From 829a475123d7b3218e88bbab1cfde2c188240807 Mon Sep 17 00:00:00 2001
From: faiber1986 <montesgomez1945@gmail.com>
Date: Fri, 6 Mar 2026 14:33:28 -0500
Subject: [PATCH] =?UTF-8?q?Archivo=20con=20interfaz=20en=20streamlit=20par?=
 =?UTF-8?q?a=20consultor=20t=C3=A9cnico.=20Para=20su=20despliegue=20es=20n?=
 =?UTF-8?q?ecesario=20instalar=20las=20dependencias=20necesarias=20en=20tu?=
 =?UTF-8?q?=20ambiente=20(en=20mi=20caso=20estoy=20usando=20uv=20como=20ge?=
 =?UTF-8?q?stor)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 week1/app.py | 73 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 73 insertions(+)
 create mode 100644 week1/app.py

diff --git a/week1/app.py b/week1/app.py
new file mode 100644
index 0000000000..ba601483de
--- /dev/null
+++ b/week1/app.py
@@ -0,0 +1,73 @@
+import streamlit as st
+import ollama
+from openai import OpenAI
+from dotenv import load_dotenv
+
+# load environment variables
+load_dotenv()
+
+# initialize OpenAI client
+openai_client = OpenAI()
+
+MODEL_GPT = 'gpt-4o-mini'
+MODEL_LLAMA = 'llama3.2'
+
+# Constants
+SYSTEM_PROMPT = "Eres un tutor técnico útil que responde preguntas sobre código Python, ingeniería de software, ciencia de datos y LLM"
+
+st.title("Tutor Técnico LLM 🤖")
+st.markdown("Haz una pregunta técnica y selecciona el modelo con el que quieres obtener tu respuesta.")
+
+# Create the layout
+model_choice = st.selectbox(
+    "Selecciona el modelo:",
+    (MODEL_GPT, MODEL_LLAMA)
+)
+
+question = st.text_area("Ingresa tu pregunta técnica aquí:", height=150)
+
+if st.button("Enviar"):
+    if not question.strip():
+        st.warning("Por favor, ingresa una pregunta antes de enviar.")
+    else:
+        user_prompt = "Por favor, da una explicación detallada de la siguiente pregunta: " + question
+        
+        messages = [
+            {"role": "system", "content": SYSTEM_PROMPT},
+            {"role": "user", "content": user_prompt}
+        ]
+        
+        st.subheader("Respuesta:")
+        
+        if model_choice == MODEL_GPT:
+            # We use an empty placeholder to stream the response
+            response_placeholder = st.empty()
+            full_response = ""
+            
+            with st.spinner("Esperando respuesta de GPT-4o-mini..."):
+                try:
+                    stream = openai_client.chat.completions.create(
+                        model=MODEL_GPT, 
+                        messages=messages,
+                        stream=True
+                    )
+                    
+                    for chunk in stream:
+                        content = chunk.choices[0].delta.content or ''
+                        full_response += content
+                        # We progressively write the response markdown
+                        response_placeholder.markdown(full_response + "▌")
+                    
+                    # Remove the cursor at the end
+                    response_placeholder.markdown(full_response)
+                except Exception as e:
+                    st.error(f"Error al llamar a OpenAI: {e}")
+                    
+        elif model_choice == MODEL_LLAMA:
+            with st.spinner("Esperando respuesta de Llama 3.2..."):
+                try:
+                    response = ollama.chat(model=MODEL_LLAMA, messages=messages)
+                    reply = response['message']['content']
+                    st.markdown(reply)
+                except Exception as e:
+                    st.error(f"Error al llamar a Ollama: {e}")