
+        "Topic: {}\n\n"
+        "Email Content: [Your email content here]"
+    ).format(topic)
+
+    response = llm.invoke(prompt)
+    return response
+
+
+@tool
+def list_calendar_events(max_results: int = 50) -> list:
+    """Use this tool to list upcoming calendar events."""
+
+    # Define los alcances que necesitamos para acceder a la API de Google Calendar
+    SCOPES = ['https://www.googleapis.com/auth/calendar']
+
+    creds = None
+
+    # La ruta al archivo token.json, que contiene los tokens de acceso y actualización
+    token_path = 'token_2.json'
+
+    # La ruta al archivo de credenciales de OAuth 2.1
+    creds_path = 'credentials_2.json'
+
+    # Cargar las credenciales desde el archivo token.json, si existe
+    if os.path.exists(token_path):
+        creds = Credentials.from_authorized_user_file(token_path, SCOPES)
+
+    # Si no hay credenciales válidas disponibles, inicia el flujo de OAuth 2.0 para obtener nuevas credenciales
+    if not creds or not creds.valid:
+        if creds and creds.expired and creds.refresh_token:
+            creds.refresh(Request())
+        else:
+            flow = InstalledAppFlow.from_client_secrets_file(
+                creds_path, SCOPES)
+            creds = flow.run_local_server(port=0)
+
+        # Guarda las credenciales para la próxima ejecución
+        with open(token_path, 'w') as token_file:
+            token_file.write(creds.to_json())
+
+    # Construye el objeto de servicio para interactuar con la API de Google Calendar
+    service = build('calendar', 'v3', credentials=creds)
+
+    # Identificador del calendario que deseas consultar. 'primary' se refiere al calendario principal del usuario.
+    calendar_id = 'primary'
+
+    # Realiza una llamada a la API para obtener una lista de eventos.
+    now = datetime.now(timezone.utc).isoformat()  # 'Z' indica UTC
+    events_result = service.events().list(
+        calendarId=calendar_id, timeMin=now, maxResults=max_results, singleEvents=True,
+        orderBy='startTime').execute()
+
+    # Extrae los eventos de la respuesta de la API.
+    events = events_result.get('items', [])
+
+    # Si no se encuentran eventos, imprime un mensaje.
+    if not events:
+        print('No upcoming events found.')
+        return
+
+    # Recorre la lista de eventos y muestra la hora de inicio y el resumen de cada evento.
+    for event in events:
+        # Obtiene la fecha y hora de inicio del evento. Puede ser 'dateTime' o 'date'.
+        start = event['start'].get('dateTime', event['start'].get('date'))
+        # Imprime la hora de inicio y el resumen (título) del evento.
+        print(start, event['summary'])
+
+    return events
+
+
+@tool
+def create_calendar_event(
+        title: str, start_time: datetime,
+        end_time: datetime, attendees: list) -> dict:
+    """Use this tool to create an event in the calendar.
+
+    Parameters:
+    - title: str - The title of the event.
+    - start_time: datetime - The start time of the event.
+    - end_time: datetime - The end time of the event.
+    - attendees: list - A list of attendee emails (required).
+
+    Returns:
+    - dict - The created event details.
+    """
+
+    if not attendees:
+        raise ValueError(
+            "El campo 'attendees' es obligatorio y no puede estar vacío.")
+
+    SCOPES = ['https://www.googleapis.com/auth/calendar']
+    creds = None
+
+    # La ruta al archivo token.json, que contiene los tokens de acceso y actualización
+    token_path = 'token_2.json'
+
+    # La ruta al archivo de credenciales de OAuth 2.0
+    creds_path = 'credentials_2.json'
+
+    # Cargar las credenciales desde el archivo token.json, si existe
+    if os.path.exists(token_path):
+        creds = Credentials.from_authorized_user_file(token_path, SCOPES)
+
+    # Si no hay credenciales válidas disponibles, inicia el flujo de OAuth 2.0 para obtener nuevas credenciales
+    if not creds or not creds.valid:
+        if creds and creds.expired and creds.refresh_token:
+            creds.refresh(Request())
+        else:
+            flow = InstalledAppFlow.from_client_secrets_file(
+                creds_path, SCOPES)
+            creds = flow.run_local_server(port=0)
+
+        # Guarda las credenciales para la próxima ejecución
+        with open(token_path, 'w') as token_file:
+            token_file.write(creds.to_json())
+
+    # Construye el objeto de servicio para interactuar con la API de Google Calendar
+    service = build('calendar', 'v3', credentials=creds)
+
+    # Validar y filtrar asistentes
+    valid_attendees = []
+    for email in attendees:
+        if isinstance(email, str) and '@' in email:
+            valid_attendees.append({'email': email})
+        else:
+            raise ValueError(f"'{email}' no es un correo electrónico válido.")
+
+    # Identificador del calendario que deseas modificar. 'primary' se refiere al calendario principal del usuario.
+    calendar_id = 'primary'
+
+    # Define el cuerpo del evento con el título, la hora de inicio y la hora de finalización
+    event = {
+        'summary': title,
+        'start': {
+            'dateTime': start_time.strftime('%Y-%m-%dT%H:%M:%S'),
+            'timeZone': 'America/Bogota',
+        },
+        'end': {
+            'dateTime': end_time.strftime('%Y-%m-%dT%H:%M:%S'),
+            'timeZone': 'America/Bogota',
+        },
+        'attendees': valid_attendees
+    }
+
+    try:
+        # Crea el evento en el calendario
+        event = service.events().insert(calendarId=calendar_id, body=event).execute()
+        print('Event created: %s' % (event.get('htmlLink')))
+    except Exception as e:
+        print(f"Error al crear el evento: {e}")
+        return {}
+
+    return event
+
+
+@tool
+def create_quick_add_event(quick_add_text: str):
+    """Use this tool to create events in the calendar from natural language, 
+    using the Quick Add feature of Google Calendar.
+    """
+    quick_add_text: str = input(
+        "- Escribe la descripcion del evento que quieres crear: ")
+    SCOPES = ['https://www.googleapis.com/auth/calendar']
+
+    creds = None
+
+    # La ruta al archivo token.json, que contiene los tokens de acceso y actualización
+    token_path = 'token_2.json'
+
+    # La ruta al archivo de credenciales de OAuth 2.0
+    creds_path = 'credentials_2.json'
+
+    # Cargar las credenciales desde el archivo token.json, si existe
+    if os.path.exists(token_path):
+        creds = Credentials.from_authorized_user_file(token_path, SCOPES)
+
+    # Si no hay credenciales válidas disponibles, inicia el flujo de OAuth 2.0 para obtener nuevas credenciales
+    if not creds or not creds.valid:
+        if creds and creds.expired and creds.refresh_token:
+            creds.refresh(Request())
+        else:
+            flow = InstalledAppFlow.from_client_secrets_file(
+                creds_path, SCOPES)
+            creds = flow.run_local_server(port=0)
+
+        # Guarda las credenciales para la próxima ejecución
+        with open(token_path, 'w') as token_file:
+            token_file.write(creds.to_json())
+
+    # Construye el objeto de servicio para interactuar con la API de Google Calendar
+    service = build('calendar', 'v3', credentials=creds)
+
+    # Identificador del calendario que deseas modificar. 'primary' se refiere al calendario principal del usuario.
+    calendar_id = 'primary'
+
+    # Crea el evento utilizando la funcionalidad Quick Add
+    event = service.events().quickAdd(
+        calendarId=calendar_id, text=quick_add_text).execute()
+
+    print('Event created: %s' % (event.get('htmlLink')))
+
+    return event
+
+
+# @tool
+# def send_message(message: str):
+#     """Use this function when you need to communicate with the user."""
+#     # Configuración del bot
+#     load_dotenv()
+#     API_TOKEN_BOT = os.getenv("API_TOKEN_BOT")
+#     bot = telebot.TeleBot(API_TOKEN_BOT)
+#
+#     bot.send_message(chat_id="5076346205", text=message)
+#
+
+@tool
+def send_message(message: str):
+    """Use this function when you need to communicate with Cristian."""
+    # Configuración del bot
+    load_dotenv()
+    API_TOKEN_BOT = os.getenv("API_TOKEN_BOT")
+    bot = telebot.TeleBot(API_TOKEN_BOT)
+
+    # Escapar caracteres especiales en Markdown
+    from telebot.util import escape_markdown
+    safe_message = escape_markdown(message)
+
+    # Enviar mensaje usando MarkdownV2
+    bot.send_message(chat_id="5076346205", text=safe_message,
+                     parse_mode="Markdown")
+
+
+@tool
+def get_company_info(prompt: str) -> str:
+    """Use this function when you need more information about the services offered by OneCluster."""
+    file_path: str = 'onecluster_info.pdf'
+
+    docs_split: list = load_split_docs(file_path)
+    embeddings_model = load_embeddins()
+    llm = load_llm_openai()
+    create_verctorstore(
+        docs_split,
+        embeddings_model,
+        file_path
+    )
+    retriever = create_retriever(
+        embeddings_model,
+        persist_directory="embeddings/onecluster_info"
+    )
+    qa = create_rag_chain(
+        llm, retriever)
+
+    # prompt: str = "Escribe un parrarfo describiendo cuantos son y  cuales son los servicios que ofrece OneCluster y brinda detalles sobre cada uno."
+    response = qa.invoke(
+        {"input": prompt, "chat_history": []}
+    )
+
+    return response["answer"]
+
+
+@tool
+def get_current_date_and_time():
+    """
+    Use this function when you need to know the current date and time.
+
+    Returns:
+        str: Current date and time in Bogotá, Colombia.
+    """
+    bogota_tz = pytz.timezone('America/Bogota')
+    current_date_and_time = datetime.now(bogota_tz)
+    return current_date_and_time.strftime('%Y-%m-%d %H:%M:%S')

langchain_tools/agents.py

@@ -0,0 +1,83 @@
+from langchain_core.tools import tool
+from langchain_community.tools.gmail.utils import (
+    build_resource_service,
+    get_gmail_credentials,
+)
+from langchain_community.agent_toolkits import GmailToolkit
+from langchain import hub
+from langchain_community.tools.tavily_search import TavilySearchResults
+from dotenv import load_dotenv
+from langchain.agents import AgentExecutor, create_openai_functions_agent
+from langchain_tools.agent_tools import (
+    multiply, redact_email, list_calendar_events,
+    create_calendar_event, create_quick_add_event,
+    send_message, get_company_info,
+    get_current_date_and_time
+)
+
+
+class AgentTools:
+
+    def load_tools(self) -> list:
+
+        toolkit = GmailToolkit()
+
+        # Can review scopes here https://developers.google.com/gmail/api/auth/scopes
+        # For instance, readonly scope is 'https://www.googleapis.com/auth/gmail.readonly'
+        credentials = get_gmail_credentials(
+            token_file="token.json",
+            scopes=["https://mail.google.com/"],
+            client_secrets_file="credentials.json",)
+        api_resource = build_resource_service(credentials=credentials)
+        toolkit = GmailToolkit(api_resource=api_resource)
+
+        # creamos la lista de herramientas de gmail
+        tools = toolkit.get_tools()
+
+        load_dotenv()
+
+        # Agregamos otras tools
+        search = TavilySearchResults(max_results=1)
+
+        tools.append(search)
+        tools.append(multiply)
+        tools.append(redact_email)
+        tools.append(list_calendar_events)
+        tools.append(create_calendar_event)
+        tools.append(send_message)
+        tools.append(get_company_info),
+        tools.append(get_current_date_and_time)
+        # tools.append(create_quick_add_event)
+
+        return tools
+
+    def load_agent(self, llm, tools):
+        instructions = """
+You are the virtual assistant of OneCluster, a company specialized in providing custom development services focused on creating personalized technological solutions for businesses and companies. Your mission is to offer a warm, friendly, and collaborative service that always reflects OneCluster's core values.
+
+**User Interactions:**
+1. **Initial Greeting:** When starting an interaction with a user, greet them courteously and identify who you have the pleasure of speaking with. Once you know the user's name, address them respectfully throughout the conversation.
+
+2. **Providing Information:** You have the ability to offer clear and detailed information about the services provided by OneCluster. Make sure to be concise yet informative, adapting the information to the user's needs.
+
+3. **Appointment Scheduling:** You are responsible for scheduling appointments for clients. Before confirming an appointment, always check the availability on OneCluster's calendar to ensure there is space, and check the current date and time so that you have a clear sense of time. Request an email address from the user to schedule the appointment.
+
+4. **Handling Unanswered Questions:** If you do not know how to answer a question, politely ask for the client's contact information and clearly identify the problem to be resolved. Then, send this information to oneclustererp@gmail.com with the subject "Unresolved customer query by the agent." Inform the client that you do not have the information at your disposal but that you can escalate the request to the support team, who will respond promptly.
+
+**Style and Tone:**
+Maintain a tone that is always friendly, approachable, and professional. Each interaction should reflect OneCluster's commitment to innovation, adaptability, and ongoing collaboration.
+"""
+
+        base_prompt = hub.pull("langchain-ai/openai-functions-template")
+
+        prompt = base_prompt.partial(instructions=instructions)
+
+        agent = create_openai_functions_agent(llm, tools, prompt)
+
+        agent_executor = AgentExecutor(
+            agent=agent,
+            tools=tools,
+            verbose=True,
+        )
+
+        return agent_executor

oc_assistant.py

@@ -0,0 +1,109 @@
+from langchain_community.tools.tavily_search import TavilySearchResults
+from typing import Annotated
+from typing_extensions import TypedDict
+from langgraph.graph.message import add_messages
+from langchain_openai import ChatOpenAI
+from dotenv import load_dotenv
+from langgraph.prebuilt import create_react_agent
+from langchain_core.prompts import ChatPromptTemplate
+from langgraph.checkpoint.memory import MemorySaver
+from langchain_tools.agent_tools import (
+    redact_email, list_calendar_events,
+    create_calendar_event,
+    get_company_info,
+    get_current_date_and_time
+)
+from langchain_community.tools.gmail.utils import (
+    build_resource_service,
+    get_gmail_credentials,
+)
+from langchain_community.agent_toolkits import GmailToolkit
+
+
+load_dotenv()
+
+
+# Inicialiamos un LLM de OpenaIA
+llm = ChatOpenAI(
+    model="gpt-4o-mini",
+    temperature=0.9
+)
+
+toolkit = GmailToolkit()
+
+# Can review scopes here https://developers.google.com/gmail/api/auth/scopes
+# For instance, readonly scope is 'https://www.googleapis.com/auth/gmail.readonly'
+credentials = get_gmail_credentials(
+    token_file="token.json",
+    scopes=["https://mail.google.com/"],
+    client_secrets_file="credentials.json",
+)
+api_resource = build_resource_service(credentials=credentials)
+toolkit = GmailToolkit(api_resource=api_resource)
+
+# creamos la lista de herramientas de gmail
+tools = toolkit.get_tools()
+
+search = TavilySearchResults(max_results=2)
+tools.append(search)
+tools.append(redact_email)
+tools.append(list_calendar_events)
+tools.append(create_calendar_event)
+tools.append(get_company_info)
+tools.append(get_current_date_and_time)
+
+system_prompt = ChatPromptTemplate.from_messages(
+    [
+        ("system", "Eres Mariana, el asistente virtual de OneCluster, una empresa de software que ofrece soluciones personalizadas. Asume el tono de J.A.R.V.I.S.: cordial, atento y con tacto en todo momento."),
+
+        # Instrucciones sobre presentación y tono
+        ("system", "Preséntate como Mariana en el primer mensaje y pregunta el nombre del usuario si no lo tienes registrado."),
+        ("system", "Si el usuario ya ha interactuado antes, usa su nombre sin necesidad de volver a preguntar."),
+        ("system", "Si el primer mensaje del usuario es una solicitud, pregúntale su nombre antes de responder si aún no lo conoces."),
+
+        # Instrucciones sobre el manejo de solicitudes y tareas
+        ("system", "OneCluster es una empresa de software especializada en desarrollo a medida. Solo responde a preguntas y solicitudes relacionadas con la empresa y sus servicios."),
+        ("system", "Si necesitas información adicional sobre la empresa, usa la función get_company_info."),
+        ("system", "Antes de enviar correos o crear eventos, muestra los detalles al usuario para que los confirme antes de ejecutar la tarea."),
+
+        # Cómo manejar preguntas fuera del alcance
+        ("system", "Si te preguntan algo no relacionado con los servicios de OneCluster, responde que solo puedes ayudar con temas relacionados con la empresa y sus soluciones."),
+
+        # Prohibición de revelar herramientas internas
+        ("system", "Evita mencionar o hacer alusión a las herramientas que utilizas internamente. Esa información es confidencial."),
+
+        # Placeholder para el contenido dinámico de la conversación
+        ("placeholder", "{messages}"),
+    ]
+)
+
+
+class State(TypedDict):
+    messages: Annotated[list, add_messages]
+    is_last_step: bool  # Cambiar a booleano si es necesario
+
+
+# Creamos el graph con el estado definido
+graph= create_react_agent(
+    model = llm, tools = tools, state_schema = State,
+    state_modifier = system_prompt,
+    checkpointer = MemorySaver()
+)
+
+config= {"configurable": {"thread_id": "thread-1", "recursion_limit": 50}}
+
+
+while True:
+    user_input = input("User: ")
+    if user_input.lower() in ["quit", "exit", "q"]:
+        print("Goodbye!")
+        break
+
+    events = graph.stream(
+        {"messages": [("user", user_input)],
+        "is_last_step": False},
+        config, stream_mode = "updates")
+
+    for event in events:
+        if "agent" in event:
+            print(f"\nAsistente: {event["agent"]["messages"][-1].content}\n")

onecluster_bot.py

@@ -0,0 +1,109 @@
+import telebot
+from dotenv import load_dotenv
+import os
+from api_openai.whisper import whisper_api, tts_api
+from langchain_tools.agent_tools import LangChainTools
+from langchain_tools.agents import AgentTools
+from langchain_core.messages import AIMessage, HumanMessage
+# from tools.scaped import scaped
+
+# Configuración del bot
+load_dotenv()
+API_TOKEN_BOT = os.getenv("API_TOKEN_BOT")
+bot = telebot.TeleBot(API_TOKEN_BOT)
+
+# Handle '/start' and '/help'
+wellcome = "¡Bienvenido! ¿Cómo puedo ayudarte?"
+
+
+@bot.message_handler(commands=['help', 'start'])
+def send_welcome(message):
+    bot.reply_to(message, wellcome, parse_mode="Markdown")
+
+
+# Creamos una lista para el historial fuera de las funciones
+history = []
+
+
+@bot.message_handler(content_types=["text", "voice"])
+def bot_mensajes(message):
+    global history  # Para acceder a la variable global 'history'
+
+    # Si el mensaje es una nota de voz
+    if message.voice:
+        user_name = message.from_user.first_name
+        file_info = bot.get_file(message.voice.file_id)
+        downloaded_file = bot.download_file(file_info.file_path)
+        file_path = "audios/nota_de_voz.ogg"
+
+        with open(file_path, 'wb') as new_file:
+            new_file.write(downloaded_file)
+
+        pregunta_usuario = whisper_api(file_path)
+        print(f"Pregunta del usuario: {pregunta_usuario}")
+        langChainTools = LangChainTools()
+        llm = langChainTools.load_llm_openai()
+
+        agentTools = AgentTools()
+        tools = agentTools.load_tools()
+        agent_executor = agentTools.load_agent(llm, tools)
+
+        respuesta_agente = agent_executor.invoke(
+            {
+                "input": pregunta_usuario,
+                "chat_history": history,
+            }
+        )
+
+        bot.send_message(message.chat.id, respuesta_agente["output"],
+                         parse_mode="Markdown")
+
+        path_voice: str = tts_api(respuesta_agente["output"])
+        with open(path_voice, 'rb') as voice:
+            bot.send_voice(message.chat.id, voice=voice)
+
+        history.append(HumanMessage(content=pregunta_usuario))
+        history.append(AIMessage(content=respuesta_agente["output"]))
+
+    # Si el mensaje es de texto
+    if message.text:
+        pregunta_usuario = message.text
+        langChainTools = LangChainTools()
+        llm = langChainTools.load_llm_openai()
+
+        agentTools = AgentTools()
+        tools = agentTools.load_tools()
+        agent_executor = agentTools.load_agent(llm, tools)
+
+        respuesta_agente = agent_executor.invoke(
+            {
+                "input": pregunta_usuario,
+                "chat_history": history,
+            }
+        )
+
+        # texto_respuesta: str = scaped(respuesta_agente["output"])
+        texto_respuesta: str = respuesta_agente["output"]
+        bot.send_message(
+            message.chat.id, texto_respuesta,
+            parse_mode="Markdown")
+
+        # Mandar mensaje de voz
+        # path_voice: str = tts_api(respuesta_agente["output"])
+        # with open(path_voice, 'rb') as voice:
+        #     bot.send_voice(message.chat.id, voice=voice)
+
+        history.append(HumanMessage(content=pregunta_usuario))
+        history.append(AIMessage(content=respuesta_agente["output"]))
+        # print(history)
+
+    # Enviar el historial después de cada interacción
+    # bot.send_message(message.chat.id, history)
+
+
+# while True:
+#     time.sleep(60)
+#     mensaje = 'Que mas pues!!'
+#     bot.send_message('5076346205', mensaje)
+
+bot.infinity_polling()

prueba_rag.py

@@ -0,0 +1,32 @@
+from rag.split_docs import load_split_docs
+from rag.llm import load_llm_openai
+from rag.embeddings import load_embeddins
+from rag.retriever import create_retriever
+from rag.vectorstore import create_verctorstore
+from rag.rag_chain import create_rag_chain
+
+dir_pdfs: str = 'documents/pdfs/'
+file_name: str = 'onecluster_info.pdf'
+file_path: str = 'onecluster_info.pdf'
+
+docs_split: list = load_split_docs(file_path)
+embeddings_model = load_embeddins()
+llm = load_llm_openai()
+create_verctorstore(
+    docs_split,
+    embeddings_model,
+    file_path
+)
+retriever = create_retriever(
+    embeddings_model,
+    persist_directory="embeddings/onecluster_info"
+)
+qa = create_rag_chain(
+    llm, retriever)
+
+prompt: str = "Dame información detallada sobre los sercivios que ofrese OneCluster."
+respuesta = qa.invoke(
+    {"input": prompt, "chat_history": []}
+)
+
+print(respuesta["answer"])

rag/embeddings.py

@@ -0,0 +1,10 @@
+from dotenv import load_dotenv
+from langchain_openai import OpenAIEmbeddings
+
+
+def load_embeddins():
+    load_dotenv()
+    # model = "text-embedding-ada-002"
+    model = "text-embedding-3-small"
+
+    return OpenAIEmbeddings(model=model)

rag/llm.py

@@ -0,0 +1,17 @@
+from dotenv import load_dotenv
+from langchain_openai import ChatOpenAI
+
+
+def load_llm_openai():
+    load_dotenv()
+    # model = "gpt-3.5-turbo-0125"
+    # model = "gpt-4o"
+    model = "gpt-4o-mini"
+
+    llm = ChatOpenAI(
+        model=model,
+        temperature=0.1,
+        max_tokens=2000,
+    )
+
+    return llm

rag/rag_chain.py

@@ -0,0 +1,40 @@
+from langchain.chains import create_history_aware_retriever
+from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
+from langchain.chains import create_retrieval_chain
+from langchain.chains.combine_documents import create_stuff_documents_chain
+
+
+def create_rag_chain(llm, retriever):
+    contextualize_q_system_prompt = """Given a chat history and the latest user question \
+    which might reference context in the chat history, formulate a standalone question \
+    which can be understood without the chat history. Do NOT answer the question, \
+    just reformulate it if needed and otherwise return it as is."""
+    contextualize_q_prompt = ChatPromptTemplate.from_messages(
+        [
+            ("system", contextualize_q_system_prompt),
+            MessagesPlaceholder("chat_history"),
+            ("human", "{input}"),
+        ]
+    )
+    history_aware_retriever = create_history_aware_retriever(
+        llm, retriever, contextualize_q_prompt
+    )
+
+    # ___________________Chain con el chat history_______________________-
+    qa_system_prompt = """You are an assistant for question-answering tasks.  \
+    Use the following pieces of retrieved context to answer the question. \
+    If you don't know the answer, just say that you don't know. \
+    The length of the answer should be sufficient to address what is being asked, \
+    meaning don't limit yourself in length.\
+    
+    {context}"""
+    qa_prompt = ChatPromptTemplate.from_messages(
+        [
+            ("system", qa_system_prompt),
+            MessagesPlaceholder("chat_history"),
+            ("human", "{input}"),
+        ]
+    )
+    question_answer_chain = create_stuff_documents_chain(llm, qa_prompt)
+
+    return create_retrieval_chain(history_aware_retriever, question_answer_chain)

rag/retriever.py

@@ -0,0 +1,16 @@
+from langchain_chroma import Chroma
+
+
+def create_retriever(embeddings, persist_directory: str):
+    # Cargamos la vectorstore
+    # vectordb = Chroma.from_documents(
+    #     persist_directory=st.session_state.persist_directory,  # Este es el directorio del la vs del docuemnto del usuario que se encuentra cargado en la session_state.
+    #     embedding_function=embeddings,
+    # )
+    vectordb = Chroma(
+        persist_directory=persist_directory,
+        embedding_function=embeddings,
+    )
+
+    # Creamos el retriver para que retorne los fragmentos mas relevantes.
+    return vectordb.as_retriever(search_kwargs={"k": 6})

rag/split_docs.py

@@ -0,0 +1,19 @@
+from langchain_community.document_loaders.pdf import PyPDFLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+import os
+
+
+def load_split_docs(file_name: str) -> list:
+
+    file_path: str = os.path.join("documents", "pdfs", file_name)
+    loader = PyPDFLoader(file_path)
+    docs: list = loader.load()
+    chunk_size: int = 2000
+    chunk_overlap: int = 300
+
+    splitter = RecursiveCharacterTextSplitter(
+        chunk_size=chunk_size, chunk_overlap=chunk_overlap
+    )
+    docs_split: list = splitter.split_documents(docs)
+
+    return docs_split

rag/vectorstore.py

@@ -0,0 +1,15 @@
+from langchain_chroma import Chroma
+import os
+
+
+def create_verctorstore(docs_split: list, embeddings, file_name: str):
+    db_name: str = file_name.replace(".pdf", "").replace(" ", "_").lower()
+
+    persist_directory: str = f"embeddings/{db_name}"
+
+    if not os.path.exists(persist_directory):
+        vectordb = Chroma.from_documents(
+            persist_directory=persist_directory,
+            documents=docs_split,
+            embedding=embeddings,
+        )

requirements.txt

@@ -0,0 +1,140 @@
+aiohappyeyeballs==2.4.0
+aiohttp==3.10.6
+aiosignal==1.3.1
+annotated-types==0.7.0
+anyio==4.6.0
+asgiref==3.8.1
+attrs==24.2.0
+backoff==2.2.1
+bcrypt==4.2.0
+beautifulsoup4==4.12.3
+blinker==1.8.2
+build==1.2.2.post1
+cachetools==5.5.0
+certifi==2024.8.30
+charset-normalizer==3.3.2
+chroma-hnswlib==0.7.3
+chromadb==0.5.3
+click==8.1.7
+coloredlogs==15.0.1
+dataclasses-json==0.6.7
+Deprecated==1.2.14
+distro==1.9.0
+durationpy==0.9
+fastapi==0.115.4
+filelock==3.16.1
+Flask==3.0.3
+flatbuffers==24.3.25
+frozenlist==1.4.1
+fsspec==2024.10.0
+google==3.0.0
+google-api-core==2.19.0
+google-api-python-client==2.131.0
+google-auth==2.29.0
+google-auth-httplib2==0.2.0
+google-auth-oauthlib==1.2.0
+googleapis-common-protos==1.63.0
+greenlet==3.1.1
+grpcio==1.67.0
+h11==0.14.0
+httpcore==1.0.5
+httplib2==0.22.0
+httptools==0.6.4
+httpx==0.27.2
+huggingface-hub==0.26.2
+humanfriendly==10.0
+idna==3.10
+importlib_metadata==8.4.0
+importlib_resources==6.4.5
+itsdangerous==2.2.0
+Jinja2==3.1.4
+jiter==0.5.0
+jsonpatch==1.33
+jsonpointer==3.0.0
+kubernetes==31.0.0
+langchain==0.3.1
+langchain-chroma==0.1.4
+langchain-community==0.3.1
+langchain-core==0.3.6
+langchain-openai==0.2.0
+langchain-text-splitters==0.3.0
+langgraph==0.2.28
+langgraph-checkpoint==1.0.11
+langsmith==0.1.128
+markdown-it-py==3.0.0
+MarkupSafe==2.1.5
+marshmallow==3.22.0
+mdurl==0.1.2
+mmh3==5.0.1
+monotonic==1.6
+mpmath==1.3.0
+msgpack==1.1.0
+multidict==6.1.0
+mypy-extensions==1.0.0
+numpy==1.26.4
+oauthlib==3.2.2
+onnxruntime==1.19.2
+openai==1.48.0
+opentelemetry-api==1.27.0
+opentelemetry-exporter-otlp-proto-common==1.27.0
+opentelemetry-exporter-otlp-proto-grpc==1.27.0
+opentelemetry-instrumentation==0.48b0
+opentelemetry-instrumentation-asgi==0.48b0
+opentelemetry-instrumentation-fastapi==0.48b0
+opentelemetry-proto==1.27.0
+opentelemetry-sdk==1.27.0
+opentelemetry-semantic-conventions==0.48b0
+opentelemetry-util-http==0.48b0
+orjson==3.10.7
+overrides==7.7.0
+packaging==24.1
+posthog==3.7.0
+proto-plus==1.24.0
+protobuf==4.25.5
+pyasn1==0.6.1
+pyasn1_modules==0.4.1
+pydantic==2.9.2
+pydantic-settings==2.5.2
+pydantic_core==2.23.4
+Pygments==2.18.0
+pyparsing==3.1.4
+pypdf==5.1.0
+PyPika==0.48.9
+pyproject_hooks==1.2.0
+pyTelegramBotAPI==4.23.0
+python-dateutil==2.9.0.post0
+python-dotenv==1.0.1
+pytz==2024.2
+PyYAML==6.0.2
+regex==2024.9.11
+requests==2.32.3
+requests-oauthlib==2.0.0
+rich==13.9.3
+rsa==4.9
+setuptools==75.2.0
+shellingham==1.5.4
+six==1.16.0
+sniffio==1.3.1
+soupsieve==2.6
+SQLAlchemy==2.0.35
+starlette==0.41.2
+sympy==1.13.3
+telebot==0.0.5
+tenacity==8.5.0
+tiktoken==0.7.0
+tokenizers==0.20.1
+tqdm==4.66.5
+typer==0.12.5
+typing-inspect==0.9.0
+typing_extensions==4.12.2
+uritemplate==4.1.1
+urllib3==2.2.3
+uvicorn==0.32.0
+uvloop==0.21.0
+watchfiles==0.24.0
+websocket-client==1.8.0
+websockets==13.1
+Werkzeug==3.0.4
+wrapt==1.16.0
+yarl==1.12.1
+zipp==3.20.2

token.json

@@ -0,0 +1 @@
+{"token": "ya29.a0AeDClZBjncDp4ZwNKNtQ5ghKHPr1IT4XkgDc9QtvhPLrFGAR84f5r5iZPCd91VB7_WoJCG3iGQS0MU1n01xdRlEjDl7wVlKjKF0H680Bdim_bzykCXn3Jj0nVVkkHDOZP7RWeP1oAfY7Vjd4qbw_VxOdOzVzG_Bc6Auy4EJINAaCgYKAcYSARASFQHGX2MipaJllxIRMLCcZb2csCZECA0177", "refresh_token": "1//05nbircha66xlCgYIARAAGAUSNwF-L9IrxbE2v7kfLwXb4u0pD6Rin7xEBOTT83DeH7t2ttfD5CDmUCyhDsOaVRMRK_r8UtdoMq8", "token_uri": "https://oauth2.googleapis.com/token", "client_id": "19011937557-bi5nh4afvg4tuqr87v6dp55qj9a9o1h2.apps.googleusercontent.com", "client_secret": "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", "scopes": ["https://mail.google.com/"], "universe_domain": "googleapis.com", "account": "", "expiry": "2024-10-30T01:16:56.882894Z"}

token_2.json

@@ -0,0 +1 @@
+{"token": "ya29.a0AeDClZChMN7SEvjp3dFVZtee2pDoqAoPFC7AWiEeIG7H6qN2HDnf7c6DcFuc--aG60e1cAnOpoKf80H8aqrFFYbF4-F4LE_vz9MY8oc21Ra9PwM16FYxGGKcM2wcrrOGaFncs9Um9_yNxzAa6MUVNq88Y_Bhpr2F2mO3o53NjQaCgYKAZESARASFQHGX2Mi7EodrKchyiyPIZ4y5Lwh0Q0177", "refresh_token": "1//05CtNC-Z3ii8qCgYIARAAGAUSNwF-L9IrbOfrB0kNACEJ5HX4T-fmdNUqsGFqn1QFlvK_1L9h0emULUS1yU85IbaNyESXZSQzHU8", "token_uri": "https://oauth2.googleapis.com/token", "client_id": "629922809906-pl9l1ipout6d5hh19ku50sfvnqgu8ir2.apps.googleusercontent.com", "client_secret": "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", "scopes": ["https://www.googleapis.com/auth/calendar"], "universe_domain": "googleapis.com", "account": "", "expiry": "2024-10-30T01:22:34.287442Z"}